
Safeguarding Democracy
EU Development at the Nexus of Elections, Information Integrity and Artificial Intelligence
Over the past decade, the importance of information integrity for democracy has become clearer, as it underpins societal trust and enables citizens to make informed choices and participate meaningfully in democratic life. Information integrity is also key to ensuring meaningful accountability on the part of representatives, and as the information space has become more polluted with false narratives and disinformation in recent years, the public space for discussion is being increasingly disrupted. The availability and sophistication of tools driven by artificial intelligence (AI) have skyrocketed over the past few years, potentially creating more acute dangers for participation and trust. Generative AI—systems that go beyond processing data to create unique synthetic content on the basis of this data—has changed the nature of online communications. Crucially, a dynamic is emerging whereby citizens may find it increasingly difficult to differentiate between what is real and what is fake, and where bad actors manipulate the online information ecosystem to produce alternative realities.
Discussions continue to take place globally regarding how governments, oversight institutions, civil society, international bodies and the private sector should respond to the challenges to democracy and elections posed by the pollution of information ecosystems. Given the growing demand for support, facilitating a resilient response can create opportunities for building international partnerships that have a positive impact.
This report investigates the key risks and opportunities associated with AI in developing contexts, mechanisms of information pollution and countermeasures being pursued, and, most importantly, the differences between the context in developing and developed nations.
The purpose of this report is to support development practitioners in identifying opportunities to promote information integrity and the ethical use of AI, while maintaining a long-term vision rooted in holistic support for democracy. The research for this report therefore took a non-prescriptive approach, drawing on insights from case studies to highlight key actors to collaborate with, principles to pursue, and components of a long-term vision for developing standards and engaging the public in finding solutions. Insights are drawn from case studies on Bangladesh, Ghana, Indonesia, Mexico, Mongolia, Pakistan and South Africa.
In the context of electoral periods, pollution of the online information space with synthetic and misleading content could create challenges for voters who wish to educate themselves and make an informed choice. Coordinated influence campaigns aimed at manipulating online information ecosystems were identified in all seven of the case studies. Between electoral periods, the misuse of these tools can contribute to increased polarization and extremism, a decay in societal trust and a reduction in people’s ability to evaluate the truth of the information they receive. Moreover, there are challenges concerning personal data privacy, with cybersecurity risks also providing openings for manipulation by bad-faith actors. In addition, false narratives frequently target women and other vulnerable minorities, multiplying dangers for these groups, a phenomenon identified especially in Mexico and Pakistan. Generative AI has the potential to accelerate these risks, as tools powered by this new technology can increase the scale and sophistication of harmful content considerably.
In spite of these risks, new communications technologies and AI also provide a number of important opportunities. They have lowered the barriers to and costs of engaging with large audiences, which can be especially helpful for new or disadvantaged actors with fewer resources. Where personal data is used transparently and responsibly, actors can tailor their communications based on advanced data analytics, while ensuring that sensitive data—for example, on race and sexual orientation—is off-limits. AI-powered tools can also help to translate communications and provide better accessibility, giving actors the means to campaign in highly repressive contexts.
Importantly, as the studies show, the manipulation of online narratives is not just about spreading advantageous false information and slandering one’s opponents; rather, it is about polluting the flow of information both online and offline with false narratives that are often spread by local elites to hold on to power.
Developing contexts can be more vulnerable to information manipulation—by both domestic and foreign actors—as they lack the state, technical and civic capacity to engage in meaningful oversight and disrupt the chain of influence. Generative AI increases the scale of these narratives, also by making it possible to create far more believable synthetic media.
It is important to note, however, that online communication tools and generative AI have also been used by those resisting repression, and the use of generative AI has not yet had a serious impact on election campaigning. AI-powered tools also show promise in enhancing oversight and monitoring of political communications by flagging disinformation and other harmful content.
The risks associated with pollution of the information ecosystem are significant: public trust may be seriously weakened, and people, being more susceptible to polarization and violent extremism, may feel less motivated to vote or participate in democratic debate. Moreover, false narratives are frequently aimed at delegitimizing the electoral administration and results, as well as undermining oversight institutions and other key mechanisms of democracy intended to ensure fair competition and checks and balances on power. The dissemination of such narratives helps to create an environment in which people are increasingly unable to differentiate between real and false information, diminishing their ability to make an informed choice and participate in political life in a meaningful way.
Amid these challenges, development practitioners should support information integrity along two main avenues: (a) by helping to establish new rules for and enforce oversight of political communications online; and (b) by complementing these rules with collaborative measures aimed at enhancing ethical behaviour and societal trust. Actions should target not only electoral management bodies (EMBs) but also oversight agencies, encouraging meaningful public engagement more broadly through multistakeholder discussions to agree upon new standards in a whole-of-society approach.
It is vital to involve diverse groups of actors in discussions to agree upon new standards and, in the long term, support solutions that address the root drivers of information pollution, working to ensure the integrity of the information space. These discussions should focus on sharing knowledge and developing legislation; increasing the capacity of oversight institutions is also crucial. It is important that these efforts look beyond the online space—traditional media are still major news sources in many countries—focusing not only on electoral periods but also on the information space between elections. This constructive approach can be especially helpful in the context of international discussions and ongoing efforts to establish agreed-upon standards for the ethical use of AI—and tools for detecting AI-generated content—by creating opportunities for further action. To support greater societal resilience and knowledge, developing electoral integrity, independent and well-funded media, and engaged citizens can help create a more receptive environment by the time these mechanisms are ready to be implemented.
Specific recommendations stemming from an analysis for development practitioners working for the European Union or its member states can be found below. One of the priorities across these recommendations is the inclusion of women, youth and other vulnerable groups in efforts of any kind. Monitoring and evaluation should obtain disaggregated data as much as possible, targeting solutions aimed at increasing the safety and inclusion of these groups. It is also important to make sure that these solutions are accessible across groups, including by encouraging the development of tools in local languages. These factors are particularly key when it comes to generative AI, as these systems are often trained on biased data that can reinforce inequalities (Tuohy-Gaydos 2024; Juneja 2024).
Recommendations for development practitioners
- Support discussions aimed at developing locally owned legislation and regulation for the use of AI-powered tools, both during and between electoral periods.
- Assist state institutions in increasing their capacity to counter digital threats, especially by driving multistakeholder discussions rooted in continual and meaningful collaboration between political representatives, government agencies, AI and social media companies, civil society and academia.
- Complement regulations in the digital sphere with efforts to target the societal factors that contribute to pollution of the information space, especially by addressing inequality and enhancing social trust, media integrity and digital literacy.
Political messaging and public discussion in the digital age
The Internet and social media have enabled the rapid spread of mass amounts of information online, the production of which has been rapidly scaled up thanks to new automated tools, especially those powered by artificial intelligence. These tools have created new opportunities for political parties and civic organizations to engage with the public; however, these same mechanisms are increasingly employed by bad-faith actors to manipulate the perceptions and decisions of ordinary citizens. The online information space in which much public discussion takes place is increasingly polluted with misinformation and disinformation, with acute consequences for electoral integrity and trust in society and democracy, in a context of increasing authoritarianism globally.
Before delving into the subject, it is worth defining the key terms central to this report: generative AI systems are machine-based systems that infer on the basis of the input they receive, for explicit or implicit objectives, how to generate outputs such as content, predictions or recommendations. The capabilities of these systems have the potential to amplify the spread of disinformation, which refers to false information intended to mislead individuals or the public, as well as to create new challenges for cybersecurity, those measures taken to safeguard data against unauthorized access and misuse.
This work addresses these three topics—generative AI, disinformation and cybersecurity—together. To counter the risks to information ecosystems in a holistic manner, development actors must take a comprehensive approach that considers the complex interplay between these factors.
Disinformation refined through the use of AI has the potential to be more believable and poignant, to be shared more quickly and widely, and to be increasingly capable of evading detection (Hanafin, Idrizi and Fischer 2024). The increasingly polluted and polarized information space that is frequently associated with elections contributes to wider risks for democracy amid backsliding and the fragmentation of society. As citizens become more confused by what they see online, trust in the media and society more broadly is increasingly damaged, which has the potential to fuel further political violence and harassment, especially of journalists and women (Stockwell et al. 2024). This combination of confusion, declining trust in the media, and increasing violence and harassment also has the potential to increase support for extremism and military rule, a particular danger in certain parts of the world where there have been military coups in recent years. Democracy supporters are increasingly looking for ways to protect against pollution of the information space in the face of the rapid speed and growing scale of AI-fuelled disinformation.
While the conjunction of AI and democracy has been widely studied in various contexts, particularly with regard to disinformation and its effects on citizens, this study aims to make two contributions to our understanding of these issues: first, to shed a light on perspectives on these issues from the Global South, as voices from the Global North have dominated discussions as digital connectivity and user numbers continue to grow (Grohmann and Ong 2024); second, to connect these perspectives to efforts by the EU and its member states to support democracy in partner countries, as the EU’s demonstrated ability as a global standard setter puts it in an ideal position to assist other countries in developing standards.
Methodology
Given the rapid development of the technologies described above and their associated opportunities and risks, this report asks the following question: how can development programming safeguard against the potential risks that come with the use of AI in electoral processes and political participation while enabling space for innovation? It investigates subquestions regarding the key risks and opportunities associated with AI in developing contexts, the mechanisms used to amplify information pollution, the countermeasures that are being pursued, and the most important contextual differences between developing and developed nations.
This report is aimed in particular at supporting development practitioners in the EU and EU member states in identifying opportunities to support information integrity and the ethical use of AI, while helping develop a holistic vision of democracy support focused on societal resilience and the development of home-grown standards and capacity. Therefore, it does not aim to be prescriptive in its suggestions of specific tools and standards. Rather, the discussion draws on insights from case studies to identify key actors to engage with, principles to pursue, and components of a long-term vision for developing standards and engaging the public in finding solutions.
The analysis aims to clarify key differences between the mechanisms through which disinformation spreads, including through the use of AI-powered tools, in developing contexts, compared with the frequently Western-centric discussion on the topic. The conclusions suggest how development planning could approach democracy and electoral support in these contexts, making recommendations for enhanced cooperation to ensure that countermeasures support free and fair democratic communication unhindered by either information pollution or repression of free speech and the media.
Our understanding of disinformation, generative AI and elections in developing contexts is based on case studies of seven of the EU’s partner countries—Bangladesh, Ghana, Indonesia, Mexico, Mongolia, Pakistan and South Africa. These case studies outline the tactics, techniques and procedures used to spread disinformation in national elections that took place in these countries in 2024, as well as the countermeasures established by legislation, government agency rules and guidance, and civil society initiatives. The full text of the case studies can be found in the annexes.
This chapter provides summaries of the case studies of Bangladesh, Ghana, Indonesia, Mexico, Mongolia, Pakistan and South Africa, the full text of which can be found in the annexes. These countries were chosen as EU development partners that held elections in 2024. Each summary begins with a short overview of the respective country’s elections and digital context, followed by a synopsis of key tactics, techniques and procedures used to spread disinformation; uses of AI; and the countermeasures being pursued by governments and civil society.
1.1. Bangladesh
Before waves of protest triggered the fall of the Awami League Government in the summer of 2024, the parliamentary election held earlier that year in Bangladesh saw the fourth consecutive appointment of Prime Minister Sheikh Hasina of the Awami League (AL). The vote saw an all-time low voter turnout, at less than 40 per cent of eligible voters, as the main opposition party, the Bangladesh Nationalist Party (BNP), called for a boycott in response to the widespread persecution and arrest of party officials. The AL responded with attempts to legitimize the elections, including by promoting their own candidates as ‘independent’ and using coercion to split the BNP, with the election ‘cementing’ the country’s move to a ‘dominant-party’ political system (Macdonald 2024).
Some 44.5 per cent of the population of Bangladesh are online, slightly below the regional average in South Asia of 51.5 per cent (Kemp 2024b). The gender distribution of users is skewed towards men, at 65.8 per cent, and this divide is particularly pronounced with regard to mobile Internet, with low levels of mobile ownership and awareness among women (Jeffrie et al. 2024; Kemp 2024b). In-person campaigning remains the most popular method, with television the most popular news source. Online news sites continue to grow rapidly, however, with the Bangladeshi Government frequently monitoring, censoring or blocking outlets (Freedom House 2023a).
1.1.1. Disinformation tactics, techniques and procedures
The tactics, techniques and procedures used to spread disinformation included coordinated networks of influence, microtargeting and cross-platform amplification. Experts noted an increase in online disinformation targeting the 2024 election with a clear intent to influence public opinion (Islam and Jamil Khan 2023). However, public opinion was manipulated not only through disinformation but also through the extensive use of coordinated networks driven by cross-posting and microtargeting without the disclosure of party affiliation.
The AL was open about investing in a coordinated online network of 8,000 people that spread pro-AL messaging online and debunked ‘disinformation’ (Bangladesh Awami League 2023). However, the use of this network was not transparent, with an analysis of online campaigning indicating that only 19 per cent of the approximately 660 pages, profiles and groups propagating pro-AL content included their affiliation, with the party in the title (Baksh et al. 2023). Both the AL and the BNP were found to have used social bot networks online, including to slander opponents, to flood opposition groups and pages with spam content, and to entice users to interact with fake profiles (Islam and Jamil Khan 2023).
Fake profiles and the use of bots leveraged online news networks to create a sense of perceived legitimacy. The AL invested in a network of so-called experts, releasing bogus information that was then reported on by legitimate news sources (AFP 2023). Both domestic and international news sources therefore inadvertently laundered pro-AL disinformation, as they attributed fake news to ‘experts’ that, upon closer investigation, did not exist. These efforts created confusion and uncertainty online without the widespread use of AI-generated content, as an analysis of campaigning in the 2024 election found that only 1.9 per cent of misleading content was AI-generated (Rahman et al. 2024).
These tactics were more effective in spreading disinformation thanks to the use of microtargeting and cross-platform amplification. The AL’s coordinated network of activists were trained to customize material based on a 2021 nationwide survey that allowed the party to tailor its communications to different regional attitudes and beliefs (Bangladesh Awami League 2023). These actors also boosted the reach of their messaging by simultaneously cross-posting content across pages, especially those imitating legitimate news sources, and used VPNs to imitate foreign IP addresses to avoid strict regulations on press sources in Bangladesh (Islam and Jamil Khan 2023).
1.1.2. Countermeasures
Countermeasures to address disinformation and to regulate AI included legislation on disinformation, content removal, platform shutdowns and digital literacy initiatives. However, critics note that many of these measures resulted in limitations on free speech and freedom of the press while having a limited impact on countering disinformation online, much of which was propagated by pro-government actors (Baksh et al. 2023; Islam and Jamil Khan 2023).
The 2018 Digital Security Act (Bangladesh 2018) prohibited the spread, through digital means, of information in Bangladesh that is offensive, false, threatening or defamatory. The act was updated in 2023 with the adoption of the Cyber Security Act. While the new act partially addressed concerns by reducing the significant penalties imposed by the 2018 act, it has been criticized as a ‘repackaging’ of the previous act. In the digital sphere, these rules are frequently enforced disproportionately relative to the violations, often targeting the opposition. Moreover, perpetrators have reportedly been subjected to inhumane treatment, resulting in a ‘culture of fear’ (Amnesty International 2023; The Business Standard 2023; Rezwan 2023).
The government frequently orders the removal of content online, shutting down platforms in the event that its orders are not followed. Content removal is ordered, purportedly to combat ‘misleading anti-state content’, with social media platforms regularly ordered to remove ‘fake’ or ‘provocative’ information (Freedom House 2023a). These orders are issued through the integrated reporting systems on the platforms themselves, demonstrating a bad-faith application of common global solutions aimed at countering disinformation.
Online news sources are routinely targeted with takedown requests, which are carried out through orders issued to domestic service providers. In 2020, online news outlets were required to register and apply for government approval, with the stated intent of preventing ‘misreporting’; those sources that did not comply were shut down in 2021 (The Daily Star 2020, 2021; Human Rights Watch 2020). This restriction has been criticized as enabling ‘total control and intervention of the government’ over the media (Transparency International Bangladesh 2020).
The government has engaged in digital literacy initiatives and made strides in regulating AI. It also supervised the establishment of a digital literacy centre, which conducts digital literacy surveys and offers online courses for different demographics (Digital Literacy Center n.d.). In 2020, a non-binding national AI strategy was released, along with a road map for future regulatory actions that include the development of an ethical and legal framework for AI, although no further measures have materialized (Information and Communication Technology Division 2020).
1.2. Ghana
Ghana’s democracy has long been regarded as one of the continent’s strongest, with a tradition of peaceful transfers of power. Parliamentary and presidential elections were simultaneously held on 7 December 2024, which saw the return to power of John Dramani Mahama—he had previously served as President from 2012 to 2017—in a presidential race dominated by debates over economic recovery, corruption and governance. The parliamentary vote featured a competitive contest between the incumbent New Patriotic Party (NPP), led by Vice President Mahamudu Bawumia, and the opposition National Democratic Congress (NDC), headed by Mahama (Yerkes and Hogan 2024). Voter turnout stood at 60.88 per cent, a sizeable decline from the 78.89 per cent recorded in the 2020 elections (International IDEA n.d.).
Ghana’s digital landscape has experienced significant growth, with Internet penetration reaching 69.8 per cent and mobile connectivity exceeding 135 per cent, indicating that many Ghanaians own multiple devices (ITA 2024; Kemp 2024c). This expansion in digital access has been accompanied by an increase in social media usage, particularly among youth aged 18 to 34, with 7.4 million users in 2024, primarily on WhatsApp and Facebook (Kemp 2024c; Statista 2024a).
Despite this growth, digital-access disparities persist along gender and geographic lines. Women and rural communities face limited access to smartphones and broadband Internet, restricting their participation in online discourse (Verity 2024; Statista 2024a). The government has implemented policies like the Digital Economy Policy to address these challenges and the Cybersecurity Act of 2020 to regulate misinformation (Parliament of Ghana 2020). However, enforcement has been criticized for selectively targeting journalists and activists under the guise of combating misinformation, fostering a ‘culture of silence’ (Darko 2022).
1.2.1. Disinformation tactics, techniques and procedures
During Ghana’s 2024 elections, political parties leveraged social media platforms to spread disinformation. Coordinated networks disseminated false narratives to manipulate public opinion, often using hashtags like #GhanaStolenElections, which were amplified by foreign-run bot accounts (Haskins 2024). Both major parties engaged in targeted misinformation campaigns, including false claims about candidates’ stances on LGBTQIA+ issues, to polarize and manipulate conservative voters (Osae-Kwapong 2024). AI-generated content was also used to create misleading audio clips and deepfake videos targeting candidates. For example, fabricated recordings falsely accused Mahama of election rigging and Bawumia of making derogatory remarks about Ghanaians (Ghana Fact-Checking Coalition 2025). Additionally, old news articles and images were repurposed as current events to mislead voters, relying on legitimate sources to make the disinformation harder to detect (Jonathan et al. 2024). Social media influencers played a significant role in spreading disinformation, often amplifying partisan narratives for compensation or ideological reasons (Jonathan et al. 2024).
1.2.2. Countermeasures
To combat electoral disinformation, civil society coalitions comprised of individual fact-checking entities like the Ghana Fact-Checking Coalition intensified their monitoring efforts. They verified over 100 instances of manipulated media during the election period, deploying real-time fact-checking desks and OSINT (open-source intelligence) analysts to debunk falsehoods quickly (Modern Ghana 2024). The Cyber Security Authority organized training sessions for journalists on identifying disinformation tactics (Abubakar 2024). However, legal frameworks like the Cybersecurity Act faced criticism for suppressing dissent under the pretext of combating misinformation. Technology interventions also played a crucial role, with platforms like TikTok partnering with local organizations to provide verified election information through in-app tools and media literacy videos (Mgwili-Sibanda 2024). AI-driven detection tools flagged synthetic media for removal before users reported it. Media literacy programmes were also implemented to educate the public on identifying misinformation. Initiatives such as Penplusbytes’ Market Roadshow educated semi-literate populations on identifying misinformation through interactive radio dramas and workshops in local languages (Bodine 2022). The United Nations Educational, Scientific and Cultural Organization (UNESCO) also trained journalists in combating election-related disinformation (Ghana News Agency 2024; United Nations Ghana 2024). These efforts highlight the importance of multistakeholder collaboration in addressing electoral disinformation while emphasizing the need for equitable digital access and transparent enforcement of regulations.
1.3. Indonesia
Indonesia’s presidential election saw then–Minister of Defence Prabowo Subianto win a majority of votes. His win was controversial, as he had been dishonourably discharged from the special forces during the military dictatorship of General Suharto, and his 2024 campaign promoted a new, softer rebrand that was criticized as potentially misleading for voters (Mao 2024).
Some 66.5 per cent of the population are online, with a fairly even proportion of men and women (Kemp 2023a). The most significant digital divide is between young and old, with 77.3 per cent of social media users being between the ages of 13 and 34 (Nurhayati-Wolff 2023). The fact that the electorate in Indonesia is notably young, with 53 per cent of eligible voters being born after 1981, helps explain why social media is heavily relied upon as a source of news (Harmadi 2023).
1.3.1. Disinformation tactics, techniques and procedures
Disinformation in the 2024 presidential election spread primarily through manipulated narratives and framing on social media, the use of coordinated networks of influence and the dissemination of AI-generated content. The use of digital tools was instrumental in Prabowo’s efforts to distance himself from his previous role as a military general during the Suharto dictatorship (Utama 2023).
Pollution of the information space has been a problem in Indonesia for more than a decade, with the term ‘buzzer’ used in political contexts since 2009 to describe those individuals collaborating with different actors to spread information aimed at shifting narratives (Dewantara et al. 2022). As microtargeting has become more sophisticated, and AI-generated content more accessible, this phenomenon has only intensified (Mediana 2023; Rayda 2024). The use of buzzers has grown into an industry of its own, spanning from volunteer networks to advertising companies, which often run ‘black campaigns’ that use disinformation to slander political candidates (Mediana 2023; Fakultas Hukum Universitas Indonesia 2024).
Candidates have capitalized on the increased popularity of video and short-form content, with several candidates live-streaming to speak directly with voters and answer questions, and participating in viral trends to connect with voters (Ratcliffe 2024; Tarigan 2024). These techniques have enabled candidates to court younger voters and cultivate a sense of intimacy and familiarity; however, messaging has focused more on the candidate’s personality than on party policy, raising concerns over less informed voters (Belinda 2023; The Economist 2024).
AI-generated content was used frequently in 2024 to support candidates’ attempts—especially by Prabowo—to depoliticize their image and communicate with voters. AI-generated images were used to great effect by the Prabowo campaign, which generated animated personas that garnered the nickname gemoy (adorable), helping to distance the candidate from his image as a military general (Utama 2023; Duffy 2024). AI-generated deepfakes were also used in online communications, including a video depicting the former dictator Suharto, who died in 2008, endorsing the Golkar party (Firdaus 2024). Deepfaked audio was used to slander the opposition through the dissemination of fake news and false content including fabricated audio of candidates expressing opinions that they did not actually voice in real life (Rayda 2024). AI-powered chatbots were also used to bring parties closer to voters, answering voters’ questions in the voices of the candidates. However, the chatbots were found to be unreliable, providing incorrect and misleading answers (Rayda 2024).
1.3.2. Countermeasures
The countermeasures taken against disinformation and to address the use of AI included a law against disinformation, government monitoring and mitigation of disinformation, cooperation between the government and social media platforms directed at content removal, AI ethics guidelines and digital literacy support.
The Indonesian legal tradition of imprisoning individuals for spreading fake, exaggerated or incomplete information intended to cause public unrest is firmly institutionalized (Djajadiningrat 2023). In 2024, new rules were introduced to address the electronic spread of misinformation that leads to material losses, incites hostility against individuals or groups based on identifying factors, or is knowingly false (Eddymurthy and Sihombing 2024). However, the enforcement of these measures has been criticized for restricting free expression (Bratajaya and Tan 2023).
The Ministry of Communication and Digital Affairs (Kominfo) monitors disinformation on social media around the clock, using both human expertise and technological tools (Inforial 2023). Kominfo communicates about these efforts on its website, bringing attention to disinformation and debunking false or misleading narratives (Kutanto 2023). The Ministry also collaborates with the General Election Supervisory Agency (Bawaslu) to counter disinformation in the context of elections (Inforial 2023).
Kominfo also sends requests to social media platforms for the removal or correction of information, on the basis of a 2020 law requiring social media platforms to self-moderate and remove unlawful content within 24 hours of receipt of a removal request or face fines and temporary or long-term shutdowns (Timmerman 2022; Freedom House 2023b). However, concerns have been raised about the frequent use of criminal sanctions and the arbitrary grounds the Indonesian Government has cited for blocking access to platforms (Freedom House 2023b; Mediana 2023).
Regarding digital literacy and AI guidelines, the government has recognized digital literacy as a core pillar in safeguarding the information environment, and has released guidelines for the ethical use of AI (Sakura and Virgiany 2024). In 2017, the government started the National Movement for Digital Literacy, a collaborative multistakeholder effort to support digital capacity and prevent the spread of disinformation, which Kominfo also highlighted as a key part of its efforts to counter disinformation in the 2024 election (ANTARA 2023a; Siberkreasi 2023). However, observers have criticized the learning materials of the movement, claiming that they encourage people to obey state orders rather than promoting critical thinking (Idris 2022). Guidelines for ethical use of AI were established through a non-binding circular in 2023 and in response to criticisms of the circular’s voluntary nature, Kominfo committed to preparing binding regulations in the future (ANTARA 2023b; Juniarto 2024).
1.4. Mexico
In the largest-ever election in the world’s most populous Spanish-speaking country, over 98 million voters were asked to cast their votes for president on 2 June 2024. Beyond selecting a new head of state, Mexicans also voted for 500 members of the Chamber of Deputies, 128 members of the Senate of the Republic, state governors and over 20,000 public service positions across state and local levels. The Mexican election cycle was marked by a record amount of political violence, both online and offline, with women especially vulnerable to abuse. With Internet penetration of over 78 per cent, Mexico hosts the second-largest audience for social media platforms across the region (Statista 2024b).
1.4.1. Disinformation tactics, techniques and procedures
The Mexican elections were accompanied by a wave of orchestrated information pollution, aimed at spreading and amplifying fake news and destructive narratives. Estimates indicate that more than half of all Internet traffic in Mexico is made up of automated bots (Imperva 2024).
A major challenge in the information environment leading up to the elections was the dissemination of misleading and false narratives by candidates and party officials themselves. Political analysts have noted that disinformation disseminated across party lines often followed a common pattern, attempting to provoke the greatest amount of political polarization possible (Echeverría 2023; Martínez Chacón 2023).
When party officials use their platforms on social media to disseminate false or deceptive narratives, the information landscape risks becoming increasingly confusing for audiences. These narratives can be picked up, amplified and kept in circulation by actors seeking to influence political perceptions (Breuer 2024). The reiteration of disinformation amplifies its reach, bloats online information spaces and lends a perceived legitimacy to inauthentic information.
Another concern in Mexico is online gender-based violence against female political candidates. In a race that saw substantial female participation and effective representation in the political arena, most notably by the two women presidential candidates leading the polls, these forms of abuse and discrimination can escalate into physical violence, including, in extreme cases, femicide (Ríos Tobar 2024).
Interestingly, most inauthentic amplification campaigns originate on X and are shared on other platforms only later. On X, actors seeking to influence public opinion use consolidated strategies to boost specific messaging with the intent of reaching as large an audience as possible. These strategies include studying the behaviours of Mexican social media users through analytics to maximize engagement, such as determining the time of day a post is more likely to garner attention. These campaigns are predominately coordinated by two different types of groups. The first type consists of independent actors who coordinate joint action in private messaging groups on WhatsApp or Telegram, employing their own networks (Aguirre 2021; Cerdeira 2024). The other type consists of actors who outsource the implementation of disinformation campaigns to companies that are paid to provide such services (Naciones Unidas México 2021). Once campaigns go viral, they are shared on other platforms, partly to circumvent regulations.
Generative AI has been deployed particularly to promote disinformation, gender-based violence online, fraudulent investment schemes and conspiracy theories. Researchers recorded several instances of deepfakes, such as altered audio clips or fabricated intimate imagery created without consent (Gamboa 2023; Méndez 2024; Del Pozo and Arroyo 2025).
1.4.2. Countermeasures
The Mexican National Electoral Institute (INE) launched a campaign, encompassing five main pillars, to combat the spread of electoral disinformation: (a) verifying information; (b) establishing bilateral cooperation between stakeholders; (c) cooperating with the media; (d) building media literacy; and (e) providing transparency and verified electoral information (INE 2024a). The campaign’s online platform serves as a fact-checking hub that utilizes multidisciplinary expertise from bilateral partnerships to debunk false or misleading information circulating online. Furthermore, the INE engaged in training missions throughout the country, collected information about disinformation narratives and generated responses to debunk them.
Moreover, as part of the campaign, the INE launched a virtual chatbot assistant intended to support the dissemination of factual and trustworthy information on the 2024 elections. In collaboration with WhatsApp, the chatbot assistant, called Inés, was able to receive information in text, image or video format that was then transmitted to fact-checking teams who provided a categorized response on the authenticity of the information. Additionally, users could ask questions and receive verified information about the elections from independent sources (INE 2024b).
Beyond their information integrity campaigns, the INE collaborated with social media platforms. In February 2024 the INE announced that it was collaborating with the Mexican civil society organization (CSO) Movilizatorio and the social media company Meta. The partnership involved the launch of a new edition of the I Am Digital (Soy Digital) educational programme, aimed at tackling electoral disinformation. The programme improves media literacy through workshops for students, civil society, the public sector and academics on how to be cognizant of disinformation (Vergara 2024). In partnership with TikTok, the INE published an election guide customized to the Mexican context, providing official data and information surrounding the elections (Valverde 2024). CSOs complemented these actions, including through independent fact-checking and media literacy initiatives (Duke Reporters’ Lab n.d.; Verificado n.d.).
The International Institute for Democracy and Electoral Assistance (International IDEA) dispatched a technical assistance mission to the INE that concluded that the INE had taken steps to respond to the growing amount of gender-based violence online. These efforts included creating a technical unit for gender equality, producing educational materials and establishing a national registry of sanctioned individuals for political gender-based violence against women (International IDEA 2024a).
1.5. Mongolia
Mongolia’s 2024 parliamentary election led to the adoption of constitutional amendments that increased parliamentary representation for opposition and smaller parties, as well as for women, youth and professionals from diverse backgrounds (AP 2024; Lkhaajav 2024; Rikkilä Tamang 2024). Despite public distrust and systemic challenges, the country remains a multiparty democracy with high levels of political engagement (Dierkes 2024).
With 72 per cent of the population actively using social media, Internet penetration in Mongolia is high. However, the governmental Communications Regulatory Commission controls much of the media landscape and has faced criticism for censorship, lack of transparency and politically influenced content regulation (Naranjargal 2018; United States Department of State n.d.). The commission has been accused of employing broad content restrictions and overseeing broadcasting licensing without public consultation; these allegations have raised issues of fairness and political influence in media access (Naranjargal 2018; United States Department of State n.d.).
1.5.1. Disinformation tactics, techniques and procedures
Political influence networks shaped media discourse during the 2024 electoral cycle. Major Mongolian broadcasters disproportionately covered the ruling Mongolian People’s Party and the opposition Democratic Party, with a notable bias towards the former (OSCE/ODIHR 2024). Television channels and online media outlets were involved in coordinated efforts to portray the ruling party in a favourable light, blurring the lines between independent journalism and party-driven narratives (OSCE/ODIHR 2024).
Given Mongolia’s geostrategic position, both its neighbours, China and Russia, played a significant role in shaping the information environment leading up to the election. Chinese-backed outlets sought to promote economic stability and regional cooperation, aligning with the objectives of the Belt and Road Initiative, while Russia reinforced messaging against Western alliances (Tsedendamba 2024).
To amplify the foreign influence from China and Russia, actors made extensive use of cross-platform dissemination. Fabricated statements made by Russian officials were quickly shared by mainstream news sources in Mongolia, ensuring that false narratives reached broad audiences (Tsedendamba 2024). The method of republishing articles across various platforms can help lend legitimacy to misleading content by creating an illusion of independent corroboration (Wilson and Starbird 2020). This cross-platform strategy can also make fact-checking efforts more difficult, as disinformation can be framed as coming from multiple sources, reinforcing its perceived credibility among the public.
A key issue in Mongolia’s information environment is the distortion of translated content. Misinformation often undergoes semantic shifts that can alter its meaning and possibly enhance its persuasive impact. While fact-checkers work to counter these effects, ensuring that the country’s nomadic population receives verified information remains a challenge (Mantas 2021). Addressing these issues requires strategies that account for both linguistic and cultural complexities.
1.5.2. Countermeasures
To combat disinformation, Mongolia has introduced legislative and regulatory measures, though these have sparked concerns about potential restrictions on dissent (Boldkhuyag et al. 2021; Volodya 2022; RSF n.d.). Efforts to improve digital literacy focus on youth, women and rural populations. Programmes are in place—including one called Digital Nation—to expand digital access and security, but gaps persist, especially among the elderly. CSOs and media initiatives play a crucial role in fact-checking and raising public awareness (NSOM, UNICEF and UNFPA 2024). The Mongolian Fact-Checking Center, for example, has launched tools like the Asuu chatbot to help verify questionable content (Meedan 2024). International organizations, including the DW Akademie and UNICEF Mongolia, have also supported local media literacy efforts, particularly during elections (Tsolmonbaatar 2019; Battulga 2023). CSOs such as the Nest Center for Journalism Innovation and Development and the Mongolian Fact-Checking Center monitor public discourse, dispel misinformation and advocate for transparency (Batdorj 2020; Meedan 2024). Their efforts, particularly in addressing health-related disinformation during the Covid-19 pandemic, underscore the importance of public trust in factual information and independent institutions in strengthening Mongolia’s democratic resilience (Battulga 2023).
1.6. Pakistan
Candidates aligned with former Prime Minister Imran Khan won the majority of votes in the 2024 general election in Pakistan, though Khan himself is barred from public office for 10 years following a 2022 no-confidence vote and indictment. However, Khan’s party, Pakistan Tehreek-e-Insaf (PTI), did not win a simple majority, and the other parties that won seats in the election were unwilling to work with the PTI in parliament, resulting in the formation of a coalition government between the Pakistan Muslim League Nawaz and the Pakistan People’s Party (Seddon and Javed 2024). The divisive result of the election and challenges with forming a coalition government have contributed to further political unrest and high levels of distrust in the electoral system, fuelled also by the PTI’s allegations of a rigged election (Gul 2024; Malik 2024).
Some 45.7 per cent of the population in Pakistan are Internet users, placing the country slightly below the South Asian regional average of 51.5 per cent (Kemp 2023b). The country has a significant gender gap in Internet users, with roughly twice as many men as women on social media, and further gaps due to the lack of affordability and digital infrastructure in rural areas (Kemp 2023b). Disinformation is nonetheless a significant issue, with high levels of disbelief in genuine political information and a majority of the public expressing uncertainty in their ability to consistently identify misinformation (Mir and Siddiqui 2022; Naeem and Rehmat 2022). Elites in the country routinely employ manipulative tactics online, with disinformation narratives targeting the credibility of the EMB especially (Shaheen 2022).
The case in question demonstrates how generative AI can supercharge existing influence campaigns, helping them bypass both online and offline countermeasures. For example, although offline measures were taken to prevent Khan from taking office, he sidestepped them with virtual rallies. Online, AI-powered tools enabled cross-platform amplification and the misuse of fact-checking initiatives. Even where AI-generated content was labelled as such, it was still useful in creating momentum for rallies, despite the low level of Urdu integration, reducing the credibility of outputs.
1.6.1. Disinformation tactics, techniques and procedures
The 2024 election featured the extensive use of disinformation and other influence tactics aimed at slandering political rivals, misdirecting voters and undermining democratic procedures. The main influence tactics identified, both ahead of and during the 2024 election, included the significant use of coordinated networks, misleading ‘fact-checking’, cross-platform amplification, microtargeting and AI-generated content. The use of social media and AI tools was key in helping the PTI to become the nation’s largest party and enabled the use of alternate forms of communication when actors faced censorship.
Coordinated networks of influence were frequently used during the election, with the PTI being the party that employed these tactics most frequently. It held virtual rallies online, engaging supporters especially through the use of AI-generated speeches that encouraged them to spread pro-party messaging further, though the party’s use of AI was not always labelled (Khan 2023; Bahl 2023). The spread of pro-party messaging was further amplified through a sizeable volunteer supporter network, especially by coordinating the use of hashtags and amplifying these messages across platforms, often reposting the same content many times or taking advantage of the priority algorithms give to sensationalist or reactionary content (Jahangir 2020; Hendrix and Jahangir 2024; Khan 2024). These tactics are also used by a professionalized force of ‘cyber troops’ that further fuel cross-platform amplification through hashtag campaigns, with coordination in recent years increasing for attacks on journalists and female politicians, including through the use of generative AI and photo manipulation (Bradshaw et al. 2021; Dawn 2023; A. Khalid 2023).
These efforts were further supported through the use of microtargeting, with messaging circulated on WhatsApp of particular concern, given the app’s encrypted nature, with evidence indicating that misinformation on WhatsApp can stay in circulation longer than on other platforms (Naeem and Rehmat 2022). AI-generated content was also used to convince voters that certain candidates had stepped down, with a targeted campaign against the PTI one day before the election (Ali 2024; Marchant de Abreu 2024). A factor that limited the use of AI was the lack of Urdu integration into large language models, making them less credible (Jahangir 2024).
1.6.2. Countermeasures
Countermeasures taken to address disinformation included several CSO initiatives as well as government-led efforts, including a national AI policy, a regulation targeting disinformation, the monitoring and removal of disinformation, and Internet shutdowns.
Civil society and journalists have made important strides against disinformation in Pakistan. CSOs have led educational programmes to improve the digital literacy of journalists and the public, with DigiMAP and the Freedom Network outlining targeted recommendations to help organizations in Pakistan effectively respond to disinformation (Naeem and Rehmat 2022). However, these efforts have faced technical and financial limitations, as well as challenges with translating English-language materials into local languages (Freedom Network 2022). CSOs have also called on the authorities to respond to disinformation by reviewing Pakistan’s digital resilience and cooperating in a multistakeholder process (A. Khalid 2023).
However, moves taken by the government to address disinformation have faced substantial criticism. In 2022, the government criminalized online ‘defamation’ of the authorities, with several other measures limiting freedom of the press and freedom of speech online under the banner of fighting disinformation (Human Rights Watch 2022; RSF 2023). The Ministry of Information and Broadcasting hosts a website to bring attention to and address disinformation; however, critics have highlighted the risk that this website will be used to target opponents and other dissenting voices (Shaheen 2022). Finally, new rules adopted in 2021 gave the government the authority to block or remove from social media information deemed indecent, blasphemous or false, or that could disrupt public order, with platforms risking nationwide blocks without compliance within 24–48 hours of receiving a removal notice (Daudpota 2021). Platform blocks and Internet shutdowns have frequently been used to silence critics and have been condemned for their impact on the public’s ability to access information during election campaigns (Access Now 2024; Hussain 2024).
1.7. South Africa
The 2024 general election in South Africa resulted in the seventh consecutive victory for the African National Congress in the post-apartheid era, although the party lost its majority for the first time and therefore entered into a coalition government with the opposition Democratic Alliance. President Cyril Ramaphosa was re-elected for a second term. Voter turnout hit a 30-year low, at 58.64 per cent, with increased political apathy and a decline in voter confidence in the electoral process, a situation particularly pronounced among youth (Gumede 2024; Cleveland and Alexander 2024).
Some 74.7 per cent of the population is online, in line with the average for Southern Africa (Kemp 2024a, 2024d). There are no major differences in Internet use between genders, while the majority of people online are between the ages of 18 and 34 (Cowling 2024). People in rural areas are less likely than those in urban areas to use the Internet, and are more likely to do so on mobile devices (Wyrzykowski 2023). While television is the most popular news source, online news sources continue to grow in popularity, especially on social media (Newman et al. 2024). Citizens are highly concerned about misinformation, with 82 per cent of respondents in a poll by the Reuters Institute Digital News Report indicating that they are concerned (Newman et al. 2024).
1.7.1. Disinformation tactics, techniques and procedures
Disinformation was spread during the 2024 election through the use of AI-generated content, coordinated networks of influence, microtargeted content, cross-platform amplification, and the manipulation of political narratives and framing using social media. AI-generated content was circulated widely during the election, causing uncertainty and confusion—despite being unsophisticated—demonstrating that even low-quality deepfakes can impact perceptions (van Damme 2024).
Coordinated networks of influence were identified across a variety of social media apps, with influencers paid and bot accounts deployed to share and amplify disinformation, especially on WhatsApp, one of the most popular apps for sharing political information in South Africa (Allen and le Roux 2024; Dionne and Liu 2024).
It is concerning that political parties were identified as one of the biggest sources of information manipulation, with common narratives including attempts to undermine the credibility of the Electoral Commission of South Africa (IEC) and xenophobic rhetoric about migrants (European Union Election Expert Mission 2024). Moreover, a network of pro-Russian bot accounts was uncovered that propagated fake news especially on X (ADDO 2024; Sguazzin 2024). Narratives intended to undermine the legitimacy of the electoral process, even when quickly and easily debunked, managed to have real-world impact, most notably with a legal challenge filed by the uMkhonto weSizwe (MK) party based on false claims of ‘missing’ votes (Allen and le Roux 2024; European Union Election Expert Mission 2024).
Both foreign and domestic actors relied especially on the spread of ‘rage bait’ material, taking authentic materials out of context to build engagement and amplify the information’s spread (Allen and le Roux 2024). The spread of certain messages was also amplified through the use of advanced data analytics and microtargeting, which made it possible to tailor messages to citizens based on their location, as well as their age, gender and other factors (Kreuser and Gitonga Theuri 2024).
Sentiment analysis shows that 88.6 per cent of posts on X were negative in tone and outlook, as accounts frequently coupled hashtags with incendiary narratives that drove user engagement (ADDO 2024). While only 18 per cent of political ads were found to have used microtargeting, they received disproportionately greater visibility than non-targeted ads, demonstrating the significance of its use even when employed in a minority of political ads (Kreuser and Gitonga Theuri 2024). Finally, cross-platform amplification played a key role in the spread of misinformation, as the unique features and audiences of different platform algorithms helped amplify the reach of misleading content (ADDO 2024; Allen and le Roux 2024).
1.7.2. Countermeasures
Countermeasures being taken against disinformation and in support of the ethical use of AI include voluntary initiatives to secure commitments and monitor disinformation, AI-powered tools to combat disinformation, legislative and regulatory measures, and digital literacy initiatives.
The major gaps left by countermeasures in the election demonstrate the need for more robust policies, greater transparency on the part of digital platforms and improved media literacy education. However, the election also demonstrated innovative mechanisms that can help to safeguard elections while respecting freedom of expression and upholding public participation through collective solutions driven by partnerships between government, the private sector and civil society actors. The Framework of Cooperation signed by key actors and the Real411 complaints platform (see Box 3.1), created by the EMB and Media Monitoring Africa, helped enhance transparency and oversight of electoral campaigning (IEC 2023).
Regulatory and legislative efforts to tackle disinformation include the Cybercrimes Act of 2020, which provides that anyone found to have knowingly disseminated false or defamatory content may be held civilly liable for defamation (Grange 2024). However, this measure has had limited impact in combating the misinformation that is rife on social media platforms, which are not subject to a clear regulatory framework during election periods. Moreover, the South African authorities have limited capacity to enforce rules in respect of social media platforms operating outside the country (European Union Election Expert Mission 2024). In addition, while the information regulator issued a guidance note for the use of personal data by political parties for the 2019 election, no such guidance was released ahead of the 2024 election, contributing to legal and regulatory gaps concerning political parties’ online communications (European Union Election Expert Mission 2024). South Africa’s Film and Publication Board plans to hold social media companies liable for information pollution online and to collaborate with the South African Artificial Intelligence Association to develop regulations on the use of AI on social media apps (Digital Watch 2024; Illidge 2024).
Turning to digital literacy efforts, media literacy is currently part of the basic curriculum in South African schools; however, a report by Africa Check indicated that there is insufficient focus on fact-checking and verifying media (Wasserman, Madrid-Morales and Rabin 2022). A more comprehensive media literacy module, however, has been developed by the Western Cape Education Department, which notably did so in partnership with Google, demonstrating the capacity to develop media literacy in collaboration with online platforms (Wasserman, Madrid-Morales and Rabin 2022). South Africa has also developed DSFOne, a national digital skills competency framework, and the Digital and Future Skills Strategy, which aligns with the EU Digital Competence Framework for Citizens, which defines digital skills across basic, intermediate and advanced levels and aims to address low digital inclusion, enhance economic opportunities and social participation, and provide targeted support for youth, women and persons with disabilities (Domingo et al. 2024).
South African CSOs have also made strides in supporting digital literacy education across the country. Media Monitoring Africa, for example, has created its own digital literacy initiatives to engage the public on ways to better spot fake news, including the Web Rangers platform to help young people and parents gain critical knowledge about online safety (Media Monitoring Africa 2024).
One of the most important findings across the case studies is that the spread of disinformation and other fake and misleading content online is not simply about creating damaging narratives but about polluting the very flow of information. This pollution of the information flow can be especially effective in shifting public opinions and attitudes and creating distrust, both in other citizens and in democratic and electoral institutions. It can also decrease individuals’ motivation to vote and participate politically at a time when voter turnout is already at record lows in the countries observed (Hanafin, Idrizi and Fischer 2024).
Disinformation is increasingly spread by for-hire networks, creating a professionalized space to shift public narratives and intensifying the challenges for platforms in detecting and countering inauthentic content and amplification (Grohmann and Ong 2024). For example, paid coordinated networks of influence were identified in the case studies in Bangladesh, Indonesia, Mexico, Pakistan and South Africa, demonstrating the increased salience of these networks in many developing contexts.
Across the case studies, there was evidence of coordinated influence campaigns—both paid and volunteer—often connected with domestic actors and political party campaign apparatuses and frequently making extensive use of fake and bot accounts. Microtargeting was less prevalent than expected in spreading party communications; leveraging the unique features of social media platforms was more common.
It is concerning that many of the unethical uses of generative AI identified in the case studies were carried out by domestic actors. The technology was used especially to smear rival political opponents and sow confusion and distrust among voters (Stockwell et al. 2024). It is therefore crucial that development practitioners recognize the role of elites in fuelling information pollution. An overview of the tactics, techniques, procedures and mechanisms involved are listed in Table 2.1.
Tactics, techniques and procedures | Country examples | Mechanisms |
---|---|---|
Coordinated influence networks | Bangladesh | Professionalized pro-party network, network of bot accounts, network of fake experts. |
Ghana | Networks promoting narratives of electoral fraud and misconduct by opposition candidates, weaponization of ethnicity and LGBTQIA+ issues to discredit candidates. | |
Indonesia | Significant use of professionalized and personal pro-party networks of fake accounts and bots, obfuscation of the original source of content to evade oversight. | |
Mexico | Significant use of coordinated networks of bot accounts across party lines (by both foreign and domestic actors), weaponization of organized-crime issues to discredit candidates. | |
Mongolia | Coordinated narratives across both traditional and online media, significant coordinated efforts to shape political narratives by China and Russia. | |
Pakistan | Significant use of volunteer and professionalized pro-party networks, coordinated networks of bot accounts, coordinated use of misleading ‘fact-checking’ to silence undesirable narratives. | |
South Africa | Coordinated networks of paid users and influencers, coordinated networks of bot accounts, weaponization of narratives to undermine electoral integrity and exploit issues with migrants, coordinated network of pro-Russian accounts. | |
Microtargeting | Bangladesh | Tailoring messaging according to regional attitudes and beliefs, training professionalized networks in reaching certain users. |
Indonesia | Inundating identified audiences with misleading content, including AI-generated content. | |
Mongolia | Uneven use of targeting across parties, complementing digital campaigns with traditional methods; leveraging by smaller parties to reach voters during a short election campaign; campaign messaging predominantly targeting men, while messaging targeting women was largely absent. | |
Pakistan | Widespread use of targeted advertising especially on Meta-owned platforms, targeting voters based on bought-and-sold contact and other personal information. | |
South Africa | Spending especially on targeted ads on Meta-owned platforms. | |
Leveraging social media platforms, including cross-platform amplification and influencers | Bangladesh | Professionalized coordinated hubs to amplify information across different pages, use of foreign IP addresses to evade oversight. |
Ghana | Out-of-context recycling of news items, significant spread of misinformation by influencers, coordinated hashtag campaigns. | |
Indonesia | Use of live-streaming and participating in viral trends to forge casual connections with young voters and depoliticize candidate images. | |
Mexico | Coordinated hashtag campaigns and artificial amplification of certain narratives through engagement with posts, amplification of messaging through private groups and chats on Meta-owned platforms and Telegram, recycling of false narratives originating on X by reposting them on other pages and platforms. | |
Mongolia | Recycling of pro-Kremlin narratives originating from sources connected with Russia. | |
Pakistan | Live-streaming virtual rallies, using coordinated hashtag campaigns, amplifying narratives across platforms to circumvent platform blocking and shutdowns, purchasing engagement to artificially increase the spread of content, leveraging private messaging services such as WhatsApp to evade fact-checking and oversight. | |
South Africa | Amplification of pro-party messaging over WhatsApp, use of paid influencers, coordinated amplification of messaging across pages and platforms and by artificially enhancing engagement, coordinated hashtag campaigns. | |
AI-generated content | Bangladesh | Foreign outsourcing of generated content (regularly manipulating images or taking things out of context is still more prevalent than deceptive synthetic depictions of political figures). |
Ghana | Synthetic audio content, significant presence of deepfakes on YouTube. | |
Indonesia | Extensive use of AI tools, use of synthetic media in campaigns to depoliticize candidate images, use of deepfake videos to depict past figures endorsing candidates or to slander political rivals, use of AI chatbots to communicate information with voters, though with unreliable accuracy. | |
Mexico | Use of synthetic audio content, use of deepfake video content to portray candidates as soft on organized crime, use of an AI-generated spokesperson, significant spread of AI-generated content to target women, including two cases of non-consensual synthetic pornography targeting women candidates. | |
Pakistan | Deepfake video content to enable a candidate to circumvent repression and give speeches virtually from prison via a synthetic version of himself, deepfake video content to falsely claim candidates were stepping down or calling for boycotts of elections, AI-generated content targeting women and transgender politicians. | |
South Africa | Significant spread of synthetic media, deepfake videos featuring false endorsements of politicians. |
Genuine democratic debate can be disrupted as citizens become increasingly unable to differentiate between real and fake information, fracturing the relationships that democracy relies on and weakening the desire to cooperate towards common goals Breuer 2024(Breuer 2024). Confusion over what is true and what is false can also sever the chain of accountability: the ‘liar’s dividend’ is increasingly of concern, whereby actors are able to dismiss real content and criticism as ‘fake news’, enabling them to evade accountability and potentially redirect criticism at the so-called perpetrators (Stockwell et al. 2024). These risks are made more acute because of cognitive biases, as people are more likely to ignore information that contradicts their beliefs and accept even ambiguous information where it confirms their beliefs (Schwieter and Gandhi 2024). A growing concern is therefore not only that citizens are exposed to disinformation but also that they are increasingly rejecting credible sources of information. As a result, the online ecosystem risks becoming an engine for division rather than a public arena where people can come together to discuss policy and a common future for society, ultimately contributing to instability and hindering democratic development.
Information ecosystems also face acute risks, given wide-ranging pressures on traditional media, combined with their continued relevance as the main source of news for many people in the developing world. The Afrobarometer (Conroy-Krutz, Amakoh and Amewunou 2024) notes that radio remains the main method of news consumption across Africa, with two-thirds of the continent’s population using it every day or a few times a week, while social media is used by slightly less than half the population. Surveys in South East Asia indicate that slightly more than half of citizens across the region access their news online, followed by television and print news, while in Latin America, two out of every five individuals rely on social media for news (Black Box 2024; LatAm Intersect 2025).
At the same time, traditional media revenues continue to decrease, and as local and regional outlets close, information vacuums are created that are quickly filled by actors peddling unconfirmed and false information (Hanafin, Idrizi and Fischer 2024). As the information ecosystem becomes increasingly overrun with false and synthetic information, it can create a vicious cycle of self-amplification, as falsehoods are repeated and spread to other sources, obscuring the original creator (Shah and Majumdar 2024).
The danger of AI-generated knowledge sources continues to grow, as highly realistic fake news stories can proliferate on convincing-looking websites as genuine sources from so-called experts that do not actually exist (Stockwell et al. 2024). This situation was observed in Bangladesh, for example, where networks of so-called experts spread false information that was then reported on by legitimate news outlets (AFP 2023), obscuring the original source and making these narratives appear more legitimate.
This erosion of trust caused by the spread of AI-generated disinformation and the blurring of legitimate and fake sources can also play a role in decreasing a population’s motivation to engage in elections. In most of the case studies, a pronounced decline in voter turnout was observed: as people lose trust in the electoral system and increasingly feel that they do not have the knowledge or tools to make an informed choice, they may pull back from voting and participating in democratic life altogether (Shah and Majumdar 2024). Conversely, where people feel that they are supported by a transparent electoral process based on responsible and complete campaign information, they are often more ready to vote and participate politically. Accordingly, it is crucial that a sustained investment is made in monitoring and analysis to inform government strategies on disinformation and AI and to avoid the risk of cracking down too strongly on content online (Stockwell et al. 2024). These legal measures should also be complemented by voluntary collaborative initiatives, especially channelling multistakeholder input towards both legal and soft-law standards.
The acute dangers posed for vulnerable groups, in particular women, by the misuse of new online communication tools also demand attention. Disinformation is increasingly aimed at women political leaders and candidates, as well as election officials and journalists, creating an atmosphere of intimidation and chilling effects on free speech (Hanafin, Idrizi and Fischer 2024). There is mounting evidence that synthetic content online disproportionately targets women, with other documented risks regarding targeted harassment of women online enabled by new digital and AI tools (Ajder et al. 2019; Brooks et al. 2021; Dunn 2021; Li, Shultz and Jankowicz 2024; Shukla 2024). In Pakistan, for example, a false narrative circulated targeting women as ‘foreign agents’ threatening traditional values, leading to their social and legal harassment (Shah and Majumdar 2024). Moreover, there is a concern that generative AI could be used to target certain groups, as the speed, scale and sophistication made possible with AI tools could translate into the disenfranchisement of entire groups (Ferrara 2024). Given the strides made in many countries to raise the status and welfare of women and other vulnerable groups, it is important that this progress is not lost, as information pollution online can emphasize traditional views of women and discourage them from running for office.
While AI-powered tools were employed at times to connect with voters, their positive impact was undermined where they were used to manipulate voter perceptions. In Indonesia, for example, the Prabowo campaign was especially successful in connecting with youth through AI-generated images of the candidate; however, these tactics were criticized as attempts to distract voters from policy-related issues and depoliticize Prabowo’s reputation (Utama 2023; Duffy 2024). Prabowo’s party also used chatbots to communicate with voters, although they were criticized for spreading inaccurate information (Rayda 2024).
The most notable use of AI-powered tools to evade government repression came from Pakistan, where the Khan campaign utilized a virtual persona to enable the candidate to campaign from prison, holding virtual rallies to energize support (Khan 2023). However, Khan’s party did not always clearly label their uses of AI (Bahl 2023).
Finally, AI-powered tools also created opportunities for CSOs to increase the transparency of election campaigning, with the African Digital Democracy Observatory in South Africa using sentiment analysis to analyse online discussions in the lead-up to the elections (ADDO 2024).
For an overview of the various opportunities and risks associated with the use of AI-powered tools in political communications see Table 2.2.
Opportunities | Risks |
---|---|
Lowered barriers and costs of communicating with mass audiences online: Easier for new actors to enter the political arena Ability of vulnerable or marginalized groups to advocate for themselves Higher capacity to campaign in repressive contexts: it is not possible to arrest an AI-powered tool | Reduced ability of citizens to understand what they are seeing online: More convincing false or synthetic content means blurring the line between what is real and fake, true and false Messaging that often seeks to delegitimize electoral and other democratic institutions or spreads harmful narratives about vulnerable groups Fractured public debate, undermined public trust, polarization Increased challenges for oversight bodies and civil society to monitor political communications, enforce rules and measure the real impact of AI Despite lowered barriers to entry, the actors with the most resources can skew the playing field and sway public opinion |
Tailored communications according to demographics: Cheap, fast and accurate translation and accessibility tools Enhanced ability to hear the desires and needs of constituents Use of clearly labelled chatbots to engage voters in two-way conversations | Data privacy risks: Lack of voter knowledge or consent regarding the use of their data Unique dangers that arise when sensitive data is used to target voters, regarding, for example, their race or sexuality Lack of access to and transparency of data used to target voters and train AI systems, creating the potential to reinforce existing biases |
Optimized use of resources: Accelerates the process of brainstorming, research and data analysis as well as the generation of drafts to inform policy positions Representatives and political parties have more time to engage with and be accessible to the public EMBs and other oversight agencies can make more efficient use of often already-limited capacity | Cybersecurity risks: AI systems continuously learn and adapt, helping bad-faith actors to evade detection Security, detection and content moderation are continuous undertakings that require constant updates For EMBs, large amounts of voter data and voting software are vulnerable to being compromised or manipulated Increased vulnerability to technical failures or supply chain disruptions |
Observed countermeasures included a number of examples of legislation passed—in Bangladesh, Ghana, Indonesia, Mongolia and Pakistan—for the stated purpose of countering misinformation or disinformation that ultimately were criticized for placing limitations on free speech, thus demonstrating the risks that come with regulating content. There were also cases (e.g. in Bangladesh, Indonesia and Pakistan) where governments used such laws to order the removal of ‘false’ information, often blocking or throttling websites—deliberately slowing access speeds—to force compliance.
The Indonesian case in particular demonstrated the risks of combining rigid standards for the removal of information deemed to be false with the communications authority’s high capacity for 24-hour monitoring of disinformation in cooperation with the EMB. Observers have criticized restrictions on free speech, pointing to the frequency of criminal sanctions imposed for failure to comply with orders to remove ‘unlawful’ content, and the arbitrary grounds justifying the government’s blocking of online platforms (Bratajaya and Tan 2023; Djajadiningrat 2023; Freedom House 2023b; Mediana 2023).
Similarly, a series of recently adopted laws in Pakistan have enabled the government to block platforms and shut down the Internet to silence dissent (RSF 2023; Access Now 2024; Hussain 2024)
Positive examples of countermeasures from the case studies mainly included digital literacy initiatives and social media–monitoring initiatives that engaged online platforms and independent civil society. In Bangladesh, a digital literacy centre was established, and the Ghanian cybersecurity authority held training sessions to support journalists in identifying disinformation tactics (Digital Literacy Center n.d.; Abubakar 2024). A 2022 initiative in Ghana demonstrated how semi-literate populations can be reached through education campaigns that focus on straightforward ideas like confirming information with another source (Bodine 2022). An especially notable partnership that drove digital and media literacy was observed in Mexico, between the EMB, civil society and online platforms (see Box 2.1). The case studies also revealed opportunities for the international community to partner with local efforts to help improve the level of digital and media literacy—for example, in Ghana and Mongolia. It is important to note that, while materials that support digital literacy efforts typically encourage critical thinking, Indonesia’s digital literacy programme has been criticized for using materials that instead promote compliance with state orders (Idris 2022).
Box 2.1. Mexico’s electoral commission partners with civil society, leverages innovation and raises awareness
In 2018, Mexico’s National Electoral Institute launched a campaign to combat the spread of disinformation on social media during electoral processes. The campaign was relaunched ahead of the 2024 elections with five main strategies: (a) verifying information; (b) encouraging bilateral cooperation between stakeholders; (c) cooperating with the media; (d) building media literacy; and (e) providing transparency and verified electoral information (INE 2024a). The campaign is a recent example where an agency partnered with platforms and civil society to provide citizens with comprehensive, accessible tools for fact-checking and verifying information.
As part of the campaign in Mexico, the INE launched a chatbot, integrated into WhatsApp, that allowed users to easily fact-check information, as well as to ask questions and receive verified information about the election sourced from a network of independent actors, including journalists (INE 2024a). Similar collaborations between the INE and platforms like Google and YouTube were aimed at strengthening digital protocols, bolstering reporting systems and increasing access to electoral information for general audiences (Lagos 2024).
Civil society initiatives—for example, the Ghana Fact-Checking Coalition and the Mongolian Fact-Checking Center—were especially helpful in identifying false narratives and carrying out real-time fact-checking (Meedan 2024; Modern Ghana 2024). The Ghana Fact-Checking Coalition was a notable example of regional civil society cooperation, as three West African organizations came together to support fact-checking in adherence with the established standards of the International Fact-Checking Network (Ghana Fact-Checking Coalition 2025). Platforms also showed a willingness to support in-app tools and media literacy efforts by engaging with local organizations, as in Ghana, Mexico and South Africa. Social media–monitoring efforts have shown promise for using AI-driven detection tools to help flag false content, including through the use of chatbots in Mexico and Mongolia. To support the monitoring of social media content in an independent and accountable manner, South Africa’s EMB partnered with online platforms and civil society, helping to drive greater transparency and oversight of online content (see Box 3.1).
There has generally not been significant progress when it comes to regulating AI. Bangladesh established a national strategy in 2020, though there is currently no evidence of progress (Information and Communication Technology Division 2020), and Pakistan’s Ministry of Information Technology and Telecommunication introduced a draft national AI policy, identifying ethical challenges regarding AI, though without any steps towards concrete action (Z. Khalid 2023; Pakistan 2022). The Indonesian Government released guidelines for the ethical use of AI, a useful first step, and committed to binding regulations in the future (ANTARA 2023b; Juniarto 2024; Sakura and Virgiany 2024).
In response to the identified challenges, development practitioners should support information integrity along two main avenues: (a) by helping to establish new rules for and enforce oversight of political communications online; and (b) by complementing these rules with collaborative measures aimed at enhancing ethical behaviour and societal trust. Actions should target not only EMBs but also oversight agencies, encouraging meaningful public engagement more broadly by engaging in multistakeholder discussions to agree upon new standards in a whole-of-society approach.
Planning can be oriented towards short, flexible, time-bound initiatives that are carried out in the context of elections, providing valuable data on the tactics and narratives surrounding disinformation and enabling a rapid response; long-term initiatives can focus on building trust and durable partnerships between public, private and civil society actors (Hanafin, Idrizi and Fischer 2024). The insights from the collected data will be key to analysing the drivers—both online and offline—of disinformation and to connecting them with efforts to increase trust in democratic institutions and the media more broadly (Schwieter and Gandhi 2024). Given the volume and rapid spread of information online, it is crucial that governments and the international development community treat Internet users as active agents of change rather than passive recipients of misleading information. Development assistance practitioners should consider actively empowering users, arming them with information and facilitating active exchanges between governments and citizens on implicit societal agreements and relationships, including on the integral role of reliable information. Programmes like Mexico’s media-literacy and trust-building initiatives are notable examples of such efforts, as they are aimed at empowering individuals to critically evaluate information and participate actively in public discourse.
This chapter will first address time-bound initiatives that are oriented towards maintaining information integrity in elections, both in terms of building the capacity of EMBs and in terms of supporting other actors and institutions that support election oversight. It will then discuss other long-term actions to build relationships between EMBs, electoral competitors, the media and civil society, and to enhance societal resilience through measures such as promoting media literacy and reducing inequality.
3.1. Establishing new rules to regulate the online ecosystem and navigating oversight capacity in implementation and enforcement
3.1.1. For EMBs
When approaching election reforms to respond to the risks posed by AI, it is essential to establish clear limits on its use in political campaigning, through both legislative and voluntary measures, and to encourage campaigners to proactively declare and label AI-generated content (CALD 2024). Another priority is to require AI communications to include content provenance features, including when shared by third parties on behalf of political campaigns (Stockwell et al. 2024).
There is also a pressing need to uphold the fundamental right to free speech while protecting the integrity of information, but there are different viewpoints on where to draw the line. For example, fact-checking can be fairly clear-cut where it concerns basic electoral information but can get murkier when it comes to matters of opinion (Hanafin, Idrizi and Fischer 2024). Where legislation focuses on content, this necessarily means that an oversight body must judge the veracity of the content and decide whether to impose sanctions or order its removal, creating the potential for abuse of the rules.
It is therefore recommended that content regulation be approached with the utmost caution, as determining where to draw the line on acceptable uses of AI—for example, to protect creative freedom in political parody—will require considerable discussion. However, the same sort of creative freedom was used to depoliticize the image of Prabowo, one of the candidates in Indonesia’s presidential election, who used 3D animated personas to distance himself from his role as a military general in a past dictatorship (Utama 2023; Duffy 2024).
It can be useful to support legislatures and regulatory agencies in tailoring existing efforts elsewhere to local contexts. Examples include the ‘ASEAN Guide on AI Governance and Ethics’ (2024), which underscores key principles tailored towards the regional context of South East Asia, or the EU’s AI Act (2024), in particular its risk-based approach. It is imperative that these efforts emphasize monitoring and the independent auditing of the use of AI by political parties, including of the funds used to produce AI-generated content, to help ensure that its use does not contribute to bias or an uneven playing field (Juneja 2024).
When approaching development planning targeting elections, it is vital to keep in mind the broad constellation of actors that engage in electoral oversight, from EMBs to media authorities to data protection agencies. Establishing dedicated channels of communication and improving coordination between institutions is crucial, ensuring that mandates are clearly defined and efforts are not duplicated. A useful priority area for strengthening connections is the relationship between the legislature—as it establishes new rules on political campaigning and communications—EMBs and other electoral oversight actors. These connections can feed into ongoing discussions to establish legal and voluntary standards—for example, by helping EMBs and other agencies to clarify how election-related rules and regulations are to be implemented.
The case studies provided useful examples on EMB guidelines: the Ghanian EMB released guidance on online campaigning, and the Indonesian communications agency outlined guidelines for the ethical use of AI (CALD 2024). The Mexican EMB developed comprehensive media literacy training and offered digital solutions for interactions with voters (Valverde 2024; Vergara 2024).
EMBs can also issue guidance to support the media, helping them to report on synthetic content or false narratives by linking them with deepfake detection tools and reporting mechanisms and ensuring open communication (Stockwell et al. 2024). Oversight authorities can also work to establish digital literacy programmes or guidelines to help the public better understand digital election processes and safeguards, helping to restore trust in the process (Cetina-Presuel et al. 2024).
Technical support for EMBs can include increasing capacity for oversight of campaign finance, social media monitoring (both through platform tools and by supporting access to commercial tools) and voluntary collaborative agreements between electoral stakeholders (Grohmann and Ong 2024). It is also useful to help EMBs to develop resilience against threats by developing capacity to gather and analyse evidence to identify and prepare for potential issues online. These efforts can be supported by more systematic data collection and analysis of the use of AI for political campaigning, including with simulations of AI-enabled influence operations (Stockwell et al. 2024). Supporting the development of AI tools can help to better allocate resources according to predicted voter turnout, reduce mistakes in voter registries and enable improved monitoring of polling stations on election day (Juneja 2024). Examples include Indonesia’s Voter List Information System, which has optimized the accuracy of voter registration and list management, and AI-powered tools that have been incorporated into candidate registration and vote counting. These tools also support real-time adaptability, enabling a rapid response to developing situations (Yu 2024). However, it is important to note that the use of AI-powered tools by political parties and EMBs must be transparent and clearly disclosed to the public at all times to ensure that their use does not damage public trust. It is also crucial that this work is accompanied by cybersecurity training and regular audits.
It is also important to support EMBs with social media monitoring and fact-checking, while remaining vigilant to ensure that any measures taken do not limit free speech.
It is therefore recommended that development actors encourage a constructive approach to monitoring online spaces that focuses on understanding the information ecosystem and correcting inaccurate information, with strict criteria for removing or blocking content. It may also be useful to create a repository of election materials that can be used to monitor trends in political communications and the spread of AI-generated materials, helping to inform EMB strategies and future public information campaigns (Stockwell et al. 2024). AI-powered tools are also currently being tested to support the monitoring and identification of disinformation and hate speech online (Tuohy-Gaydos 2024). Such initiatives can also be supported by linking EMBs with independent CSOs and helping to build their technical capacity and resources.
EMBs can, for example, partner with online platforms and civil society in ‘pre-bunking’ programmes, which teach voters to recognize and resist misinformation and disinformation and put in place mechanisms to alert electoral authorities and quickly respond (Cetina-Presuel et al. 2024). This approach was demonstrated in Pakistan, where the CSOs Soch Fact Check and AFP Fact Check have worked to counter disinformation (Shah and Majumdar 2024). Soch Fact Check signed a third-party fact-checking agreement with Meta, helping it to directly flag misleading or false content, which the platform can then take action on. This kind of partnership can also be useful in balancing online fact-checking with potential threats to free speech.
3.1.2. For oversight more broadly
Turning from support for EMBs to establishing effective legislation on data privacy, online campaigning and the ethical use of AI, it is vital for support to take a non-prescriptive approach, bringing stakeholders together to develop rules tailored to a country’s context. A collective multistakeholder process that involves consultations with diverse voices, upholds human rights and strikes a balance between various competing interests in the online sphere is recommended (Bereskin 2023). Moreover, such an approach will help to ensure that technological solutions can be implemented in a responsible and transparent way as they are developed and become more widely accessible.
The case studies highlight the risk that anti-disinformation efforts could lead to the suppression of free speech; incorporating the voices of the most vulnerable and developing layered oversight measures can help in addressing this risk. The strictest care is recommended in approaching legislation that specifies both acceptable and unacceptable online behaviour, as legislative approaches that are overly concerned with policing speech on social media have tended to result in further authoritarian capture (Grohmann and Ong 2024). For example, following UNESCO’s recent attempt to encourage the publication of guidelines on content regulation, 40 non-governmental organizations from across the Asia-Pacific region signed a letter emphasizing the dangers of such recommendations in authoritarian contexts (Open Net 2023). Their response underscores the need for continued vigilance against the potential misuse of rules governing online speech.
Where AI-specific rules are established, attention should be given to maintaining a human-centric perspective and emphasizing the principles of transparency, accountability, explainability, fairness and inclusivity (Bontcheva et al. 2024; Dawson and Ball 2024). Data privacy and security are also crucial; notably, it is important to limit the use of personal data for microtargeting for political purposes, especially regarding sensitive categories of information such as race and sexuality (Bieber et al. 2024). It is important that these standards are made specific and clear in how they apply to elections to prevent grey areas in which unchecked behaviour can lead to significant impacts (Yu 2024). Measures to promote greater transparency may include disclosure requirements on the use of AI and the underlying data, with rapidly developing provenance standards also showing the potential to improve transparency when it comes to the original content creator (Bereskin 2023). These principles may help to ensure sufficient oversight of the use of AI and that citizens are better able to understand the content they interact with.
Ensuring human oversight is a key step in supporting the transparent, accountable and explainable use of AI systems. Human oversight is also helpful in making necessary adjustments to counter bias and ensure that high-quality training data is used (Bieber at al. 2024; CALD 2024). When the reasoning for an AI system’s outputs is unclear, it becomes more challenging to improve the system, identify potential bias and maintain trust in the system on the part of users (Tuohy-Gaydos 2024). There is increasing evidence that a reduction in societal trust fuels false narratives online. Approaches to AI ethics should therefore ensure that ethical principles are applied in order to reassure the public that AI is being used responsibly.
A useful starting point is to help legislators in identifying and addressing gaps in existing legislation that enable the misuse of AI in political processes. Key areas to consider, as outlined by Yu (2024), include the following:
- a lack of transparency requirements mandating the disclosure of AI use, including in voter targeting, content personalization and party strategy optimization, which would help to support accountability for party communications;
- the absence of data protection measures that encompass the highly sophisticated data analysis and profiling made possible through AI-powered tools to ensure that sensitive data is not exploited without users’ consent or knowledge; and
- gaps in regulatory and ethical standards regarding both binding and voluntary standards aimed at maintaining a level playing field in elections and ensuring meaningful public deliberation both online and offline.
Discussions can also be guided by the risk-based approach developed in the EU AI Act, the world’s first comprehensive legal framework regulating AI systems. Central to this act is an approach that categorizes AI systems based on the potential harm they can cause and imposes obligations depending on the risk level. Since AI systems and tools intended to influence election outcomes are considered high-risk, they are subject to strict requirements, including risk assessments, transparency measures such as labelling and robust testing (Rabitsch and Calabrese 2024). Such tools include those that use AI to generate political advertising, profile voters and automate decision making to influence voter behaviour. Recognition of the potential impact of such tools can help to establish common ground among stakeholders that wish to develop a risk-based approach tailored to local contexts.
It is essential to keep the following two limitations in mind: first, existing tools developed to detect the use of AI are unreliable; second, the rapid development of technology and multitude of actors involved in enforcement make implementation and enforcement a persistent challenge (DGA Group and Albright Stonebridge Group 2024). Notably, more progress has been made in training specialized detectors that target specific kinds of manipulation, but a universal detection model is not currently technically feasible. Moreover, the danger of false positives may only further fracture the information landscape and intensify distrust (Bontcheva et al. 2024; Schwieter and Gandhi 2024). It is also important to note that the vast majority of these tools have been trained on English-language models, greatly limiting the possibility for their application in developing contexts (Bontcheva et al. 2024). Countries’ ability to counteract risks online therefore hinges especially on their resources and technical capacity and the nature of their information ecosystem, making tailored legislation and regulations all the more important.
It is also crucial that these reforms incorporate monitoring processes that obtain gender-disaggregated data that is used for in-depth analysis on how pollution of the information space and misuse of AI disproportionately impact women and other vulnerable populations. It is essential to apply a cross-cutting gender perspective across all priorities, given the complexities of gender and how it intersects with other group identities. Such an approach can help vulnerable groups build the capacity to organize and advocate for themselves even in the face of false narratives (Hanafin, Idrizi and Fischer 2024). Gender-disaggregated data can help in developing AI tools to support women leaders; several projects already exist that support the tracking, detection, reporting and removal of sexist and harassing content online (Tuohy-Gaydos 2024).
A final consideration for regulation targeting AI ethics in the developing context is the matter of ‘AI sovereignty’, or the ability of nations to independently govern their own AI systems. Attention should therefore be given to developing the capacity of partner countries to harness AI for their own development in the economic, social and security spheres without becoming overly dependent on foreign providers (CALD 2024). Such efforts can be supported by helping countries to develop their own AI strategies or by strengthening existing initiatives, such as with standard-setting efforts in Bangladesh, Indonesia and Pakistan.
3.2. Complementing new rules with a whole-of-society approach and soft-law solutions
3.2.1. For EMBs
In approaching long-term planning regarding information integrity during electoral periods, a whole-of-society approach is crucial. Alongside building oversight capacity, civic engagement and participation also need targeted action to build checks and balances, not only for traditional CSOs but also for other actors, such as influencers, community leaders, academics, journalists and human rights defenders (Hanafin, Idrizi and Fischer 2024). A central goal of these efforts should be to encourage proactive transparency and ethical campaigning to complement formal legislative standards, especially through developing voluntary collaborative measures such as codes of conduct (see Box 3.1).
Key features of codes of conduct can include commitments both on fair and transparent campaigning and communication and on internal party measures. Existing examples have included measures calling on parties to refrain from discriminatory statements, to disclose relevant information on political advertising, to limit the use of sensitive personal data or foreign funding in placing ads, to promote internal compliance with commitments, to build cybersecurity resilience, and to conduct a post-election review (International IDEA 2024b). These agreements are especially useful in supporting a sense of trust between parties, oversight agencies and the media to commit to transparency and to help raise awareness among voters and improve their knowledge, creating a useful foundation for raising ethical standards and increasing ambitions.
Codes of conduct can also create useful synergies with tech companies, some of which have collaborated on the development of voluntary standards: for example, the Tech Accord to Combat Deceptive Use of AI in 2024 Elections is a voluntary framework that had the following goals: (a) preventing the spread of misinformation and disinformation; (b) ensuring the provenance of content; (c) detecting AI-generated content; (d) providing responsive protection; (e) putting evaluation mechanisms in place; (f) raising public awareness; and (g) improving resilience (Soon and Quek 2024).
However, it is important that soft-law solutions are complemented by targeted regulatory and legislative measures, as the lack of enforcement powers associated with voluntary measures necessarily limits their impact.
Public awareness campaigns organized by EMBs can build on these efforts, providing guidance for how individuals can critically engage with the content they see online, helping them to access trusted sources of verified electoral information, understand the risks of using AI chatbots and find out where they can report content (Stockwell et al. 2024). It is also useful to inform the public about the establishment of mechanisms for reporting false content and direct channels for submitting complaints regarding actions taken by platforms (Cetina-Presuel et al. 2024). These measures can be supported by promoting media integrity, digital literacy and civic education more broadly.
Box 3.1. Cooperation between online platforms, EMBs and civil society
In 2023 South Africa’s Electoral Commission collaborated with Google, Meta, TikTok and the civil society organization Media Monitoring Africa (MMA) to sign a framework of cooperation, making commitments to act in good faith, respect the data privacy of users and take certain measures against disinformation (IEC 2023). The fact that this framework was agreed sufficiently in advance of the country’s planned general election enabled stakeholders to prepare systems to ensure compliance with the commitments outlined in the framework.
Stakeholders took advantage of the time available to establish several key mechanisms—an independent three-member commission to review reported cases of misinformation and to make recommendations to the IEC, which could in turn request that the malicious content be removed (IEC 2023). Meta established a South Africa–specific elections operations centre that brought together experts from both Meta and South Africa to identify threats in real time, with an elections resource centre also established to provide users with information on how to secure their accounts and identify trusted sources of information (Idé Siddo 2024). The Google News Initiative partnered with Africa Check, an independent fact-checking organization, to provide tools to citizens and journalists to help assess the veracity of information online (Ndlovu 2024). Finally, the IEC established the Real411 complaints platform, administered by the MMA, enabling the public to report online misinformation (European Union Election Expert Mission 2024).
These efforts resulted in greater transparency and oversight of online content, demonstrating both the usefulness of crowdsourced data and the potential of AI-powered tools to counter the spread of disinformation online. The Real411 platform both acted as an early-warning system and provided specific information on which news items were most closely tied to false content, providing fact-checkers and digital literacy educators with key data on the most helpful approaches (Bird and Smith 2024). The platform received approximately 200 complaints, 47 per cent of which were deemed information manipulation, revealing the scale of disinformation during the election (European Union Election Expert Mission 2024).
The partnership between the Google News Initiative and Africa Check helped to supply AI-powered tools such as an alerts function, which flagged fake news on social media; a search function, which enabled journalists to find factual claims more easily; and a transcription function, which furnished accurate records of campaign speeches (Ndhlovu 2024; Ndlovu 2024).
However, there were also notable limitations to the effort. As with other voluntary agreements to tackle electoral disinformation, the cooperation framework’s non-binding nature limited its impact, and the fact that X did not agree to be party to the agreement was also a major limitation (Ndlovu 2024). Although WhatsApp—an encrypted messaging application owned by Meta—was party to the agreement, there were limitations on the degree to which it could track the flow of misinformation (Dionne and Liu 2024).
Finally, a review of social media platforms’ efforts revealed that more fact-checkers and monitors who are fluent in South Africa’s various official languages were needed to enhance early detection of misinformation, in addition to more efforts to counter cross-platform misinformation (Ngubane n.d.; European Union Election Expert Mission 2024).
3.2.2. For oversight more broadly
As the case studies demonstrated, pollution of the information space can be driven by a variety of actors and societal factors, which also influence the ways in which AI-generated content is leveraged online, requiring a long-term perspective in planning. While the exact political, legal and social factors across countries will differ, one commonality across the case studies that drives false narratives is inequality. There is increasing evidence that addressing inequality, especially economic and digital inequality, is crucial to ensuring that citizens are able to assess the information they receive, advocate for their own interests and reject extremist narratives (Grohmann and Ong 2024). In Mexico, for example, Breuer (2024) identified a number of structural conditions that contributed to the spread of disinformation during the country’s elections—persistent poverty and socio-economic inequalities; high levels of violent crime and corruption, especially organized crime; a highly concentrated media market with low levels of transparency; gaps in guarantees to access to information; low trust in government overall; and broader disenchantment with both the political establishment and democracy.
Another form of inequality to consider is the outsized power that online platforms hold, with smaller nations often struggling to receive cooperation or compliance from the companies that own these platforms. In the long term, development efforts should also aim to establish trust and collaboration between political organizations, tech developers and independent auditors, which can be leveraged for collaboration on AI systems (CALD 2024). Supporting active, engaged and effective legislatures can help in these efforts, as representatives form crucial links between the public, political party apparatuses, regulatory agencies and online platforms. Members of parliament can be instrumental in engaging the public to raise awareness and improve media literacy and can help to forge cross-border connections with counterparts in other countries, sharing knowledge and best practices, and supporting a harmonized response that can create regional resilience (Bereskin 2023). Supporting capacity building and partnerships between public institutions, the media and civil society can also help to build trust and a culture of openness, establishing collaborative initiatives to promote access to reliable and accurate sources (Breuer 2024). Finally, supporting worker and media rights, especially ensuring whistleblower protections and freedom-of-information acts, can help to empower everyday citizens to take action where they witness the spread of disinformation or the unethical use of AI (Grohmann and Ong 2024).
Building access to quality information and journalism and supporting media literacy and civic education to help the public evaluate information and take informed decisions are key tools in the arsenal of societal resilience against false narratives. The role of a free, independent and well-funded media is therefore more crucial than ever, both in terms of its role as a watchdog to keep representatives accountable and in keeping citizens informed. Journalists are ideally placed to counter information manipulation online, especially through targeted support for training in fact-checking, the strategic use of open government data and freedom-of-information requests, and data analysis (Breuer 2024). Increasing the safety of journalists is of the utmost importance, and ensuring that traditional, non-commercial, local and community media are not forgotten—which may be important sources of information for certain segments of the population—is also a priority (Breuer 2024). However, governments may not be receptive to initiatives to support free media, as they themselves may be subjected to critical reporting, underscoring the importance of engaging in a continuous constructive dialogue with governments on improving conditions for human rights.
The media landscapes in many countries have rapidly changed in recent years, meaning that journalists may need help adapting in order to understand how and with whom to engage, as well as how to navigate the increasingly influential sphere of non-traditional media sources (Hanafin, Idrizi and Fischer 2024). In the long term, supporting coalition building in the media can have a significant positive impact: the resulting networks can be leveraged at critical moments, helping organizations grapple with the massive scale and sophistication of communications, as well as in the face of attempts to repress speech or manipulate electoral results.
It is essential to ensure support for a balanced media landscape and assistance to help media organizations become self-sustainable, as an overreliance on state funding or overseas development aid could leave these organizations vulnerable to undue influence or disruption (Hanafin, Idrizi and Fischer 2024). Securing the participation of different members, especially in highly polarized contexts, may also be a challenge.
Citizens are also in need of support in navigating the rapidly changing information ecosystem. Measures to support media integrity should be complemented with civic education efforts to improve media literacy, both in schools and through training and awareness-raising campaigns targeting adults. These measures can teach individuals to evaluate sources of information, understand the mechanisms behind targeted and AI-generated content, and more easily differentiate between credible and false information (Yu 2024). These measures should also focus on digital inclusion, supporting efforts to ensure that the public is aware of and has access to digital tools in general in an equal, effective and safe manner (Hanafin, Idrizi and Fischer 2024). It is a priority that media literacy and digital inclusion efforts go hand in hand, as progress in one will be hindered by stagnation in the other—for example, widespread access to digital tools may exacerbate online challenges if citizens do not understand what they are doing.
In supporting the development of media literacy campaigns, it is important to engage local actors that can support a tailored programme based on the disinformation landscape in the respective country. Since organized crime has been identified as a major source of disinformation in Mexico, for example, media literacy efforts should therefore include information specific to that context (Breuer 2024). Media literacy campaigns can also leverage partnerships with social media platforms to help users to understand what they are seeing, underscoring the importance of building relationships between local representatives, oversight agencies and online platforms. Finally, it is crucial that gender is established as a cross-cutting issue in these efforts, as increasing the inclusion of women in the digital, societal and political spheres and countering targeted online violence and harassment are key to protecting the rights of women and girls (Hanafin, Idrizi and Fischer 2024).
Box 3.2. Regional and international cooperation to achieve information integrity and the ethical use of AI in EU partner countries
A critical component of a comprehensive approach towards supporting information integrity and the ethical use of AI is fostering durable networks between different stakeholders—regionally and internationally—that are rooted in trust, knowledge sharing and regular discussions. Efforts should especially be made both to address the cross-border nature of disinformation and to support countries in counterbalancing the significant power differential between developing nations and tech giants.
Regional networks involving key government agencies, in particular EMBs and other electoral oversight bodies, and legislatures are useful components of this approach (Bereskin 2023; Cetina-Presuel et al. 2024). Building trust between oversight bodies can bolster information sharing on AI threats and other electoral vulnerabilities, helping countries to keep up with increasingly sophisticated cyberattacks by collectively anticipating challenges and coordinating responses (Yu 2024). Promoting collaboration between civil society and academia can also be useful in supporting innovative research that can both drive the development of new technologies for detecting synthetic media and countering online hate speech and facilitate media literacy and public awareness campaigns.
Regional cooperation can also help in harmonizing legal and policy approaches to disinformation and political uses of AI, especially cooperation with existing regional bodies that are already working on the topic. The ‘ASEAN Guide on AI Governance and Ethics’, for example, has created a useful starting point for further discussions in the Asia-Pacific region, demonstrating that such initiatives can support the development of strong common legal frameworks tailored to the unique political, social and technological contexts in a region (CALD 2024). The harmonization of legal frameworks on political communications and the use of AI can close loopholes and prevent foreign actors from exploiting gaps in regulations, especially during elections (Yu 2024). Pooled resources can also help developing nations to overcome limitations in their technical and financial capacities.
It is also important to adopt an international perspective when fostering cooperation among partner countries. A key finding of the research for this report is that, while disinformation and generative AI affect countries in the Global North and the Global South alike, the context in which their impact occurs is very different. Therefore, taking a global perspective when advocating for common solutions and localizing regional and national solutions based on shared lessons—from both the North and South—are priorities (Grohmann and Ong 2024).
This report has considered the information ecosystem during 2024 elections in seven of the EU’s developing partner countries in respect of new methods of online communications enabled by online tools. The case studies revealed a variety of mechanisms for communicating with voters; there is evidence that many of these mechanisms were used to manipulate voters. In particular, coordinated networks of influence worked across the information ecosystem to amplify certain narratives in all seven countries observed, leveraging the unique features of different social media platforms and microtargeting voters to various extents. AI-generated content spread widely in online discussions in the lead-up to elections in six of the seven countries—all but Mongolia—most notably in Indonesia. Meanwhile, government efforts to counter disinformation played a major role in suppressing dissent and had a chilling effect on free speech, with positive examples of measures frequently coming from civil society–led initiatives on digital literacy and social media monitoring.
Based on the analysis provided in this report, it is recommended that development practitioners work along two main avenues: (a) by guiding the development of new rules and building capacity for oversight of political communications online that focus more on understanding and correcting false information, with strict criteria for taking down or blocking content; and (b) by complementing these rules with collaborative measures aimed at enhancing ethical behaviour and societal trust, especially by addressing inequality and engaging in civic education and digital and media literacy initiatives.
It is vital to take a broad view of the information ecosystem and the various ways in which different factors, particularly related to disinformation, AI and cybersecurity, interact to impact the integrity of elections and meaningful public discussion. With political discussion increasingly taking place online, as more and more of the developing world begins using social media, narratives are frequently crafted long before official campaign seasons begin, with candidates potentially smearing an opponent’s identity before they can even begin to advocate for themselves (Shah and Majumdar 2024). Developing legislation targeting information integrity and the ethical use of AI and addressing those structural societal factors that serve to drive disinformation should go hand in hand.
Amid the unprecedented speed and scale of AI-related developments, making rules and establishing cooperation in the online sphere are continual processes. Leveraging multistakeholder alliances and international and regional partnerships for knowledge sharing and evidence-based decision making and investing in monitoring, evaluation and feedback mechanisms to help guide improvements are key. Establishing two-way communication and partnerships between governments, civil society and online platforms is also a priority. The EU and its member states are ideally placed to build relationships and have productive discussions, with the EU leveraging its position as an international standard setter and development partner that can ensure mutual benefits for all stakeholders.
As multistakeholder solutions are explored, it is important to pay attention to establishing trust between stakeholders wherever possible, as it is fundamental for building democratic values and effective public institutions (Schwieter and Gandhi 2024). It is also crucial that the impacts of information pollution on women and other vulnerable minorities are kept in mind at all times, including by collecting disaggregated data on different groups, such as young people or rural populations. This information should also be used to ensure that solutions are made accessible across countries’ diverse populations and to help ensure that the increased use of AI-powered tools does not result in unintended bias. In this way, development planning can help to foster a holistic, long-term, whole-of-society approach that can help restore trust in society and democracy.
Specific recommendations from the analysis for development practitioners—whether in the EU or its member states—are outlined in the remaining sections of this chapter.
4.1. Recommendations for development practitioners
- Support discussions aimed at developing locally owned legislation and regulation for the use of AI-powered tools, both during and between electoral periods.
Rules and standards established offline should apply equally online, and they should be established in light of the unique circumstances of AI-powered communications. As with regulating misinformation and disinformation, a balanced approach helps to ensure that harmful practices are reduced, while protecting legitimate uses and freedom of speech. It is also crucial that any discussions on these issues aim to understand the various categories of risk that may apply at different times, noting the importance of AI ethics both during and between electoral periods.
Several universal principles can guide these discussions—transparency, accountability, fairness and inclusivity, privacy and explainability. These principles are key to ensuring trust in these technologies and to maintaining societal trust more broadly, as they can help to ensure meaningful understanding and effective oversight of AI. It is also useful to encourage AI sovereignty throughout these conversations, as supporting a country’s ability to independently develop, control and regulate AI systems without relying on foreign technology, infrastructure or policies can help to navigate the global power imbalances associated with tech giants.
Discussions can focus on developing local standards for risk assessment of digital infrastructure, aimed at both cybersecurity risks and societal risks, with the EU’s AI Act providing a useful starting point.
Initial measures that could be encouraged can include the development of legislation on AI transparency and disclosure, notably on clear labelling and provenance mechanisms. Such mechanisms should ensure that information is provided about how an AI technology was used, and the public should be provided with this information in a timely manner and in a concise, machine-readable format.
The development of voluntary measures such as codes of ethics to help identify where there is already agreement and find areas for future cooperation may also be beneficial.
These discussions must include communities at the local level that can actively participate in shaping legislation and regulation for the use of AI tools, not only during electoral periods but continuously, to ensure that these technologies align with public values and protect local interests. As we see in the case studies, engagement of local communities must go hand in hand with making AI systems more transparent for non-technical stakeholders. These communities can be instrumental in evaluating AI systems at the local level, reflecting community needs and providing public feedback.
- Assist state institutions, such as EMBs and parliaments, in increasing their capacity to counter digital threats, especially by driving multistakeholder discussions rooted in continual, meaningful collaboration between political representatives, government agencies, AI and social media companies, civil society and academia.
The scale and complexity of issues associated with disinformation and AI require collaboration across a range of actors with expertise in diverse areas. Crucially, this collaboration should be aimed not only at knowledge sharing but especially at building trust between actors and institutions to encourage proactive measures and a culture of openness. At the international level, it is important to support efforts to learn from the experiences of the Global South, as knowledge gathering on approaches to online regulation have traditionally been dominated by Western voices.
A key component of collaborative efforts should be helping journalists, civil society and academia to join forces to help bridge gaps in capacity and resources in the face of the fast-paced and complex developments taking place in this area. Cross-sectoral collaboration has the potential to make a significant impact—for example, by connecting behavioural scientists with data scientists to gain insights into the actual impacts of AI-driven disinformation, which can then be communicated to the public by the media and civil society. Such insights can help not only to drive evidence-based decision making but also to ensure that the dangers of AI are not overblown, a phenomenon that has also been shown to damage societal trust.
Concerted efforts should be made to connect EMBs with other government agencies and with civil society and media initiatives to help them accomplish the enormous task of regulating elections both online and offline. Cooperation may also encompass public outreach initiatives, especially by facilitating the formation of partnerships between the government, civil society and academia to help people evaluate information online and understand where to find accurate information, particularly on voting processes. Various electoral stakeholders, in particular political parties and online platforms, should establish shared codes of conduct to encourage ethical behaviour and to set standards where laws may place an undue limitation on free speech.
Technical support for EMBs should help determine oversight responsibilities for online political campaigning and its funding, particularly by elaborating clear rules through a combination of legal and voluntary measures on what is and is not acceptable behaviour. It is also useful to explore potential ways that AI can be used to support EMBs throughout the election cycle, such as by reinforcing the distribution of election information to reduce disparities and promote inclusion in the voting process, and by supporting efforts to monitor social media to detect misinformation and specific content that might violate election laws. A priority is that, in establishing these mechanisms, regular third-party audits are envisioned to provide the basis for regular adjustments and to correct for bias, in as transparent a manner as possible to help ensure public trust in the electoral process.
These efforts may also incorporate social media–monitoring and fact-checking initiatives that focus on improving individuals’ understanding of the online information space, with strict criteria for taking down content. It is important that these initiatives are transparent regarding the purpose of monitoring, its scope, and the safeguards that are in place to ensure public trust in these initiatives and to avoid overreach. It is recommended that these initiatives take place during specific electoral periods, as governments may lack the capacity for continuous monitoring. Moreover, this sort of monitoring gives rise to surveillance concerns.
- Complement regulations in the digital sphere with efforts to target the societal factors that contribute to pollution of the information space, especially by addressing inequality and enhancing social trust, media integrity and digital literacy.
Countering disinformation and supporting the ethical use of generative AI are global issues for which there is still ongoing international discussion, and a number of technological solutions are also being pursued to support information integrity online. Given the rapidly developing pace of technology, any solutions put in place will need constant review and updates. The task is not only to suppress disinformation and poor-quality information, while promoting a healthy information ecosystem, but also to redefine the role of the recipients of information online.
Therefore, it is recommended that practitioners invest in actions to work closely on citizenship-building initiatives, thus building societal resilience against information pollution. Such initiatives should target both the online and offline media spaces, as narratives that start online may be inadvertently spread by traditional media. Integral to these efforts is the goal of increasing people’s access to reliable news sources and restoring their trust in the media, one another and democracy more broadly, especially by supporting transparency and critical engagement with the media.
Initiatives in this area include supporting civic education tailored towards specific actors and drivers that are unique to the country in question and raising public awareness of the possibilities for online manipulation. It is also essential to carry out media literacy and public awareness campaigns, both in the context of elections and between electoral periods, helping citizens to recognize and resist false narratives, as well as to improve their understanding of and trust in digital election processes.
It is also crucial to support strong, free and independent media, not only for commercial outlets but also for non-commercial and local media. Ensuring the safety of journalists and countering harassment and violence towards them is paramount. It is also important to enable journalists, civil society and academia to access information, not only from government sources but also from online platforms, especially regarding algorithmic transparency.
Practical steps include engaging institutions such as legislatures, oversight agencies (including EMBs) and CSOs to partner on civic education efforts aimed at improving digital literacy. These actors should also be included in an analysis of inequality and drivers of disinformation and societal distrust to better understand the next steps that should be taken. Forming partnerships between civil society and local media may also be beneficial, as such partnerships can help build capacity for oversight of political communications and engaging citizens.
Abubakar, I., ‘Cyber Security Authority educates journalists on misinformation and disinformation’, 3News, 10 October 2024, <https://3news.com/news/cyber-security-authority-educates-journalists-on-misinformation-and-disinformation>, accessed 8 March 2025
Access Now, ‘#KeepItOn in Pakistan: Ensure open and secure Internet for all during the 2024 elections’, 6 February 2024, <https://www.accessnow.org/press-release/keepiton-pakistan-2024-elections>, accessed 25 February 2024
African Digital Democracy Observatory (ADDO), ‘Unravel South Africa’s election chaos and Zuma’s influence’, Medium, 9 May 2024, <https://disinfo.africa/unravel-south-africas-election-chaos-and-zuma-s-influence-b9e9880ff3d3>, accessed 11 February 2025
Agence France-Presse (AFP), ‘Fake experts drive disinformation before Bangladesh polls’, France 24, 7 September 2023, <https://www.france24.com/en/live-news/20230907-fake-experts-drive-disinformation-before-bangladesh-polls>, accessed 5 March 2025
Aguirre, S., ‘¡Ay, no! La desinformación ya llegó a Telegram: Este es el “modus operandi”’ [Oh no! Disinformation has already reached Telegram: This is the modus operandi], Animal Político, 24 March 2021, <https://www.animalpolitico.com/tendencias/actualidad/canales-desinformacion-telegram-como-funcionan>, accessed 17 May 2025
Ajder, H., Patrini, G., Cavalli, F. and Cullen, L., ‘The State of Deepfakes: Landscape, Threats, and Impact’, DeepTrace, September 2019, <https://regmedia.co.uk/2019/10/08/deepfake_report.pdf>, accessed 17 April 2025
Ali, F., ‘How deepfakes, AI and fake news impacted Pakistan’s elections 2024’, Samaa TV, 12 February 2024, <https://www.samaa.tv/208739652-how-deepfakes-ai-and-fake-news-impacted-pakistan-s-elections-2024>, accessed 21 February 2024
Allen, K. and le Roux, J., ‘Under the Influence? Online Mis/Disinformation in South Africa’s May 2024 Election’, Institute for Security Studies, Southern Africa Report 61, December 2024, <https://issafrica.s3.amazonaws.com/uploads/pages/1734073957062-sar-61.pdf>, accessed 11 February 2025
Amnesty International, ‘Bangladesh: Government must remove draconian provisions from the Draft Cyber Security Act’, 31 August 2023, <https://www.amnesty.org/en/latest/news/2023/08/bangladesh-government-must-remove-draconian-provisions-from-the-draft-cyber-security-act>, accessed 6 March 2025
ANTARA, ‘Govt provides AI training to people through digital literacy movement’, 7 December 2023a, <https://en.antaranews.com/news/300489/govt-provides-ai-training-to-people-through-digital-literacy-movement>, accessed 20 February 2025
—, ‘Govt to prepare special AI regulation’, 22 December 2023b, <https://en.antaranews.com/news/301644/govt-to-prepare-special-ai-regulation>, accessed 19 February 2025
Associated Press (AP), ‘Post-communist generation is hoping for a new era of democracy in Mongolia’, NY1, 5 July 2024, <https://ny1.com/nyc/all-boroughs/ap-top-news/2024/07/05/post-communist-generation-is-hoping-for-a-new-era-of-democracy-in-mongolia>, accessed 17 February 2025
Association of Southeast Asian Nations (ASEAN), ‘ASEAN Guide on AI Governance and Ethics’, February 2024, <https://asean.org/wp-content/uploads/2024/02/ASEAN-Guide-on-AI-Governance-and-Ethics_beautified_201223_v2.pdf>, accessed 1 July 2025
Bahl, V., ‘Pakistan political turmoil: These “PTI protest” images aren’t real’, France 24, 17 May 2023, <https://www.france24.com/en/tv-shows/truth-or-fake/20230517-pakistan-political-turmoil-these-pti-protest-images-aren-t-real>, accessed 14 February 2025
Baksh, S. W., Mungra, D., Ahasan, N., Pariwala, V., Jain, R., Surve, A., Banerjee, P., Mustofi Eron, S., Mehta, S. and Rashid Diya, S., ‘How Facebook has become a political battleground in Bangladesh’, Tech Global Institute, 19 December 2023, <https://infolab.techglobalinstitute.com/how-facebook-has-become-a-political-battleground-in-bangladesh>, accessed 4 March 2025
Bangladesh, People’s Republic of, Digital Security Act 2018, 8 October 2018, <https://natlex.ilo.org/dyn/natlex2/natlex2/files/download/110029/BGD110029.pdf>, accessed 17 April 2025
Bangladesh Awami League, ‘Awami League plans big social media campaign for upcoming election’, 15 October 2023, <https://albd.org/articles/news/41092/Awami-League-plans-big-social-media-campaign-for-upcoming-election>, accessed 4 March 2025
Batdorj, B., ‘Social media and public discourse in Mongolia’, Friedrich-Ebert-Stiftung, 27 May 2020, <https://asia.fes.de/news/social-media-and-public-discourse-in-mongolia.html>, accessed 18 February 2025
Battulga, B., ‘Real-life success: Overcoming COVID-19 through healthy life style and vaccination – How exercise and vaccination can help us win COVID-19’, UNICEF Mongolia, 1 February 2023, <https://www.unicef.org/mongolia/stories/real-life-success-overcoming-covid-19-through-healthy-life-style-and-vaccination>, accessed 19 February 2025
Belinda, Y., ‘Is K-Pop Indonesian political parties’ latest tool to entice young voters?’, The Jakarta Post, 21 March 2023, <https://www.thejakartapost.com/culture/2023/03/21/is-k-pop-indonesian-political-parties-latest-tool-to-entice-young-voters.html>, accessed 13 February 2025
Bereskin, C., ‘Parliamentary Handbook on Disinformation, AI and Synthetic Media’, Organization of American States and Commonwealth Parliamentary Association, 2023, <https://www.cpahq.org/media/sphl0rft/handbook-on-disinformation-ai-and-synthetic-media.pdf>, accessed 17 April 2025
Bieber, C., Heesen, J., Grunwald, A. and Rostalski, F., ‘AI in the Super Election Year of 2024’, Lernende Systeme, White Paper, 2024, <https://www.plattform-lernende-systeme.de/files/Downloads/Publikationen_EN/ES_2024/ES_AI_in_the_super_election_year_of_2024.pdf>, accessed 17 April 2025
Bird, W. and Smith, T., ‘Real411 is ready to process your complaints about election misinformation and disinformation’, Daily Maverick, 5 May 2024, <https://www.dailymaverick.co.za/article/2024-05-05-2024-electionstackling-misinformation-disinformation>, accessed 11 February 2025
Black Box, ‘From print to pixels: Understanding media consumption preferences in Southeast Asia’, 2 August 2024, <https://www.blackbox.com.sg/perspectives/from-print-to-pixels-understanding-media-consumption-preferences-in-southeast-asia>, accessed 17 April 2025
Bodine, A., ‘Semiliterate Ghanaians learn media and information literacy’, DW Akademie, 2 December 2022, <https://akademie.dw.com/en/mil-for-illiterate-ghanaians/a-63955851>, accessed 8 March 2025
Boldkhuyag, P., Batmunkh, O., Tserensodnom, L., lkhagvasuren, G. and Munkhtugs, O., ‘Media Freedom Report 2020’, Globe International Center, 3 May 2021, <https://www.gic.mn/public/docs/freedom_report/media_freedom_report_2020_en.pdf>, accessed 16 February 2025
Bontcheva, K., Papadopoulous, S., Tsalakanidou, F., Gallotti, R., Dutkiewicz, L., Krack, N., Teyssou, D., Nucci, F. S., Spangenberg, J., Srba, I., Aichroth, P., Cuccovillo, L. and Verdoliva, L., ‘Generative AI and Disinformation: Recent Advances, Challenges, and Opportunities’, Vera.AI, White Paper, February 2024, <https://edmo.eu/wp-content/uploads/2023/12/Generative-AI-and-Disinformation_-White-Paper-v8.pdf>, accessed 17 April 2025
Bradshaw, S., Campbell-Smith, U., Henle, A., Perini, A., Shalev, S., Bailey, H. and Howard, P. N., Country Case Studies. ‘Industrialized Disinformation: 2020 Global Inventory of Organized Social Media Manipulation’, Oxford Internet Institute, 2021, <https://demtech.oii.ox.ac.uk/wp-content/uploads/sites/12/2021/03/Case-Studies_FINAL.pdf>, accessed 14 February 2025
Bratajaya, Y. and Tan, D., ‘Criminalization of disinformation threatens freedom of expression’, The Jakarta Post, 27 November 2023, <https://www.thejakartapost.com/opinion/2023/11/27/criminalization-of-disinformation-threatens-freedom-of-expression.html>, accessed 13 February 2025
Breuer, A., ‘Information Integrity and Information Pollution: Vulnerabilities and Impact on Social Cohesion and Democracy in Mexico’, German Institute of Development and Sustainability, IDOS Discussion Paper 2/2024, <https://www.idos-research.de/uploads/media/DP_2.2024.pdf>, accessed 8 April 2025
Brooks, T., Princess, G., Heatley, J., J. J., Kim, S., M. S., Parks, S., Reardon, M., Rohrbacher, H., Sahin, B., S. S., J. S., O. T. and R. V., ‘Increasing Threat of Deepfake Identities’, United States Department of Homeland Security, 2021, <https://www.dhs.gov/sites/default/files/publications/increasing_threats_of_deepfake_identities_0.pdf>, accessed 17 April 2025
Business Standard, The, ‘“Digital Security Act succeeds in creating culture of fear”’, 14 January 2023, <https://www.tbsnews.net/bangladesh/digital-security-act-succeeds-creating-culture-fear-speakers-567734>, accessed 7 March 2025
Cerdeira, P., ‘¿Cómo funcionan los ‘Bots’ y cómo saber quiénes están detrás de ellos?’ [How do bots work, and how can we find out who’s behind them?], MVS Noticias, 22 February 2024, <https://mvsnoticias.com/entrevistas/2024/2/22/como-funcionan-los-bots-como-saber-quienes-estan-detras-de-ellos-627813.html>, accessed 15 May 2025
Cetina-Presuel, R., Dutton, W., Eskiadi, I. G., Gauthier, C., Gordon, D., Krebsz, M., Jost, I., Lallana, E. C., Mamattah, R., Orembo, L., Slachmuijlder, L., Santos, N., Sengupta, A., Yu, C. W. and Volkmer, I., ‘2024 Global Election Datafication SafeGuarding: Holistic Collaboration Policies between Digital Platform Regulators and Electoral Authorities’, UNESCO, I4T G20 Policy Brief, September 2024, <https://www.unesco.org/sites/default/files/medias/fichiers/2024/09/t20_brazil_i4t_knowledge_s_policy_brief.pdf?hub=71542>, accessed 17 April 2025
Cleveland, G. and Alexander, W., ‘ANC loses majority after 30 years’, Friedrich Naumann Foundation for Freedom, 3 June 2024, <https://www.freiheit.org/liberal-workshop-south-africa/anc-loses-majority-after-30-years>, accessed 11 February 2025
Conroy-Krutz, J., Amakoh, K. and Amewunou, K., ‘Africa’s Shifting Media Landscapes: Digital Media Use Grows, but so do Demographic Divides’, Afrobarometer, Dispatch No. 800, 30 April 2024, <https://www.afrobarometer.org/wp-content/uploads/2024/04/AD800-PAP14-Digital-media-use-grows-across-Africa%5ELJ-but-so-do-demographic-divides-Afrobarometer-29april24.pdf>, accessed 17 April 2025
Council of Asian Liberals and Democrats (CALD), ‘AI in Elections in East and Southeast Asia: Opportunities, Challenges, and Ways Forward for Democrats and Liberals’, December 2024, <https://www.freiheit.org/sites/default/files/2024-12/cald-ai-policy-paper.pdf>, accessed 17 April 2025
Cowling, N., ‘Distribution of social media users in South Africa as of January 2024, by age group and gender’, Statista, 23 February 2024, <https://www.statista.com/statistics/1100988/age-distribution-of-social-media-users-south-africa>, accessed 11 February 2025
Daily Star, The, ‘Registration of online portals of newspapers, TV channels: TIB warns of govt control over media’, 1 September 2020, <https://www.thedailystar.net/registration-online-versions-newspapers-tv-news-portals-tib-warns-govt-control-over-media-1954193>, accessed 6 March 2025
—, ‘BTRC blocks BSS while weeding illegal news sites out’, 29 September 2021, <https://www.thedailystar.net/news/bangladesh/crime-justice/news/btrc-botches-attempt-comply-hc-order-2186196>, accessed 7 March 2025
Darko, S. A., ‘Ghana’s law on publication of false news is vague and easily abused’, Africa at LSE, 25 February 2022, <https://blogs.lse.ac.uk/africaatlse/2022/02/25/ghana-law-publication-of-fake-news-vague-easily-abused-free-speech>, accessed 8 March 2025
Daudpota, F., ‘Understanding Pakistan’s 2021 social media rules: Suppression of constitutional right to freedom of expression through prior restraint censorship of online content’, last revised 29 November 2021, <http://dx.doi.org/10.2139/ssrn.3597721>
Dawn, ‘Troll armies’, 16 August 2023, <https://www.dawn.com/news/1770418>, accessed 20 February 2025
Dawson, A. and Ball, J., ‘Generating Democracy: AI and the Coming Revolution in Political Communications’, Demos, January 2024, <https://demos.co.uk/wp-content/uploads/2024/01/Generating-Democracy-Report-1.pdf>, accessed 17 April 2025
Del Pozo, C. M. and Arroyo, D. R., ‘Generative AI and Its Influence on Mexico’s 2024 Elections’, Friedrich Naumann Foundation for Freedom, February 2025, <https://www.freiheit.org/publikation/generative-ai-and-its-influence-mexicos-2024-elections-0>, accessed 8 April 2025
Dewantara, J. A., Syamsuri, S., Wandira, A., Afandi, A., Hartati, O., Cahya, N., Sapitri, P. and Nurgiansah, T. H., ‘The role of buzzers in social media in guiding public opinion regarding political choices’, Journal of Etika Demokrasi, 7/3 (2022), pp. 681–94, <https://doi.org/10.26618/jed.v7i3.8103>
DGA Group and Albright Stonebridge Group, ‘AI and Elections: Impact and Implications of AI Generated Content’, August 2024, <https://dgagroup.com/wp-content/uploads/2024/08/AI_and_Elections_Assessing_the_impact_and_implications_of_AI_generated_content_for_democratic_elections_1_.pdf>, accessed 17 April 2025
Dierkes, J., ‘Mongolia’s MPP maintains power with strategic gains’, East Asia Forum, 23 July 2024, <https://doi.org/10.59425/eabc.1721772000>
Digital Literacy Center, [n.d.], <https://digitalliteracy.gov.bd>, accessed 6 March 2025
Digital Watch, ‘South Africa’s proposed misinformation regulations spark debate’, 10 April 2024, <https://wp.dig.watch/updates/south-africas-proposed-misinformation-regulations-spark-debate>, accessed 11 February 2025
Dionne, K. Y. and Liu, S., ‘Understanding the fake news about South Africa’s elections’, Good Authority, 27 May 2024, <https://goodauthority.org/news/disinformation-misinformation-fake-news-about-south-africas-2024-elections>, accessed 11 February 2025
Djajadiningrat, V. P., ‘Indonesia digital policy against distribution of information disorder: A better policy for sustainable society’, ES3 Web of Conferences, 426 (2023), <https://doi.org/10.1051/e3sconf/202342602029>
Domingo, E., Muscat, S., Arnold, S., Salzinger, M. and Chetty, P., ‘The Geopolitics of Digital Literacy and Skills Cooperation with Africa’, European Centre for Development Policy Management, Discussion Paper No. 369, June 2024, <https://ecdpm.org/application/files/8717/1921/9245/The-geopolitics-digital-literacy-skills-coorperation-with-Africa-ECDPM-Discussion-Paper-369-2024.pdf>, accessed 11 February 2025
Duffy, K., ‘AI in context: Indonesian elections challenge GenAI policies’, Council on Foreign Relations, 13 February 2024, <https://www.cfr.org/blog/ai-context-indonesian-elections-challenge-genai-policies>, accessed 17 April 2025
Duke Reporters’ Lab, Fact-checking sites around the world, [n.d.], <https://reporterslab.org/fact-checking>, accessed 16 May 2025
Dunn, S., ‘Women, not politicians, are targeted most often by deepfake videos’, Centre for International Governance Innovation, 3 March 2021, <https://www.cigionline.org/articles/women-not-politicians-are-targeted-most-often-deepfake-videos>, accessed 17 April 2025
Echeverría, M., ‘Antecedent conditions of disinformation in Mexico: A theoretical framework’, Proceedings of the International Conference on Communication and Applied Technologies (2023), pp. 332–43, <https://doi.org/10.2991/978-94-6463-254-5_33>
Economist, The, ‘Tiktok is a key battleground in Indonesia’s election’, 1 February 2024, <https://www.economist.com/asia/2024/02/01/tiktok-is-a-key-battleground-in-indonesias-election>, accessed 15 February 2025
Eddymurthy, I. A. and Sihombing, A. K., ‘Legal update: Indonesia takes big step forward with new revision of electronic information and transactions law’, Lexology, 17 January 2024, <https://www.lexology.com/library/detail.aspx?g=ed376f51-5ed6-4610-8cef-150c61651bbf>, accessed 7 April 2025
Electoral Commission of South Africa (IEC), ‘Electoral Commission partners with social media giants to combat disinformation in 2024 national and provincial elections’, 3 July 2023, <https://www.elections.org.za/pw/News-And-Media/News-List/News/News-Article/Electoral-Commission-partners-with-social-media-giants-to-combat-disinformation-in-2024-National-and-Provincial-Elections?a=AISDGvpz75ps1usOfX7oimHCQG6/AToNAzCQK374oSg=>, accessed 11 February 2025
European Union Election Expert Mission, ‘South Africa, National and Provincial Elections, 29 May 2024, Final Report’, June 2024, <https://www.eods.eu/library/EU%20EEM%20ZAF%202024%20FR.pdf>, accessed 11 February 2025
Fakultas Hukum Universitas Indonesia, ‘Negative campaigns and black campaigns: What’s the difference?’, 10 October 2024, <https://law.ui.ac.id/negative-campaigns-and-black-campaigns-whats-the-difference>, accessed 7 April 2025
Ferrara, E., ‘Charting the Landscape of Nefarious Uses of Generative Artificial Intelligence for Online Election Interference’, University of Southern California, HUMANS Lab, Working Paper No. 2024.1, <https://doi.org/10.2139/ssrn.4883403>
Firdaus, A., ‘Fake Suharto video fuels debate on AI use in Indonesian election campaign’, Benar News, 12 January 2024, <https://www.benarnews.org/english/news/indonesian/suharto-deepfake-used-in-election-campaign-01122024135217.html>, accessed 13 February 2025
Freedom House, ‘Freedom on the Net 2023: Bangladesh’, 2023a, <https://freedomhouse.org/country/bangladesh/freedom-net/2023>, accessed 4 March 2025
—, ‘Freedom on the Net 2023: Indonesia’, 2023b, <https://freedomhouse.org/country/indonesia/freedom-net/2023>, accessed 7 April 2025
Freedom Network, ‘Countering disinformation efforts need urgent boost in Pakistan’, 13 December 2022, <https://www.fnpk.org/countering-disinformation-efforts-need-urgent-boost-in-pakistan>, accessed 19 February 2025
Gamboa, V., ‘Nace “iXóchitl”, vocera especial de la precandidate, creada con inteligencia artificial’ [Introducing ‘iXóchitl’: An AI-generated spokesperson for the presidential pre-candidate], El Universal, 17 December 2023, <https://www.eluniversal.com.mx/elecciones/nace-ixochitl-vocera-especial-de-la-precandidata-creada-con-inteligencia-artificial>, accessed 15 May 2025
Ghana Fact-Checking Coalition, ‘Countering Electoral Disinformation: Lessons from Ghana’s 2024 Elections’, January 2025, <https://www.fact-checkghana.com/wp-content/uploads/2025/02/GFC-Report-Brochure-Final-View-1.pdf>, accessed 6 March 2025
Ghana News Agency, ‘Elections 2024: UNESCO trains journalists in countering hate speech, misinformation on social media’, BusinessGhana, 30 June 2024, <https://www.businessghana.com/site/news/general/309430/Elections-2024%3A-UNESCO-trains-journalists-in-countering-hate-speech%2C-misinformation-on-social-media>, accessed 8 March 2025
Grange, H., ‘Combatting fake news and disinformation is a civil duty’, Good Governance Africa, 19 January 2024, <https://gga.org/combatting-fake-news-and-disinformation-is-a-civil-duty>, accessed 11 February 2025
Grohmann, R. and Ong, J. C., ‘Disinformation-for-hire as everyday digital labor: Introduction to the special issue’, Social Media + Society, 10/1 (2024), <https://doi.org/10.1177/20563051231224723>
Gul, A., ‘Official resigns, admits tampering in Pakistan’s controversy-marred vote’, Voice of America, 17 February 2024, <https://www.voanews.com/a/official-resigns-admits-tampering-in-pakistan-s-controversy-marred-vote-/7491919.html>, accessed 20 February 2025
Gumede, N., ‘Trust deficit drives youth voter apathy’, Africa Practice, 13 March 2024, <https://africapractice.com/trust-deficit-drives-youth-voter-apathy>, accessed 11 February 2025
Hanafin, N., Idrizi, Z. and Fischer, J., Information Integrity for Electoral Institutions and Processes: Reference Manual for UNDP Practitioners (Oslo, Norway: United Nations Development Programme Global Policy Centre for Governance, 2024), <https://www.undp.org/policy-centre/governance/publications/information-integrity-electoral-institutions-and-processes-reference-manual-undp-practitioners>, accessed 17 April 2025
Harmadi, S. H. B, ‘2024 voter demographics’, Kompas, 28 November 2023, <https://www.kompas.id/baca/english/2023/11/27/en-demografi-pemilih-2024>, accessed 13 February 2025
Haskins, C., ‘Phony X accounts are meddling in Ghana’s election’, Rest of World, 15 November 2024, <https://restofworld.org/2024/ghana-election-ai-bots-x-twitter>, accessed 6 March 2025
Hendrix, J. and Jahangir, R., ‘Pakistan and the intersection of tech & elections’, Tech Policy Press, 24 February 2024, <https://www.techpolicy.press/pakistan-and-the-intersection-of-tech-elections>, accessed 24 February 2025
Human Rights Watch, ‘Bangladesh: Repeal abusive law used in crackdown on critics’, 1 July 2020, <https://www.hrw.org/news/2020/07/01/bangladesh-repeal-abusive-law-used-crackdown-critics>, accessed 6 March 2025
—, ‘Pakistan: Repeal amendment to draconian cyber law’, 28 February 2022, <https://www.hrw.org/news/2022/02/28/pakistan-repeal-amendment-draconian-cyber-law>, accessed 19 February 2025
Hussain, A., ‘X faces restrictions in Pakistan amid protests over alleged vote rigging’, Al Jazeera, 19 February 2024, <https://www.aljazeera.com/news/2024/2/19/x-faces-restrictions-in-pakistan-amid-protests-over-alleged-vote-rigging>, accessed 21 February 2025
Idé Siddo, B., ‘How Meta is preparing for the 2024 South African elections’, Meta, 12 April 2024, <https://about.fb.com/news/2024/04/how-meta-is-preparing-for-the-2024-south-african-elections>, accessed 11 February 2025
Idris, I., ‘Indonesia’s misinformation program undermines more than it teaches’, Monash University, 28 February 2022, <https://lens.monash.edu/@politics-society/2022/02/28/1384482/indonesias-misinformation-program-undermines-more-than-it-teaches>, accessed 31 February 2025
Illidge, M., ‘Plan to regulate AI and misinformation in South Africa’, My Broadband, 11 May 2024, <https://mybroadband.co.za/news/internet/535839-plan-to-regulate-ai-and-misinformation-in-south-africa.html>, accessed 11 February 2025
Imperva, ‘2024 Bad Bot Report’, 2024, <https://www.imperva.com/resources/resource-library/reports/2024-bad-bot-report-report-ty?lang=EN&asset_id=6912&gated=1>, accessed 15 May 2025
Inforial, ‘2024 election: Bawaslu joins forces with communications ministry to tackle hoaxes’, The Jakarta Post, 16 October 2023, <https://www.thejakartapost.com/adv/2023/10/16/2024-election-bawaslu-joins-forces-with-communications-ministry-to-tackle-hoaxes.html>, accessed 12 February 2025
Information and Communication Technology Division, Government of the People’s Republic of Bangladesh, ‘National Strategy for Artificial Intelligence’, March 2020, <https://ird.gov.bd/sites/default/files/files/ird.portal.gov.bd/miscellaneous_info/46fc550e_bdbb_4ec4_b2ca_1da23e8ccab0/National%20Strategy%20for%20Artificial%20Intellgence%20-%20Bangladesh%20.pdf>, accessed 17 April 2025
Instituto Nacional Electoral (INE), ‘Estrategias’ [Strategies], 2024a, <https://centralelectoral.ine.mx/wp-content/uploads/2024/03/Estrategias_Certeza-2024.pdf>, accessed 16 May 2025
—, ‘Te presentamos a “Inés”, la asistente virtual del INE para combater la desinformación sobre las elecciones 2024’ [Meet ‘Inés’, the INE’s virtual assistant to combat misinformation about the 2024 elections], Central Electoral, 26 April 2024b, <https://centralelectoral.ine.mx/2024/04/26/te-presentamos-a-ines-la-asistente-virtual-del-ine-para-combatir-la-desinfomacion-sobre-las-elecciones-2024>, accessed 17 May 2025
International IDEA, Voter Turnout Database, ‘Ghana’, [n.d.], <https://www.idea.int/data-tools/data/question-country?question_id=9189&country=84&database_theme=293>, accessed 10 April 2025
—, Misión de Acompañamiento Técnico Electoral al Instituto Nacional Electoral de México, Proceso Electoral Federal y los Procesos Electorales Locales Concurrentes 2023–2024, ‘Informe Final’ [Electoral Technical Assistance Mission to the National Electoral Institute of Mexico, 2023–2024 Federal Electoral Process and Concurrent Local Elections, ‘Final Report’], 1 July 2024a, <https://repositoriodocumental.ine.mx/xmlui/handle/123456789/175918>, accessed 17 April 2025
—, Codes of Conduct on the Rise: Fair and Ethical Political Campaigning Online, Policy Brief (Stockholm: International IDEA, 2024b), <https://doi.org/10.31752/idea.2024.23>
International Trade Administration (ITA), Ghana Digital Economy, 20 September 2024, <https://www.trade.gov/country-commercial-guides/ghana-digital-economy>, accessed 6 March 2025
Islam, Z. and Jamil Khan, M., ‘Spin doctors go into overdrive ahead of polls’, The Daily Star, 12 November 2023, <https://www.thedailystar.net/news/bangladesh/politics/news/spin-doctors-go-overdrive-ahead-polls-3467331>, accessed 5 March 2025
Jahangir, R., ‘Life in PTI’s social media bubble’, Dawn, 16 August 2020, <https://www.dawn.com/news/1574529>, accessed 20 February 2025
—, ‘Online rallies, AI script: New threats loom ahead of 2024 polls’, Dawn, 16 January 2024, <https://www.dawn.com/news/1806163/online-rallies-ai-script-new-threats-loom-ahead-of-2024-polls>, accessed 19 February 2025
Jeffrie, N., Bahia, K., Lindsey, D., Ouattara Boni, A.-N., Sibthorpe, C. and Zagdanski, J., ‘The Mobile Gender Gap Report’, GSMA, May 2024, <https://www.gsma.com/r/wp-content/uploads/2024/05/The-Mobile-Gender-Gap-Report-2024.pdf>, accessed 7 April 2025
Jonathan, S., Sanusi, S. A., Akinbola, T. E. and Nwankwo, E. G., ‘Information Ecosystems ahead of Ghana’s 2024 Elections: Assessing Misinformation/Disinformation Threats and Patterns, Voter Competence and Media Readiness’, Centre for Journalism Innovation and Development, 2024, <https://thecjid.org/document/information-ecosystems-ahead-of-ghanas-2024-elections-assessing-misinformation-disinformation-threats-and-patterns-voter-competence-and-media-readiness>, accessed 6 March 2025
Juneja, P., Artificial Intelligence for Electoral Management (Stockholm: International IDEA, 2024), <https://doi.org/10.31752/idea.2024.31>
Juniarto, D., ‘AI in the 2024 Indonesian election: Is having ethical guidelines enough?’, The Jakarta Post, 12 February 2024, <https://www.thejakartapost.com/opinion/2024/02/10/ai-in-the-2024-indonesian-election-is-having-ethical-guidelines-enough.html>, accessed 20 February 2025
Kemp, S., ‘Digital 2023: Indonesia’, DataReportal, 9 February 2023a, <https://datareportal.com/reports/digital-2023-indonesia>, accessed 13 February 2025
—, ‘Digital 2023: Pakistan’, DataReportal, 13 February 2023b, <https://datareportal.com/reports/digital-2023-pakistan>, accessed 13 March 2025
—, ‘Internet Use in 2024’, DataReportal, 31 January 2024a, <https://datareportal.com/reports/digital-2024-deep-dive-the-state-of-internet-adoption>, accessed 11 February 2025
—, ‘Digital 2024: Bangladesh’, DataReportal, 23 February 2024b, <https://datareportal.com/reports/digital-2024-bangladesh>, accessed 4 March 2025
—, ‘Digital 2024: Ghana’, DataReportal, 23 February 2024c, <https://datareportal.com/reports/digital-2024-ghana>, accessed 6 March 2025
—, ‘Digital 2024: South Africa’, DataReportal, 23 February 2024d, <https://datareportal.com/reports/digital-2024-south-africa>, accessed 11 February 2025
Khalid, A., ‘Election illusions: Deepfakes & disinformation in Pakistan’, Digital Rights Foundation, 13 November 2023, <https://digitalrightsfoundation.pk/election-illusions-deepfakes-disinformation-in-pakistan>, accessed 21 February 2025
Khalid, Z., ‘Pakistan’s draft national AI strategy is a hodgepodge of technospeak’, Centre for Strategic and Contemporary Research, 21 June 2023, <https://cscr.pk/explore/themes/politics-governance/pakistans-draft-national-ai-policy-is-a-hodgepodge-of-technospeak>, accessed 11 June 2025
Khan, H., ‘The influence of fake news in Pakistan’s 2024 general election’, Stratheia, 24 January 2024, <https://stratheia.com/the-influence-of-fake-news-in-pakistans-2024-general-election>, accessed 20 February 2025
Khan, N., ‘Ex-PM Khan’s party says “millions” attended virtual rally despite social media disruption’, Arab News Pakistan, last updated 18 December 2023, <https://www.arabnews.pk/node/2427636/pakistan>, accessed 10 June 2025
Kreuser, C. and Gitonga Theuri, J., ‘Contextualising Political Advertising Policy to Political Micro-targeting in the South Africa 2024 Elections’, African Internet Rights Alliance, November 2024, <https://aira.africa/wp-content/uploads/2024/11/Contextualising-Political-Advertising-Policy-to-Political-Micro-Targeting-in-South-Africa-2024-Elections.pdf>, accessed 11 February 2025
Kutanto, H., ‘Peran DJPPI Kominfo Dalam Mencegah Hoax Pemilu 2024 [The Role of DJPPI Kominfo in Preventing Hoaxes in the 2024 Election]’, Kominfo, 7 December 2023, <https://web.archive.org/web/20241208040512/https://djppi.kominfo.go.id/news/peran-djppi-kominfo-dalam-mencegah-hoax-pemilu-2024>, accessed 30 June 2025
Lagos, A., ‘Transparencia México: Recursos para combatir la desinformación durante el proceso electoral [Transparency Mexico: Resources to combat disinformation during the electoral process], Wired, 12 March 2024, <https://es.wired.com/articulos/transparencia-mexico-recursos-para-combatir-la-desinformacion-durante-el-proceso-electoral>, accessed 17 April 2025
LatAm Intersect, ‘How social media took over Latin America’s headlines’, 17 January 2025, <https://latamintersectpr.com/how-social-media-took-over-latin-americas-headlines>, accessed 17 April 2025
Li, E. R., Shultz, B. and Jankowicz, N., ‘Deepfake Pornography Goes to Washington: Measuring the Prevalence of AI-Generated Non-consensual Intimate Imagery Targeting Congress’, American Sunlight Project, 11 December 2024, <https://static1.squarespace.com/static/6612cbdfd9a9ce56ef931004/t/67586997eaec5c6ae3bb5e24/1733847451191/ASP+DFP+Report.pdf>, accessed 17 April 2025
Lkhaajav, B., ‘Mongolia’s political parties showcase candidates for the 2024 parliamentary election’, The Diplomat, 22 May 2024, <https://thediplomat.com/2024/05/mongolias-political-parties-showcase-candidates-for-the-2024-parliamentary-election>, accessed 17 February 2025
Macdonald, G., ‘A Perilous Moment for Bangladesh’s Democracy’, United States Institute of Peace, 4 March 2024, <https://www.usip.org/publications/2024/03/perilous-moment-bangladeshs-democracy>, accessed 5 March 2025
Malik, M., ‘PTI, police face off in several cities amid protests against “rigging’’’, Dawn, 18 February 2024, <https://www.dawn.com/news/1815124/pti-police-face-off-in-several-cities-amid-protests-against-rigging>, accessed 19 February 2025
Mantas, H., ‘The Mongolian Fact Checking Center fights misinformation in a country where the word doesn’t exist’, Poynter, 10 August 2021, <https://www.poynter.org/fact-checking/2021/the-mongolian-fact-check-center-fights-misinformation-in-a-country-where-the-word-doesnt-exist>, accessed 16 February 2025
Mao, F., ‘Prabowo Subianto: The tainted ex-military chief who will be Indonesia’s new leader’, BBC, 15 February 2024, <https://www.bbc.com/news/world-asia-68237141>, accessed 15 February 2025
Marchant de Abreu, C., ‘Artificial intelligence and deepfakes take over Pakistan’s elections’, France 24, 9 February 2024, <https://www.france24.com/en/tv-shows/truth-or-fake/20240208-artificial-intelligence-and-deepfakes-takeover-pakistan-elections>, accessed 22 March 2025
Martínez Chacón, M., ‘Expertos ven una “avalancha” de noticias falsas en futura elección presidencial de México’ [Experts see an ‘avalanche’ of fake news in Mexico’s upcoming presidential election], Los Angeles Times, 22 August 2023, <https://www.latimes.com/espanol/mexico/articulo/2023-08-22/expertos-ven-una-avalancha-de-noticias-falsas-en-futura-eleccion-presidencial-de-mexico>, accessed 15 May 2025
Media Monitoring Africa, ‘Digital literacy project Web Rangers launches WhatsApp channel for parents’, Press Council of South Africa, 5 September 2024, <https://presscouncil.org.za/2024/09/05/web-rangers-media-literacy-for-children>, accessed 9 February 2025
Mediana, ‘Campaigns on social media: Buzzers and influencers are increasingly “fluid’’’, Kompas, 27 December 2023, <https://www.kompas.id/baca/english/2023/12/26/en-kampanye-di-media-sosial-sosok-pendengung-dan-pemengaruh-semakin-cair>, accessed 14 February 2025
Meedan, ‘Meedan and Nest Center launch Mongolia’s first fact-checking tipline’, 17 May 2024, <https://meedan.com/post/meedan-and-nest-center-launch-mongolias-first-fact-checking-tipline>, accessed 19 February 2025
Méndez, C., ‘Difunden “fotografía íntima” de Mariana Rodríguez creada con IA; alista denuncia’ [‘Intimate photograph’ of Mariana Rodríguez created with AI released; ready to file a complaint], El Heraldo de México, 22 January 2024, <https://heraldodemexico.com.mx/elecciones/2024/1/22/difunden-fotografia-intima-de-mariana-rodriguez-creada-con-ia-alista-denuncia-571656.html>, accessed 17 April 2025
Mgwili-Sibanda, F., ‘TikTok reinforces commitment to platform integrity with key initiatives ahead of Ghana elections’, TikTok Newsroom, 26 November 2024, <https://newsroom.tiktok.com/en-africa/ghana-elections>, accessed 8 March 2025
Mir, A. and Siddiqui, N., ‘Losing Facts to Fiction: Nationalism, Misinformation and Conspiracy Theories in Pakistan’, United States Institute of Peace, 17 November 2022, <https://www.usip.org/publications/2022/11/losing-facts-fiction-nationalism-misinformation-and-conspiracy-theories>, accessed 21 February 2025
Modern Ghana, ‘Ghana Fact-Checking Coalition records over 100 false publications’, 12 December 2024, <https://www.modernghana.com/news/1365129/election2024-ghana-fact-checking-coalition-recor.html>, accessed 6 March 2025
Naciones Unidas México, ‘La desinformación se ha convertido en una industria y puede causar más muertes que la COVID-19: experto’ [Disinformation has become an industry and could cause more deaths than COVID-19: expert], 16 July 2021, <https://coronavirus.onu.org.mx/la-desinformacion-se-ha-convertido-en-una-industria-y-puede-causar-mas-muertes-que-la-covid-19-experto>, accessed 17 May 2025
Naeem, W. and Rehmat, A., ‘Countering Disinformation in Pakistan: Lessons and Recommendations for Digital Journalism’, Freedom Network and DigiMAP, December 2022, <https://www.fnpk.org/wp-content/uploads/2022/12/Countering-Disinformation-Report-1.pdf>, accessed 30 July 2025
Naranjargal, K., ‘Internet freedom in Mongolia’, FOCUS, 91 (2018), <https://www.hurights.or.jp/archives/focus/section3/2018/03/internet-freedom-in-mongolia.html>, accessed 16 February 2025
National Statistics Office of Mongolia (NSOM), United Nations Children’s Fund (UNICEF) and United Nations Population Fund (UNFPA), ‘Mongolia Social Indicator Sample Survey-2023: Key Indicators Report’, July 2024, <https://mics.unicef.org/sites/mics/files/2024-08/Mongolia%202023%20SISS-MICS%20Key%20Indicators%20Report_English.pdf>, accessed 16 February 2025
Ndhlovu, L., ‘Combating disinformation in South Africa’s elections’, Development and Cooperation, 14 October 2024, <https://www.dandc.eu/en/article/during-recent-elections-south-africa-disinformation-spread-rapidly-discrediting-political>, accessed 11 February 2025
Ndlovu, N., ‘AI deepfakes and SA’s fight to protect the 2024 election’, Tech Central, 30 January 2024, <https://techcentral.co.za/ai-deepfakes-fight-protect-2024-election/238743>, accessed 11 February 2025
Newman, N., Fletcher, R., Robertson, C. T., Arguedas, A. R., and Nielsen, R. K., ‘Reuters Institute Digital News Report 2024’, Reuters Institute for the Study of Journalism, 2024, <https://reutersinstitute.politics.ox.ac.uk/sites/default/files/2024-06/RISJ_DNR_2024_Digital_v10%20lr.pdf>, accessed 11 February 2025
Ngubane, S., ‘Countering disinformation in South Africa’s elections: The role of social media platforms’, Republic of South Africa, Government Communication and Information System, [n.d.], <https://www.gcis.gov.za/gcis.gov.za/countering-disinformation-in-south-africa-s-elections-the-role-of-social-media-platforms>, accessed 11 February 2025
Nurhayati-Wolff, H., ‘Breakdown of social media users by age and gender in Indonesia as of January 2021’, Statista, 30 January 2023, <https://www.statista.com/statistics/997297/indonesia-breakdown-social-media-users-age-gender>, accessed 30 May 2025
Open Net, ‘Open Net and about 40 NGOs push back against UNESCO Internet regulation guideline’, 31 May 2023, <https://www.opennetkorea.org/en/wp/4710>, accessed 17 April 2025
Organization for Security and Co-operation in Europe, Office for Democratic Institutions and Human Rights (OSCE/ODIHR), ‘Mongolia, Parliamentary Elections, 28 June 2024, ODIHR Election Observation Mission, Final Report’, 13 December 2024, <https://www.osce.org/files/f/documents/e/c/583375_2.pdf>, accessed 15 February 2025
Osae-Kwapong, J., ‘Ghana’s election 2024: Technology and the changing nature of political competition’, Centre for International Governance Innovation, 5 December 2024, <https://www.cigionline.org/articles/ghanas-election-2024-technology-and-the-changing-nature-of-political-competition>, accessed 6 March 2025
Pakistan, Islamic Republic of, Ministry of Information Technology and Telecommunication, ‘Draft National Artificial Intelligence Policy’, 2022, <https://moitt.gov.pk/SiteImage/Misc/files/National%20AI%20Policy%20Consultation%20Draft%20V1.pdf>, accessed 17 April 2025
Parliament of Ghana, Cybersecurity Act, 2020 (Act 1038), <https://csdsafrica.org/wp-content/uploads/2021/08/Cybersecurity-Act-2020-Act-1038.pdf>, accessed 6 March 2025
Rabitsch, A. and Calabrese, S., ‘The EU’s Artificial Intelligence Act and Its Impact on Electoral Processes: A Human Rights-Based Approach’, European Partnership for Democracy, September 2024, <https://epd.eu/news-publications/the-eus-artificial-intelligence-act-and-its-impact-on-electoral-processes-a-human-rights-based-approach>, accessed 17 April 2025
Rahman, S., Dip, S., Karim, F. and Naser, S., ‘Whispers amidst the Echo Chamber: Decoding Bangladesh’s 2024 Election Disinformation’, Center for Critical and Qualitative Studies and Earki, January 2024, <https://www.fact-watch.org/whisperers-amidst-the-echo-chamber-decoding-bangladeshs-2024-election-disinformation>, accessed 10 June 2025
Ratcliffe, R., ‘From military leader to “harmless grandpa”: The rebranding of Indonesia’s Prabowo’, The Guardian, 9 January 2024, <https://www.theguardian.com/world/2024/jan/09/indonesia-election-prabowo-subianto-rebranding-kidnapping-accusations>, accessed 7 April 2025
Rayda, N., ‘Indonesia elections 2024: How AI has become a double-edged sword for candidates and election officials’, Channel News Asia, 4 February 2024, <https://www.channelnewsasia.com/asia/ai-disinformation-deepfakes-indonesia-elections-4091296>, accessed 14 February 2025
Reporters Without Borders (RSF), Mongolia, [n.d.], <https://rsf.org/en/country/mongolia>, accessed 10 June 2025
—, ‘Threats to press freedom in Pakistan from hastily adopted laws ahead of elections’, 11 August 2023, <https://rsf.org/en/threats-press-freedom-pakistan-hastily-adopted-laws-ahead-elections>, accessed 22 February 2025
Rezwan, ‘The government of Bangladesh quietly passed the new Cyber Security Act 2023’, Global Voices, 19 September 2023, <https://globalvoices.org/2023/09/19/the-government-of-bangladesh-quietly-passed-the-new-cyber-security-act-2023>, accessed 6 March 2025
Rikkilä Tamang, L., ‘In Mongolia, electoral reform ushers in a record number of women to parliament’, International IDEA, 11 July 2024, <https://www.idea.int/news/mongolia-electoral-reform-ushers-record-number-women-parliament>, accessed 15 February 2025
Ríos Tobar, M., Violencia Política de Género en la Esfera Digital en América Latina [Gender-Based Political Violence in the Digital Sphere in Latin America] (Stockholm: International IDEA, 2024), <https://doi.org/10.31752/idea.2024.93>
Sakura, Y. and Virgiany, M., ‘Ethical guidelines on use of artificial intelligence (AI) in Indonesia’, Herbert Smith Freehills Kramer, 12 February 2024, <https://www.herbertsmithfreehills.com/notes/tmt/2024-02/ethical-guidelines-on-use-of-artificial-intelligence-ai-in-indonesia>, accessed 17 April 2025
Schwieter, C. and Gandhi, M., ‘Vilify, Ridicule, Disinform: Political Communication and Media Trust in the Age of Generative AI’, Friedrich-Naumann-Stiftung für die Freiheit, Policy Paper, February 2024, <https://www.freiheit.org/sites/default/files/2024-03/a4_policy-paper_fimi-ai_en-3-1.pdf>, accessed 17 April 2025
Seddon, S. and Javed, F., ‘Pakistan election: PMLN and PPP reach agreement on coalition government’, BBC, 20 February 2024, <https://www.bbc.com/news/world-asia-68354256>, accessed 21 February 2025
Sguazzin, A., ‘Pro-Russian X accounts turn to touting Zuma’s MK Party’, Tech Central, 2 April 2024, <https://techcentral.co.za/pro-russian-x-accounts-zuma-mk-party/242310>, accessed 11 February 2025
Shah, N. and Majumdar, A., ‘The Making of Misinformed Choice: Digital Technologies in Election Cycles – Insights from Nine Countries in Asia’, Bertelsmann Stiftung, 2024, <https://www.bertelsmann-stiftung.de/fileadmin/files/user_upload/The_Making_of_Misinformed_Choice_Digital_technologies_in_election_cycles._Insights_from_nine_countries_in_Asia.pdf>, accessed 17 April 2025
Shaheen, S., ‘Fake news, escalation, and polarization: Pakistan’s disinformation vulnerabilities’, South Asian Voices, 12 May 2022, <https://southasianvoices.org/fake-news-escalation-and-polarization-pakistans-disinformation-vulnerabilities>, accessed 23 February 2025
Shukla, V., ‘Deepfakes and elections: The risk to women’s political participation’, Tech Policy Press, 29 February 2024, <https://www.techpolicy.press/deepfakes-and-elections-the-risk-to-womens-political-participation>, accessed 17 April 2025
Siberkreasi, ‘Wakil Menteri Komunikasi dan Informatika: Kolaborasi jadi Kunci Penting Wujudkan Masyarakat Indonesia yang Berdaya dan Cakap Digital’ [Deputy Minister of Communication and Informatics: Collaboration is the Key to Realizing an Empowered and Digitally Capable Indonesian Society], 12 August 2023, <https://web.archive.org/web/20230924211756/https://gnld.siberkreasi.id/wakil-menteri-komunikasi-dan-informatika-kolaborasi-jadi-kunci-penting-wujudkan-masyarakat-indonesia-yang-berdaya-dan-cakap-digital>, accessed 20 February 2025
Soon, C. and Quek, S., ‘Safeguarding Elections from Threats Posed by Artificial Intelligence’, Institute of Policy Studies, IPS Working Papers No. 56, August 2024, <https://lkyspp.nus.edu.sg/docs/default-source/ips/ips-working-paper-no-56_safeguarding-elections-from-threats-posed-by-artificial-intelligence.pdf>, accessed 17 April 2025
Statista, ‘Social media in Ghana – Statistics & facts’, 28 June 2024a, <https://www.statista.com/topics/9778/social-media-in-ghana>, accessed 6 March 2025
—, ‘Internet usage in Mexico – Statistics & facts’, 22 November 2024b, <https://www.statista.com/topics/3477/internet-usage-in-mexico>, accessed 10 June 2025
Stockwell, S., Hughes, M., Swatton, P. and Bishop, K., ‘AI-Enabled Influence Operations: The Threat to the UK General Election’, Centre for Emerging Technology and Security, Briefing Paper, 28 May 2024, <https://cetas.turing.ac.uk/publications/ai-enabled-influence-operations-threat-uk-general-election>, accessed 17 April 2025
Tarigan, E., ‘Indonesian presidential hopefuls are trying social media, K-pop to win young voters. Will it work?’, AP News, 8 February 2024, <https://apnews.com/article/indonesia-presidential-election-kpop-young-voters-73227fb99b0ed70f704c9f0f759622ce>, accessed 15 February 2025
Timmerman, A., ‘Indonesia will enforce laws on content moderation with tight response time and harsh fines, documents show’, Rest of World, 16 August 2022, <https://restofworld.org/2022/indonesia-social-media-regulations>, accessed 20 February 2025
Transparency International Bangladesh, ‘Directive to national newspapers and TV channels to re-register respective online portals an example of blueprint for ultimate control of the media; TIB expresses grave concern about people’s right to free flow of information and freedom of opinion’, 1 September 2020, <https://www.ti-bangladesh.org/articles/press-release/6126>, accessed 10 April 2025
Tsedendamba, B., ‘Caught between two giants: How Mongolia’s democracy can counter authoritarian propaganda’, Power 3.0, 25 June 2024, <https://www.power3point0.org/2024/06/25/caught-between-two-giants-how-mongolias-democracy-can-counter-authoritarian-propaganda>, accessed 16 February 2025
Tsolmonbaatar, T., ‘Mongolia: Truly Media provides a new fact-checking program’, DW Akademie, 30 October 2019, <https://akademie.dw.com/en/fact-check-mongolia-a-new-verification-program-with-truly-media/a-51054358>, accessed 19 February 2025
Tuohy-Gaydos, G., ‘Artificial Intelligence (AI) in Action: A Preliminary Review of AI Use for Democracy Support’, Westminster Foundation for Democracy, Policy Paper, September 2024, <https://www.wfd.org/sites/default/files/2024-09/wfd_2024_ai_in_action_-_final.pdf>, accessed 17 April 2025
United Nations, ‘Information Integrity on Digital Platforms’, Our Common Agenda, Policy Brief 8, June 2023, <https://www.un.org/sites/un2.un.org/files/our-common-agenda-policy-brief-information-integrity-en.pdf>, accessed 2 July 2025
United Nations Ghana, ‘Countering hate speech and misinformation- UNESCO kickstarts regional trainings for journalist ahead of 2024 elections’, 21 June 2024, <https://ghana.un.org/en/272297-countering-hate-speech-and-misinformation-unesco-kickstarts-regional-trainings-journalist>, accessed 8 March 2025
United States Department of State, ‘Mongolia 2021 Human Rights Report’, [n.d.], <https://www.state.gov/reports/2021-country-reports-on-human-rights-practices/mongolia>, accessed 2 March 2025
Utama, V. R., ‘Will the “gemoy” tactic be effective in wooing Gen Z voters?’, The Jakarta Post, 6 December 2023, <https://www.thejakartapost.com/opinion/2023/12/06/will-the-gemoy-tactic-be-effective-in-wooing-gen-z-voters.html>, accessed 15 February 2025
Valverde, D., ‘TikTok measures against disinformation in Mexico’s 2024 elections’, Mexico Business News, 15 April 2024, <https://mexicobusiness.news/tech/news/tiktok-measures-against-disinformation-mexicos-2024-elections>, accessed 15 April 2025
van Damme, P., ‘Disinformation, governance and the South African election’, ISS African Futures, 19 March 2024, <https://futures.issafrica.org/blog/2024/Disinformation-governance-and-the-South-African-election>, accessed 11 February 2025
Vergara, R., ‘INE y Meta lanzan programa “Soy Digital” para combatir desinformación en las elecciones’ [INE and Meta launch the ‘I Am Digital’ programme to combat disinformation during the elections], Proceso, 9 February 2024, <https://www.proceso.com.mx/nacional/2024/2/9/ine-meta-lanzan-programa-soy-digital-para-combatir-desinformacion-en-las-elecciones-323731.html>, accessed 17 May 2025
Verificado, ‘Cursos y talleres [Courses and workshops], [n.d.], <https://verificado.com.mx/cursos-y-talleres>, accessed 16 May 2025
Verity, J., ‘Data Codebook for a Round 9 Afrobarometer Survey in Ghana’, Afrobarometer, April 2024, <https://www.afrobarometer.org/wp-content/uploads/2024/05/GHA_R9.Codebook_21May24.final_.pdf>, accessed 10 June 2025
Volodya, B., ‘Law on cyber security’, Mondaq, 26 April 2022, <https://www.mondaq.com/security/1186842/law-on-cyber-security>, accessed 19 February 2025
Wasserman, H., Madrid-Morales, D. and Rabin, F., ‘An Assessment of Media Literacy and Fact-Checking Training Needs in South African Schools and Universities’, Africa Check, March 2022, <https://africacheck.org/sites/default/files/media/documents/2022-05/An%20assessment%20of%20media%20literacy%20and%20fact-checking%20training%20needs%20in%20South%20African%20schools%20and%20universities.pdf>, accessed 11 February 2025
Wilson, T. and Starbird, K., ‘Cross-platform disinformation campaigns: Lessons learned and next steps’, Harvard Kennedy School Misinformation Review, 14 January 2020, <https://doi.org/10.37016/mr-2020-002>
Wyrzykowski, R., ‘Rural–urban digital divide still poses a challenge in South Africa’, Opensignal, 13 December 2023, <https://www.opensignal.com/2023/12/13/rural-urban-digital-divide-still-poses-a-challenge-in-south-africa>, accessed 11 February 2025
Yerkes, S. and Hogan, E., ‘Ghana’s presidential contest shows why democracy requires more than well-run elections’, Carnegie Endowment for International Peace, 19 December 2024, <https://carnegieendowment.org/emissary/2024/12/ghana-election-democracy-challenges-lessons?lang=en>, accessed 3 March 2025
Yu, C., ‘How will AI steal our elections?’, Center for Open Science, 2024, <https://doi.org/10.31219/osf.io/un7ev>
Introduction
Before waves of protest triggered the fall of the Awami League (AL) Government in the summer of 2024, Bangladesh held its 12th general election, resulting in the fourth consecutive appointment of incumbent Prime Minister Sheikh Hasina of the AL in January 2024 (Spinelli 2024). The AL’s win was undermined by an election boycott by opposition parties, most notably the coordinating primary opposition party, the Bangladesh Nationalist Party (BNP). The BNP described the 2024 electoral process as a ‘sham vote’, and its resignation effectively removed any competitive forces challenging the AL (Rahman 2023). The BNP’s withdrawal took place in the wake of a government crackdown on opposition movements that resulted in a report by the Civicus Monitor (2024) that downgraded the civic space in Bangladesh to ‘closed’, amid reports that some 25,000 opposition leaders and supporters had been arrested in the lead-up to the elections. In response to the boycott, the AL attempted to create an appearance of electoral legitimacy by using deceptive tactics, such as promoting its own candidates as ‘independent’ and splitting the BNP through coercion (Riaz 2024). Analysts and activists expressed concerns at the lack of transparency and fairness in Bangladesh’s democratic procedures, with the United States Institute of Peace stating that the 2024 election ‘cemented the country’s transition to a “dominant-party” political system’ (Macdonald 2024).
Voter turnout reached a new low in the country’s democratic history, as less than 40 per cent of the 120 million eligible voters cast ballots. Aside from the boycott, the low turnout was likely a product of a series of arson attacks on local buses and trains that prevented voters from reaching polling stations on election day (Karim 2024).
Digital context
As of early 2024, less than half of Bangladesh’s population had access to the Internet, with a penetration rate of 44.5 per cent, slightly below the regional average in South Asia of 51.5 per cent (Kemp 2024b). Most Internet users—equating to 30.4 per cent of the total population—are also active on some form of social media. Social media connectivity is escalating at a rapid pace, with the most popular platform, Facebook, gaining 9.7 million users between early 2023 and the beginning of 2024, an increase of 22.3 per cent. The gender distribution on social media platforms skews towards men, who make up 65.8 per cent of total users (Kemp 2024b). One factor impacting this gender imbalance is the significant mobile gender gap in the country, with notably low levels of mobile ownership, mobile Internet usage and mobile Internet awareness among Bangladeshi women (Jeffrie et al. 2024).
Generally, in-person campaigning remains the preferred method of outreach for the country’s political parties due to the comparatively low levels of Internet penetration. Television is, by a significant margin, the most popular source of news among Bangladesh’s population. Nonetheless, digital forums are increasingly important for presenting a diversity of voices in the political sphere, and social media platforms have taken second place in popularity rankings of news sources (Digitally Right and International Media Support 2021). Social media platforms have become essential for political opposition as government-led persecution and silencing of critics shrink offline civic spaces (Castro 2024).
Online freedom is not unrestricted, however, as the Bangladesh Telecommunication Regulatory Commission regularly monitors, censors and blocks outlets that publish criticism of the government—particularly in response to politically charged incidents (Freedom House 2023a).
The increasing importance of the online information landscape has also opened the floodgates for an abundance of misleading information. Experts documented an uptick in disinformation that coincided with the 2024 election, which was clearly intended to influence public opinion ahead of the polls (Islam and Jamil Khan 2023).
Tactics, techniques and procedures
Coordinated networks of influence
Bangladesh’s Government frequently employs influence networks to regulate and manipulate the online information landscape. The AL has been overt in its adoption of such networks, posting an announcement on its website about the launch of a digital campaign that would employ nearly 8,000 people to disseminate pro-Hasina messaging, debunk ‘disinformation’ and counter propaganda against the party on social media (Bangladesh Awami League 2023). The effort focused primarily on Facebook, where content is frequently shared between various pages and groups. An analysis of online campaigning in Bangladesh showed that only 19 per cent of pages, profiles and groups that propagated pro-AL messages included the party name in their title, the exclusion of which risks obfuscating the affiliations of political advertisers (Baksh et al. 2023). In addition, both the AL and BNP utilized social bot networks for promotion, frequently slandering their political opponents through misleading and deceptive content.
To gain attention and engagement, actors commonly either spammed messaging groups and pages by sending a large number of consecutive messages or fabricated identities or credentials to entice audiences to interact with fake accounts. In such cases, misleading profiles might have falsely claimed to be news agencies or members of the opposition. They also employed bait-and-switch techniques by initially posting images of attractive women to entice users and then shared political disinformation (Islam and Jamil Khan 2023).
Aside from social media, online news outlets have been utilized to artificially boost the spread of content promoting the AL. An investigation found that a ‘network of so-called experts’ (AFP 2023), consisting of 35 names, have collectively published more than 700 articles spread over at least 60 domestic and international news sites since 2022. An analysis of the author profiles shows that they have completely untraceable identities with stolen headshots, fabricated names and erroneous accreditations, often connecting them to internationally prestigious institutions. Articles also attribute made-up quotes that support their agenda to real professionals who have not taken part in interviews or made statements. This type of content is then further inadvertently spread by legitimate news publications that have insufficiently verified the information integrity in the original article, further disseminating deceptive information and effectively laundering disinformation (AFP 2023).
Microtargeting
The AL’s coordinated network of 8,000 activists were specifically trained to customize material for the residents in their neighbourhoods and subsequently disseminate it through local Facebook groups. The campaign strategy was based on an internally conducted nationwide survey which collected data that would help maximize the trustworthiness of campaign material in different regions, effectively personalizing political advertisements in accordance with regional attitudes and beliefs (Bangladesh Awami League 2023).
The seeds for the AL’s electoral social media campaign were planted in 2021, when the party committed to training 100,000 ‘cyber warriors’ that were meant to reach a total of 20 million Facebook users in the country. A reported 660 Facebook pages and groups spread centrally coordinated messages to smaller online communities, many of which were not explicitly affiliated with the party. One additional boost to help with outreach was to feature endorsements from celebrities and public figures, who effectively acted as influencers on social media platforms (Baksh et al. 2023).
Cross-platform amplification
Pro-AL actors have created centrally coordinated hubs on Facebook where they publish disinformation that is simultaneously cross-posted to a network of nearly identical groups and pages. These pages often deceptively imitate news sources to gain legitimacy but lack any identifiable information, such as the original publisher or the source of the information. Since online news outlets in Bangladesh require a government-sanctioned licence to operate, they often circumvent this requirement by using a foreign IP address. Foreign IP addresses are also commonly used to evade prosecution by domestic law enforcement, as the act of spreading disinformation online is criminalized in the country (Islam and Jamil Khan 2023).
AI-generated content
Although the practice is not yet widespread, the production of AI-generated content has given some actors seeking to exert influence over public perceptions new tools to reiterate narratives. Ahead of the 2024 election, pro-government news outlets and influencers outsourced the production of AI-generated content to be used in electoral campaigns. This content was usually produced outside the country by paid entities, whose services were available for as low as USD 24 a month (Parkin 2023).
AI-generated content used during the election commonly depicted political figures, such as party or government officials, making statements that they did not actually make, which had implications for their trustworthiness, eligibility as a candidate or stance on political matters (Castro 2024).
As of 2024, deepfake images and video were estimated to account for only a fraction of the overall electoral disinformation in Bangladesh. A review of electoral disinformation following the election found that only 1.9 per cent of misleading content was deemed to be AI-generated; manipulating images, posting clickbait thumbnails and taking clips out of context were much more frequent disinformation techniques (Rahman et al. 2024).
Countermeasures
Digital Security Act
In the autumn of 2018, the Bangladeshi authorities adopted the Digital Security Act (DSA), which aimed to stop the spread of disinformation by prohibiting the act of intentionally publishing information that is offensive, false or threatening (section 25) or defamatory (section 29) on any digital platform. Violations of the act carry penalties of imprisonment for a maximum of three years or fines of up to BDT 300,000–500,000 (approximately USD 2,500–4,150) (People’s Republic of Bangladesh 2018).
The act has been harshly criticized by human rights organizations, as the punishment handed down under the act is often disproportionate to the offence and indiscriminately directed at political opponents, journalists and others criticizing the government and the AL (Riaz 2021; Amnesty International 2023). Individuals charged under the DSA are subjected to inhumane treatment, and enforcement is used to discourage free and open discussions on online forums, thus creating a ‘culture of fear’ (The Business Standard 2023). Enforcement has even been directed at minors. One prominent example is that of a 15-year-old who was sent to juvenile detention for allegedly defaming Prime Minister Hasina on Facebook (Human Rights Watch 2020).
In 2023, the DSA was amended and renamed the Cyber Security Act (CSA). Although the amended act lessens incarceration penalties, it has largely been denounced as a repackaging of the DSA. Critics have pointed out that several sections of the CSA infringe freedom of expression and opinion (Rezwan 2023).
Content removal
The Bangladeshi authorities have repeatedly engaged in content removal with the self-proclaimed intent of combating ‘misleading anti-state content’ (Freedom House 2023a). These efforts primarily target online news outlets, which since 2020 have been required to apply for government approval and registration (Human Rights Watch 2020). Censorship is generally carried out through orders issued to domestic service providers, which are legally required to comply.
Social media platforms are also impacted by censorship efforts, as the Bangladesh Telecommunication Regulatory Commission (BTRC) regularly requests the removal of ‘fake’ or ‘provocative’ information. Removals are facilitated by the draft BTRC Regulation for Digital, Social Media and OTT Platforms, introduced in 2021, which requires social media platforms to remove false or misleading material within 72 hours of receipt of a removal order (Bangladesh Telecommunication Regulatory Commission 2021).
Content takedowns are also conducted through platform-integrated reporting systems, such as the YouTube copyright claims system, thereby leveraging mechanisms in bad faith to combat alleged misinformation (Freedom House 2023a). Between August and November of 2023, the Cyber and Special Crime Division of the Bangladeshi police shut down at least 700 groups, profiles and pages on social media for allegedly spreading rumours (Islam and Jamil Khan 2023).
Platform shutdowns
In 2020 the High Court of Bangladesh ordered that all online news outlets obtain state approval and registration—a requirement already applicable to offline news publishers. The stated intent was to prevent ‘misreporting’ (The Daily Star 2020). Following its decision, the High Court ordered shutdowns of all unregistered online sources in 2021 (The Daily Star 2021).
The restriction has been criticized by Transparency International Bangladesh (2020) for stifling the information environment and allowing the ‘total control and intervention of the government over the media’. It has also been credited for a lack of digital media initiatives focusing on politics in the country, suggesting that online outlets shift their focus to less controversial topics to avoid the risk of being blocked by the BTRC (Digitally Right and International Media Support 2021).
Leading up to the 2024 election, several secretaries from the EMB of Bangladesh proposed shutting down Facebook on the day of the election to prevent the spread of misleading information. The matter was discussed within the parliament and in talks with Meta. Ultimately, the service remained available during the electoral period (Prothom Alo 2023).
Digital literacy initiatives
The Bangladeshi Computer Council has established a digital literacy centre under the supervision of the Bangladeshi Government’s Communication and Technology Department. The centre acts as both a research hub and an educational facility, conducting digital literacy surveys of the population and offering online courses for different demographics. The project is currently in its initial stages, formulating a national curriculum based on data gathered from civic volunteers. In general, the only available information on the project is found on the centre’s website, indicating that it is not widely known outside the country. However, the centre claims to have trained 371,650 students at 10,783 schools (Digital Literacy Center n.d.).
AI regulation
In 2020 the Information and Communication Technology (ICT) Division of Bangladesh’s Ministry of Posts, Telecommunications and Information Technology released a non-binding National Strategy for Artificial Intelligence, calling for a positive outlook for ‘embracing [AI] for the digitalization of the nation’ (National Strategy for Artificial Intelligence 2019–2024). The document touches upon AI ethics, albeit only briefly, in the area of data privacy protection. It does not discuss how AI might influence the information landscape or provide any concrete suggestions to address ethical concerns in the legislature.
The Information and Communication Technology Division also presented a road map for future regulatory actions, including the development of an ethical and legal framework for AI, intended for release between 2022 and 2024 (Information and Communication Technology Division 2020). These measures have not yet materialized, and the lack of such regulatory frameworks has been criticized by academics. Critics have noted the need for national and localized guidelines due to the self-contained nature of Bangladeshi communities because of the country’s cultural and linguistic diversity. This diversity inhibits the penetration of general efforts to combat disinformation, thereby allowing disinformation to circulate unhindered within closed communities (Yesmin 2023).
Conclusion
Bangladesh serves as an example illustrating that the tactics, techniques and procedures that drive disinformation often generate greater insecurity in the information environment than mere inaccuracies in content. The impact of these deceptive tactics becomes evident as they are employed to influence public opinion even in cases where information is not necessarily false or misleading.
The microtargeting strategies used by the AL’s coordinated network of online activists and the use of cross-posting to Facebook pages that do not disclose their affiliation are examples of how actors can manipulate public perceptions without spreading outright falsehoods. In addition, it is often hard to discern what is true and what is false, as misleading information may be based on exaggeration, omission or overemphasis.
Disinformation is often intertwined with factual claims during campaigns, making it appear credible, thereby allowing it to spread more widely. Ultimately, the blending of truth and falsehood raises the question of where to draw the line between legitimate information, misinformation and disinformation.
In Bangladesh, the blurred boundaries between legitimate information and disinformation are leveraged to justify ‘campaigns against disinformation’. These campaigns violate the rights to freedom of speech and opinion through legislation such as the DSA and CSA. The authorities use these laws to stamp out dissent under the guise of preventing the dissemination of ‘falsehoods’ or ‘disruptions of public order’. Government attempts to control the media and the information landscape thereby directly restrict the availability, diversity and trustworthiness of information in ways that interfere considerably with the quality of democratic processes and the integrity of elections.
Introduction
Ghana’s presidential and parliamentary elections held on 7 December 2024 saw the return to power of John Dramani Mahama, who had previously served as President from 2012 to 2017. The elections featured a competitive contest between the incumbent New Patriotic Party (NPP), led by Vice President Mahamudu Bawumia, and the opposition National Democratic Congress (NDC), headed by Mahama (Yerkes and Hogan 2024). The campaign was marked by debates over economic recovery, corruption and governance, with both parties seeking to address the concerns of an electorate increasingly frustrated with economic stagnation and high unemployment (Osae-Kwapong 2024b; Electoral Commission of Ghana 2024).
Voter turnout stood at 60.88 per cent, a sizeable decline from the 78.89 per cent recorded in the 2020 elections (International IDEA n.d.a). Analysts attributed this drop to growing voter apathy, particularly among young people, who feel disillusioned with the political leadership and the lack of substantial policy changes (Verity 2024; International IDEA n.d.d). Additionally, concerns over electoral integrity and the spread of disinformation—particularly false claims about voter suppression, rigged results and candidate misconduct—had a significant negative impact on public perceptions of the election process (Penplusbytes 2024; MFWA 2025).
Ghana has long been regarded as one of Africa’s strongest democracies, with a history of peaceful transitions of power. However, the 2024 elections underscored persistent challenges, including disinformation campaigns aimed at influencing voter behaviour and undermining trust in electoral institutions. Reports from CSOs and international observers highlighted the need for stronger electoral safeguards and digital literacy initiatives to counter the effects of misinformation (Boakye, Bchir and Missoh 2024; African Union Election Observer Mission 2024). Addressing these challenges will be critical to maintaining the credibility of Ghana’s democratic institutions and ensuring broader voter engagement in future elections.
Digital context
Ghana’s digital landscape has experienced significant growth, reflecting a broader trend of increasing Internet and mobile connectivity across the country. The number of Internet users has reached 24.06 million, representing an Internet utilization rate of 69.8 per cent (Kemp 2024c). This widespread connectivity has been facilitated by collaborative efforts between the government and private sector to provide near-universal broadband coverage (ITA 2024).
Mobile connectivity continues to be the principal driving force behind Ghana’s digital adoption, with the mobile phone use rate exceeding 135 per cent (ITA 2024), an indication that many Ghanaians own multiple mobile devices. In addition, social media usage has also seen a notable increase. There were 7.4 million social media users in Ghana in January 2024, up from 6.6 million the previous year (Kemp 2024c). While WhatsApp and Facebook remain the most dominant platforms, others—including Instagram and LinkedIn—have also gained traction (Kemp 2024c; Statista 2024a). The majority of social media users are between the ages of 18 and 34, highlighting the youth-driven nature of digital engagement in the country.
However, disparities in digital access persist along gender and geographic lines. Women are significantly less likely than men to own smartphones or access digital platforms, limiting their participation in online discourse and their economic opportunities (Verity 2024; Statista 2024a). Similarly, urban areas enjoy better Internet infrastructure and higher connectivity rates, while rural regions face challenges such as limited broadband access (ITA 2024; Statista 2024a). These disparities affect the ability of citizens to access electoral information and participate in civic activities.
The Ghanaian Government has recognized the importance of digital transformation and has implemented policies to support the growth of the digital economy. Initiatives such as the Digital Economy Policy and Strategy aim to address challenges and leverage opportunities within the digital landscape (Republic of Ghana 2024b). Additionally, forums like the Ghana Internet Governance Forum focus on fostering a secure, inclusive and sustainable digital economy (Republic of Ghana 2024a).
Ghana has also implemented measures to regulate the online space and combat misinformation, particularly in the context of elections. The Cyber Security Authority, under the Ministry of Communications and Digitalisation, oversees digital security efforts, including policies aimed at addressing online misinformation (Ghana Cyber Security Authority n.d.). The country’s Cybersecurity Act of 2020 provides a legal framework for monitoring online activity and prosecuting cyber-related offences, including the dissemination of false information (Parliament of Ghana 2020). However, CSOs have raised concerns that these regulations, while intended to curb disinformation, have been selectively enforced to suppress dissenting voices. Critics argue that broad and vague provisions within the Cybersecurity Act and the Electronic Communications Act (2008) have been used to target journalists, activists and opposition figures under the pretext of combating misinformation (Darko 2022). The targeting of dissenting voices has contributed to a growing ‘culture of silence’, where fear of legal repercussions discourages open political discourse. Overall, Ghanaians are mistrustful of information originating from social media, with 65 per cent believing that users spread false information ‘sometimes or often’ (Osae-Kwapong 2024a).
Disinformation tactics, techniques and procedures
Coordinated networks of influence
During Ghana’s 2024 elections, organized networks disseminated false or misleading narratives to manipulate public opinion. Groups affiliated with political parties leveraged social media, particularly X (formerly Twitter), to spread coordinated messages amplifying claims of electoral fraud and candidate misconduct (MFWA 2025). Mahama, the opposition party candidate, was targeted as ‘corrupt’ and a ‘thief’ by groups that sought to promote the incumbent party candidate, Bawumia. Investigations revealed that some of these networks extended beyond Ghana’s borders, with foreign-run social media accounts amplifying divisive political content. A report published by Rest of World found that automated bot accounts on X pushed thousands of posts using the hashtag #GhanaStolenElections, many of which originated from IP addresses outside Ghana (Haskins 2024).
Ghana’s 2024 elections unfolded in a fiercely competitive political environment where both the ruling NPP and the opposition NDC employed sophisticated digital strategies to shape voter perceptions. The NPP, determined to break Ghana’s historical eight-year cycle of power transitions, and the NDC, leveraging public dissatisfaction with governance, both relied on coordinated online influence campaigns to discredit opponents and reinforce their own narratives (Osae-Kwapong 2024a). One example of this strategy was the deployment of targeted misinformation campaigns on social media. On 25 and 26 November competing ads surfaced on Facebook, each falsely portraying the opposing party’s presidential candidate as pro-LGBTQIA+, despite both candidates having publicly opposed same-sex relationships. These duelling narratives were designed to manipulate voters who prioritized conservative social values, illustrating how disinformation can be weaponized to confuse and polarize the electorate (Osae-Kwapong 2024a).
AI-generated content
The 2024 elections in Ghana saw a significant increase in AI-generated disinformation, which played a critical role in shaping voter perceptions. The Ghana Fact-Checking Coalition identified multiple instances where generative AI was used to create misleading audio recordings and synthetic media. In one case, an AI-generated audio clip falsely claimed that Mahama, the opposition candidate, was planning to rig the elections, while another fabricated a recording of Vice President Bawumia allegedly calling Ghanaians ‘weak minds’ (Ghana Fact-Checking Coalition 2025). Additional falsified recordings purported to show the main presidential candidates engaging in private conversations and instigating their followers to commit violent acts (MFWA 2025). These AI-generated falsehoods circulated widely on WhatsApp and Facebook, creating confusion and fuelling political tensions. AI was also employed to alter images and fabricate videos to misrepresent political events, with deepfakes having a significant presence on YouTube (Jonathan et al. 2024). These tactics illustrate how AI-driven disinformation was used to manipulate narratives, erode trust in institutions and polarize the electorate.
Manipulating political narratives and framing through social media
One of the most prevalent disinformation tactics in Ghana’s 2024 elections was the manipulation of framing through the use of recycled news items taken out of context. Political actors and disinformation agents frequently repurposed old news articles, images and videos, presenting them as current events to mislead voters and fuel political tensions. Recycled images from past elections were digitally altered and reintroduced as new evidence of alleged voting irregularities (Jonathan et al. 2024). This tactic was particularly effective because it relied on legitimate sources, making it more difficult for the public to immediately identify the deception (Ghana Fact-Checking Coalition 2025).
This form of disinformation underscores the challenges of verifying online content in a fast-paced digital environment. While fact-checking organizations worked to debunk misleading claims, the widespread use of recycled news items highlighted the urgent need for greater media literacy and improved digital verification tools to prevent misinformation from influencing voter behaviour (Kportufe 2025).
Social media influencers as disinformation agents
Social media influencers played a critical role in the dissemination of disinformation during Ghana’s 2024 elections, leveraging their large followings to spread misleading narratives. Analysis of disinformation trends from 2021 to 2024 shows that influencers accounted for 25 per cent of false or misleading content, second only to political figures, who contributed 40 per cent (Jonathan et al. 2024). In their 2025 report, the Ghana Fact-Checking Coalition also found that social media influencers were significant purveyors of disinformation and played a major role in shaping partisan narratives and manipulating the public’s perception of reality. The influence of these figures exists in a grey area between paid political advertising and organic discourse. Some were compensated by political parties, while others amplified disinformation for ideological or personal gain (Jonathan et al. 2024). Given that politicians were also primary sources of falsehoods—responsible for 40 per cent of misleading content—the challenge of accountability remains significant.
Weaponization of ethnicity and LGBTQIA+ issues
The 2024 Ghanaian elections witnessed the strategic deployment of ethnic and LGBTQIA+ issues as tools for political manipulation. Both major political parties, the NPP and the NDC, engaged in tactics that exploited social and cultural sensitivities to sway voter perceptions and deepen political divisions. Disinformation campaigns framed these issues in ways that resonated with existing societal biases, intensifying polarization and shaping electoral narratives (Jonathan et al. 2024).
Ethnic identity has long played a role in Ghanaian politics, but the 2024 elections saw an increase in disinformation tactics that targeted ethnic sentiments. Misinformation campaigns sought to manipulate ethnic allegiances, particularly between the Akan—Ghana’s largest ethnic group, concentrated in the Ashanti and Eastern Regions and historically aligned with the NPP—and the Ewe, who are primarily based in the Volta Region and have traditionally supported the NDC (LaMonica 2024). One notable instance involved a flyer that falsely attributed a divisive ethnic remark to Mahama that was strategically disseminated across social media applications with the likely aim of alienating Akan voters from the NDC (Jonathan et al. 2024).
In addition, such false claims were further amplified by local-language radio stations. A notable case involved a presenter on a radio station owned by a prominent NPP executive who falsely informed his audience that voters supporting Mahama and other opposition candidates would cast their votes on a separate day after the official election date of 7 December (Prempeh 2024). This misinformation, delivered in a local dialect, was an attempt to suppress opposition turnout. The broadcast was later removed from the station’s social media accounts after authorities arrested the presenter for the ‘publication of false news’ (Prempeh 2024).
LGBTQIA+ issues were also manipulated in campaign messaging, with candidates and their supporters framing opponents as sympathetic to LGBTQIA+ rights—a politically sensitive issue in Ghana. Disinformation campaigns falsely claimed that both Mahama and Bawumia secretly supported LGBTQIA+ rights despite their public opposition (Osae-Kwapong 2024a). Throughout the campaign, political parties sought to use this issue as an electoral weapon to mobilize conservative voters. The ruling NPP’s candidate, Bawumia, publicly committed to signing a proposed anti-LGBTQIA+ bill into law, while his opponent, Mahama, accused Western countries of attempting to impose LGBTQIA+ rights on Ghana (Thomas Reuters Foundation 2024). These statements reinforced a climate where misinformation about LGBTQIA+ policies thrived, with fabricated claims circulating on social media while politicians led rallies motivated by homophobic chants (Ghana Fact-Checking Coalition 2025).
Countermeasures
Fact-checking initiatives and engagement of CSOs
To combat electoral disinformation, several CSO-led fact-checking coalitions intensified their monitoring efforts during the election period. The Ghana Fact-Checking Coalition, which operates under the auspices of the Media Foundation for West Africa and is comprised of Fact-Check Ghana, Dubawa Ghana and GhanaFact, verified more than 100 instances of manipulated media, ranging from AI-altered candidate speeches to fabricated images depicting election-related violence (Modern Ghana 2024). The coalition operated real-time fact-checking desks, where journalists and analysts collaborated to debunk falsehoods before they gained traction online (Africa Check 2024). Furthermore, the Centre for Journalism Innovation and Development deployed a team of 19 fact-checkers, observers and open-source intelligence analysts to monitor and combat the spread of disinformation (Ismaila 2024). These efforts demonstrated the potential for impactful civil society–led fact-checking initiatives where different CSOs and volunteers can pool their resources to monitor disinformation and AI-generated content online.
Regulatory measures and government initiatives
Ghana’s Electronic Communications Act (2008) and Cybersecurity Act (2020) penalize the spread of false information, but enforcement has been selective, raising concerns that these laws are being used to suppress legitimate political discourse (Parliament of Ghana 2020; Darko 2022). Critics argue that vague provisions in these laws allow the authorities to criminalize dissent, leading to accusations of state overreach, the suppression of press freedom, and the creation of a culture of silence, where fear of legal repercussions stifles open discourse and investigative journalism (Darko 2022).
Recognizing the threat posed by misinformation, the Electoral Commission of Ghana took proactive measures ahead of the 2024 elections by reinforcing its communication channels, urging media outlets and the public to verify information before sharing (Joy News 2024). In addition, the Cyber Security Authority organized training sessions for media professionals across the country. These sessions aimed to deepen journalists’ understanding of cybersecurity risks, disinformation tactics and fact-checking strategies to prevent the unintentional amplification of false narratives (Abubakar 2024).
Technology- and platform-based interventions
Social media platforms implemented fact-checking, content moderation and media literacy initiatives to curb misinformation during Ghana’s 2024 elections. TikTok launched an in-app election centre with the Electoral Commission of Ghana and Dubawa, providing authoritative election information in English and Twi (Mgwili-Sibanda 2024). To counter AI-generated disinformation, TikTok provided media literacy videos, introduced automatic labelling for synthetic media and partnered with Code for Africa for fact-checking. Election-related content was labelled and linked to verified sources, while AI-driven detection tools helped to flag and remove misinformation before it was reported by users.
Digital literacy initiatives
Recognizing the threat of disinformation, several organizations launched media and information literacy initiatives to equip Ghanaians with the skills to identify and resist false narratives. Penplusbytes’ Market Roadshow brought media literacy training directly to semi-literate and illiterate populations in marketplaces and bus stations through interactive radio dramas in multiple languages, demonstrating common misinformation tactics and providing simple verification steps via flyers (Bodine 2022). CSOs, including the Media Foundation for West Africa and Penplusbytes, expanded these efforts by implementing community workshops, digital literacy programmes and media awareness campaigns to help voters counter disinformation (ITA 2024; Kwafo 2024; Penplusbytes 2024; MFWA 2025). Other organizations, such as TikTok, also contributed to digital literacy programmes (Mgwili-Sibanda 2024; Ghana Fact-Checking Coalition 2025). Additionally, the UNESCO Regional Office in Ghana, in collaboration with local partners, launched training programmes for journalists, equipping them with tools to detect and combat election-related misinformation (Ghana News Agency 2024; United Nations Ghana 2024).
Despite progress on media literacy, challenges persist. Rural communities and economically disadvantaged groups remain vulnerable due to limited Internet access and lower levels of digital literacy (ITA 2024). While broadband coverage has expanded, high costs and poor service quality continue to hinder equitable access to reliable information.
Conclusion
Ghana’s 2024 general elections highlighted both the resilience of the country’s democratic institutions and the significant challenges posed by electoral disinformation. The widespread use of AI-generated content, social media manipulation and the weaponization of identity politics demonstrated the evolving nature of misinformation campaigns in elections. While fact-checking initiatives, civil society engagement and regulatory measures sought to counter these threats, concerns regarding digital literacy and legal enforcement endure. Moving forward, strengthening media literacy, enhancing technological interventions and ensuring transparent enforcement of misinformation regulations will be critical in safeguarding the integrity of future elections.
Introduction
Over 200 million eligible voters cast their ballot in Indonesia’s general election on 14 February 2024, resulting in the election of then–Minister of Defence Prabowo Subianto, of the Gerindra party, as president with 58.6 per cent of the popular vote (Le Monde 2024). Prabowo has been a polarizing figure since he acknowledged responsibility for the kidnapping and torture of democracy activists. He was dishonourably discharged from military service in 1998 and went into self-imposed exile before re-entering Indonesian politics in 2004 (Watmough 2024). After several failed presidential campaigns in previous election periods, he attempted to introduce a new, softer political persona, distancing himself from his military background. This rebranding came under scrutiny for potentially misleading voters who were unaware of his background (Mao 2024). Prabowo’s election was aided by an endorsement from incumbent President Joko Widodo, who held a high approval rating. Perceptions of Widodo’s support were reinforced when his son, Gibran Rakabuming Raka, controversially joined Prabowo’s campaign as his running mate, a move viewed by some observers as fraudulent due to the undue influence Gibran’s uncle, Anwar Usman, who was the chief justice of Indonesia's Constitutional Court (Fikri 2023; Harbowo 2024; Maulia 2024).
Indonesia’s 2024 election had a voter turnout of 82.39 per cent, just below the average regional rate of 84 per cent across South East Asia (International IDEA n.d.b). A distinctive characteristic of the Indonesian electorate is the dominance of younger demographics, with 53 per cent of eligible voters being born between 1981 and 2007 (Harmadi 2023). The high proportion of younger voters contributed to the effectiveness of Prabowo’s attempts to depoliticize his image, as most of these voters may not remember or had not yet been born during Prabowo’s time as a military general.
Digital context
Indonesia has 185.3 million Internet users, accounting for 66.5 per cent of the total population; 75 per cent of all Internet users are active on some type of social media. The gender split between users is fairly even, with 46.5 per cent of users identifying as women and 53.5 per cent as men (Kemp 2023a). The differences between age groups are much greater, as 77.3 per cent of social media users are between the ages of 13 and 34 (Nurhayati-Wolff 2023).
Young voters have played a central role in shaping Indonesia’s contemporary electoral information landscape. As newer generations that are more proficient with digital technologies grow, social media has become one of the main sources of electoral information. Today, the social media platforms TikTok, YouTube and Instagram are the leading sources of information for political campaigns among Millennials and Generation Z. Aside from television and radio, the platforms hold the top spots even among older voters (Statista n.d.). Heavy reliance on semi-regulated social media platforms makes information systems vulnerable to tactics used by disinformation actors, who have historically had a noticeable impact on the country’s elections (Masduki 2021). The use of emerging technologies is widespread in campaigning, which has quickly adapted to the country’s 185.3 million netizens. AI-powered tools, social media and online pundits are increasingly utilized in the strategies of official and unofficial actors seeking to shape public opinion ahead of an election.
Disinformation tactics, techniques and procedures
Coordinated networks of influence
The phenomenon of political influencers and amplifiers is a well-established facet of Indonesia’s political landscape. The term ‘buzzer’ has circulated in political contexts since 2009 and is used to describe a set of local collaborators that, for differing incentives, team up with actors to produce and disseminate political information and build narratives (Dewantara et al. 2022). They achieve their goals by microtargeting audiences that are then inundated with content, often boosted by AI-generated material, spread by personal networks of fake accounts and social bots (Mediana 2023; Rayda 2024). Collaboration and coordination between buzzers are imperative, as messaging is amplified by consolidating the networks that belong to singular buzzers composed of fake accounts (Rayda 2024).
Over time, the industry has grown and today represents a range spanning from volunteer groups to advertising companies (Mediana 2023). These networks are often responsible for ‘black campaigns’, where political hoaxes and disinformation aimed at slandering political candidates are disseminated through open media channels—tactics that are illegal under Indonesian electoral law (Fakultas Hukum Universitas Indonesia 2024). Black campaigns played a part in the 2024 election cycle as well, but they demonstrated a shift to incorporate newer technologies, such as deepfake images and video that spread hateful and bigoted messages about candidates or falsely accused them of claims they had not made. Buzzers also impacted the success of Prabowo’s gemoy campaign, as they employed their networks to artificially amplify the candidate’s virality (Iannone 2024). The Asian Network for Free Elections (ANFREL 2024a) noted in its final election observation report the ‘pervasiveness’ of disinformation during the election, pointing to the need for a ‘multi-pronged approach to address information disorder’.
Indonesia is one of the largest markets for social media in the world, with more than 153.7 million users aged 18 or above, accounting for 79.5 per cent of eligible voters (Kemp 2023a). How the public engages with social media is shifting, however, as the popularity of short-form video content is increasing, with TikTok growing faster than any other social media platform (Kominfo n.d.).
The expansion into live- and short-form video has opened up new avenues for candidates to interact with voters directly, fostering a perceived intimacy and familiarity with their audience. These perceptions are facilitated through social media streaming, which several candidates have used to communicate directly with voters and answer questions in a casual manner (Tarigan 2024). Another technique is to participate in viral trends, a notable example being a dance based on a circulating video clip of presidential candidate Prabowo (Ratcliffe 2024). These viral moments are seldom created by campaign coordinators; instead, they are co-opted from existing trends, giving them a grassroots quality, which imparts a sense of authenticity to the candidate.
A consistent objective in these depoliticized uses of campaign platforms on the Internet is the courting of younger demographics with lower levels of political sophistication and engagement. Experts and members of the Indonesian public have raised concerns about the pivot away from policy-centric campaigning towards entertainment, as it risks preventing voters from staying informed on matters relevant to the election (Belinda 2023).
In addition, criticism has also been directed at the depoliticization of candidates’ images, as narratives are being displaced from policy to personality, which could mislead voters and affect their choices (The Economist 2024). Analysis in the aftermath of the 2024 general election indicated that these narratives and techniques helped bring about Prabowo’s victory, highlighting their ability to turn negative discussions into positive ones online and energize support from young voters (Sam, Yusuf and Rahayu 2024; Muhtadi and Muslim 2024; Wulandari 2024).
AI-generated content
Deepfakes, AI chatbots and other types of generative content were widely incorporated into electoral campaigning for the 2024 election. In the same manner that candidates employed social media conventions to influence public perceptions, deepfakes were used to depoliticize and curate their public images. All three presidential candidates employed AI technologies in campaigning, but Prabowo’s campaign was most prolific in its use (Lamb, Potkin and Teresia 2024). Prabowo used generative AI to fabricate 3D animated personas of himself and his running mate. His character garnered the nickname gemoy, an Indonesian slang expression meaning ‘adorable’, which Prabowo endorsed in an attempt to distance himself from his past as a military general during the Suharto dictatorship (Utama 2023).
AI was also used to spread overt falsities. A controversial case was a deepfake video that reached 4.2 million views, depicting the former dictator Suharto, who died in 2008, endorsing the Golkar party (Firdaus 2024). Deepfakes were also used to slander political rivals, as several candidates were the targets of audio deepfakes that misled listeners into believing that the subject was speaking in a different language, expressing opinions or taking part in conversations they were not present for (Rayda 2024). Finally, several parties created AI-powered chatbots that were trained to provide responses to voters’ questions on political matters in voices similar to those of each party’s respective presidential candidate. The bots proved to be unreliable sources of information, as they provided answers that were incorrect and misleading, such as inventing principles of Indonesian state philosophy (Rayda 2024).
Countermeasures
Law against disinformation
Indonesia has several laws that were enforced during the election to curb the spread of disinformation. The most commonly applied principles (articles 14 and 15 of Law No. 1 of 1946 concerning the Criminal Law Regulations) entail imprisonment for spreading fake, exaggerated or incomplete information to intentionally cause, or with reasonable suspicion that the information may cause, public unrest (Djajadiningrat 2023).
In early 2024 the Indonesian Government revised the Electronic Information and Transactions Law to include an article with three new clauses that impact the spread of disinformation. Article 28 prohibits the distribution of electronic information containing hoaxes or misinformation that leads to material losses in electronic transactions or that incites hostility against individuals or groups based on identifying factors; it also disallows the intentional dissemination of electronic information that is known to be false which could cause social unrest (Eddymurthy and Sihombing 2024). The enforcement of these laws has come under criticism for restricting freedom of expression and for including a vague definition of ‘disinformation’ that could result in arbitrary application of the law (Bratajaya and Tan 2023).
Government monitoring and mitigation of disinformation
Indonesia’s Ministry of Communication and Digital Affairs (Kominfo) monitored disinformation during the 2024 election. Monitoring was conducted around the clock by a unit that combined human operators and technological tools to identify potential disinformation on social media platforms (Inforial 2023). Kominfo used its website as a platform to bring attention to and fact-check circulating hoaxes aimed at undermining the functioning of state operations and the electoral system (Kutanto 2023). The Ministry also collaborated with the General Election Supervisory Agency (Bawaslu) to coordinate social media monitoring and takedowns related to the 2024 election, with the explicit intent of minimizing polarization in the political climate (Inforial 2023). The Ministry also pointed to civil society–led initiatives, such as one by the Safer Internet Lab, led by the Indonesian think tank Centre for Strategic and International Studies in collaboration with Google Indonesia, as key ‘digital guides’ for countering disinformation (Tenggara Strategies 2023).
Alongside individual checks, these initiatives regularly report the total number of hoaxes within a set period as part of a disinformation overview. Leading up to the 2024 election, the deputy minister of Kominfo declared the Ministry’s commitment to taking comprehensive measures to protect Indonesia’s digital space from deceptive content (ANTARA 2024a).
Finally, Bawaslu created a task force together with the Indonesian Broadcasting Commission, the General Elections Commission and the Press Council to monitor the spread of unlawful content on social media in political campaigning surrounding the 2024 election (ANTARA 2023a)
Government relations with social media and content removal
The General Election Supervisory Agency’s efforts also included collaboration with social media platforms ahead of the election to secure information accessibility and promote freedom of speech. An example of such a measure was the correction of an error wherein searches for the vice presidential candidate Mahfud MD on X (formerly Twitter) erroneously yielded no results (ANTARA 2024b).
In 2022 Kominfo stepped up enforcement of Ministerial Regulation No. 5, a law introduced in 2020 that requires social media platforms to self-moderate (Timmerman 2022). Article 5 of the regulation gives platforms 4–24 hours to remove content deemed ‘unlawful’ by government authorities. If they fail to comply, platforms risk fines of up to USD 33,000 per violation. Failures to remove ‘negative’ content from social media have previously resulted in both temporary and long-term shutdowns of public access to websites and applications (Freedom House 2023b).
The Indonesian Government has come under scrutiny for its large number of requests to remove content from social media, as no appeals system is in place. Between 2018 and 2024, Kominfo blocked access to approximately 4.8 million pieces of negative content online (Wiguna 2024).
Concern has also been raised regarding the frequent practice of imposing criminal sanctions following content removal (Mediana 2023). In addition, criticism has been directed at the arbitrary grounds for the government-imposed blocks of Internet platforms, as the regulations that enable monitoring and takedowns are in large part not based on democratic principles and procedures (Freedom House 2023b).
AI ethical guidelines
Kominfo issued a circular in late 2023 that provided guidelines on the ethical aspects of AI use. Part of the guidelines took into consideration the dangers to democracy posed by AI as a tool for spreading disinformation (Indonesia Business Post 2023). Following the circular’s release, questions were voiced regarding the circular’s lack of binding power, meaning the recommendations lacked the power necessary to have a meaningful impact (Juniarto 2024). Along with the release of the circular, Kominfo announced a commitment to prepare binding regulations on AI in the future (ANTARA 2023c).
Digital literacy support
The Indonesian Government has recognized digital literacy as a core pillar in safeguarding the information environment against disinformation. To ensure this bulwark, it played a role in the creation of the National Movement for Digital Literacy (known as Siberkreasi) to strengthen digital capacity and prevent the spread of disinformation. The programme started in 2017 and had enrolled over 24 million citizens as of late 2023. The deputy minister of Kominfo has hailed the project’s collaborative structure as essential for its success. Among its 60 contributors are actors from academia, the private sector (including the owners of social media platforms), governmental bodies and civil society (ITU n.d.; Siberkreasi 2023). Recently, the scope of the curriculum has expanded to include literacy on AI, adapting to recent developments in the information landscape. Kominfo also highlighted the initiative as part of a campaign to combat electoral disinformation surrounding the 2024 election (ANTARA 2023b). However, Siberkreasi has also attracted criticism from independent voices, questioning its learning materials. The criticism has centred on the digital ethics and digital culture segments of the programme, which, according to the critics, put greater emphasis on obedience to the state over critical thinking and media literacy (Idris 2022).
Conclusion
During the past decade, Indonesia has faced challenges in securing free and democratic access to information surrounding elections. Challenges primarily arise due to the dynamic nature of disinformation tactics, which continuously adapt alongside emerging technologies. This evolution is evident from the initial deployment of coordinated networks of influence, dating back to the late 2000s, to the widespread utilization of deepfakes and short-form video content during the 2024 election cycle. Today, the reliance on social media for information is common among the relatively young Indonesian electorate, and political campaigning has shifted to align with this new paradigm. Accordingly, actors seeking to sway public opinion, both overtly and covertly, have turned to digital tools, such as AI and live-streaming, to extend the reach and penetration of their influence. On the other side of the same coin, state regulation of disinformation has been extensive and, according to many critics, disproportionate and undemocratic—often accredited to vague definitions of what constitutes disinformation.
Introduction
In the largest-ever election in the world’s most populous Spanish-speaking country, over 98 million voters were asked to cast their votes for president on 2 June 2024. Beyond selecting a new head of state, Mexicans also voted for 500 members of the Chamber of Deputies, 128 members of the Senate of the Republic, state governors and over 20,000 public service positions across state and local levels. In a presidential race dominated by female candidates, the incumbent party candidate and frontrunner, Claudia Sheinbaum, won 59.4 per cent of the votes (Millan, Kao and Saito 2024). She campaigned on a centre-left platform, pledging to continue the policies of outgoing President Andrés Manuel López Obrador, whose Morena party has been a dominant force in Mexican politics (Millan, Kao and Saito 2024). A former mayor of Mexico City, Sheinbaum is also a climate scientist and former youth activist. Her campaign focused on economic reforms and public safety, two of the most pressing concerns for Mexican citizens.
Presidents in Mexico are elected through a plurality vote and serve a single six-year term. Voter turnout for the election stood at 61.04 per cent (International IDEA n.d.c).
The 2024 election cycle was marked by a record high amount of political violence, with 330 recorded incidents targeting political figures; at least 95 incidents led to one or more reported deaths (Pellegrini and Breda 2024). A report by Armed Conflict Location and Event Data demonstrates that political violence occurs primarily at the local level, targeting candidates, former officials and political affiliates. Organized-crime groups were identified as the primary perpetrators of these attacks (Pellegrini and Breda 2024).
In addition to physical violence, the election period saw a surge in gender-based abuse, particularly online. Mexico has one of the highest femicide rates in the world, and activists criticized the lack of campaign promises to tackle domestic violence (Jaime 2024). The electoral observation mission of the Organization of American States (2024) reported that the Mexican EMB had received 199 complaints regarding gender-based violence against women. Moreover, the Wilson Center (Calderón 2024) noted an increase in complaints of gender-based violence against women in recent years, especially in the context of elections, as well as incidents of sexist comments and stereotypes of women both online and offline. There is also evidence that political violence is not only becoming more common but is also being committed against women in more varied forms and in more severe ways, ranging from online threats to physical assault to even murder (Orozco 2024c). The two leading female candidates, Sheinbaum of Morena and Xóchitl Gálvez of the National Action Party, faced significant harassment on social media. Sheinbaum, in particular, was subjected to anti-Semitic slurs and false claims that she was born in Hungary—rumours she had to repeatedly refute (Klepper 2024).
Several factors contributed to the spread of disinformation during the elections, including growing distrust of the news media, widespread violence committed by drug cartels, rapid increases in social media usage coupled with a lag in digital literacy and the deliberate spreading of falsehoods by political figures themselves (Breuer 2024; Klepper 2024). As political figures are some of the main actors spreading disinformation online, there is a systemic lack of incentives to combat false information by regulatory means.
Digital context
Mexico’s digital landscape has changed rapidly, with Internet penetration reaching 83.2 per cent at the beginning of 2024 (Kemp 2024d). While digital infrastructure has expanded in urban centres, a significant divide remains between urban and rural areas. To bridge this gap, former President López Obrador launched an initiative to expand Internet access in rural communities. His administration claims to have increased rural Internet coverage from 10.8 million people in 2018 to 17.5 million in 2024, representing a 60 per cent rise in connectivity (AMLO 2024).
Alongside Brazil, Mexico has the most Internet users in Latin America: as of 2024, 74.63 per cent of Latin America’s population had Internet access (Statista 2024c). Social media usage in Mexico has surged, with 99.9 per cent of the population above the age of 18 on social media, with a near even spread of users across genders (Kemp 2024d). The most widely used social media platforms in Mexico are Meta-owned WhatsApp, Facebook and Instagram, with 65 per cent of users being under the age of 35 (Statista 2024c).
Given that over 61 per cent of Mexicans receive their news through social media (UNESCO and IPSOS 2023), the integrity of online information plays a crucial role in shaping public perception of electoral processes. The spread of misinformation remains a pressing issue. However, digital literacy remains a challenge. The National Institute of Statistics and Geography reported in 2019 that 20 per cent of Mexican families were ‘digitally illiterate’, meaning that they did not own a computer—not due to a lack of resources but because they did not understand how computers work (INEGI 2019). Internet speeds in Mexico are also very low compared with global averages, reaching only 60 per cent of fixed Internet speeds and 50 per cent of mobile Internet speeds compared to global averages (Kemp 2024d).
While Mexicans largely agree that governments and regulators are required to put trust and safety measures in place during election campaigns to protect the integrity of elections (UNESCO and IPSOS 2023), regulation efforts have had little success thus far. Despite some agreements signed between the authorities, such as the National Electoral Institute (INE), and platform owners like Google or Meta to detect disinformation operations and educate voters during elections, efforts to mitigate the spread of disinformation online remain limited. These agreements have followed a largely self-regulatory model that relies on the social responsibility of the platforms themselves (Echeverría 2023).
CSOs active in Mexico claim that AI is promoted within the country through an economic and social development narrative that has failed to involve CSOs in a structured dialogue with state entities on commonly agreed rules concerning the regulation of AI systems (AlSur 2024). Crucially, former President López Obrador frequently and actively engaged in disseminating disinformation, often aimed at discrediting journalists, demonstrating a pattern observed globally by scholars whereby attacks on information integrity are used by autocrats to help maintain and expand their power (Sinanoglu and Breuer 2024).
Disinformation tactics, techniques and procedures
Coordinated networks of influence
It is estimated that more than half of all Internet traffic in Mexico is made up of automated bots, many of which are responsible for spreading and amplifying fake news and destructive narratives (Imperva 2024). During the country’s previous election, in 2018, bots were extensively employed to orchestrate and disseminate electoral disinformation campaigns with the intent of influencing the outcome of key races (Broderick and Arredondo 2018; Echeverría 2023).
Attempts to interfere in the 2018 election through the use of bot campaigns have been linked to both foreign and state actors (Freedom House 2023c). Domestic bot usage is not linked to specific political actors but is rather a common practice across party lines (DFRLab 2018). Campaigns generally employ bots to artificially amplify messages through post engagement or boost the virality of discussion topics by making hashtags trend. One such example is the hashtag #LastiriSi, which was promoted to support the Institutional Revolutionary Party’s Senate candidate Juan Carlos Lastiri Quiros. Over the course of two days, the hashtag appeared in 31,000 posts on X (formerly Twitter). Some 24,718 of these hashtag mentions originated from just 30 accounts, meaning the hashtag was mentioned on average 824 times per user in less than 48 hours. A vast majority of the accounts had little to no prior activity on the social media platform, indicating that it is unlikely they were linked to an authentic user (DFRLab 2018).
The risk of interference by bots in the 2024 elections has been stressed by several Mexican officials, including the president of the ruling Morena party, Mario Delgado. In an open letter to the executive representatives for Meta, TikTok and X in Latin America, Delgado urged the companies to act against the ‘dirty war’ being waged against the Morena party’s 2024 presidential candidate, Sheinbaum, and then–President López Obrador (Aguilar 2024; Delgado 2024). One of the most prominent falsehoods spread as part of this so-called dirty war accused Sheinbaum of having ties to drug trafficking, which was circulated using the hashtag #NarcopresidenteAMLO (InSight Crime 2024).
A metadata analysis of 10 million X posts using hashtags related to the campaign found that more than 500,000 posts contained the same false claim. These posts also originated from accounts located outside Mexico and used stolen profile pictures (Gutiérrez Aranda 2024). Morena filed a complaint with the INE against the opposition presidential candidate Gálvez, accusing her of financing foreign bots as part of a digital strategy to promote her presidential campaign and slander political rivals from Morena. Gálvez denied the accusations, insisting that her supporters were authentic and that Morena’s claims were exaggerated (Animal Político 2024).
Strategic use of disinformation to disrupt the information environment
The Morena party and former President Lopéz Obrador have also been criticized by political observers and academics for disseminating unsubstantiated and false claims about opposition candidates (Martínez Chacón 2023b). Several pieces of disinformation have been identified by independent fact-checking organizations on social media, including false claims about the campaign promises of opposition candidates. Political analysists have highlighted that the dissemination of disinformation across party lines shares a common strategy of attempting to maximize political polarization (Echeverría 2023; Martínez Chacón 2023a). When party officials use their platforms on social media to disseminate false or deceptive narratives, the information landscape risks becoming increasingly confusing for audiences. This problem is particularly pervasive online, as narratives can be picked up, amplified and kept in circulation by actors seeking to influence political perceptions (Breuer 2024). Reiteration of disinformation amplifies its reach, bloats online information spaces and provides perceived legitimacy to inauthentic information.
Cross-platform amplification
Although inauthentic amplification campaigns in Mexico often originate on X, most disinformation gains a foothold once it appears on other platforms. X hosts 18.02 million Mexican users, which is low in comparison with the 90.20 million Facebook or 83.10 million YouTube users (Kemp 2024d).
On X, actors seeking to influence public opinion use consolidated strategies to boost specific messaging with the intent of reaching as large an audience as possible. These strategies include studying the behaviours of Mexican social media users through analytics to maximize engagement, such as determining the time of day a post is more likely to garner attention. These campaigns are predominately coordinated by two different types of groups. The first type consists of independent actors who coordinate joint action in private messaging groups on Telegram or WhatsApp, employing their own networks (Aguirre 2021; Cerdeira 2024). The other type consists of actors who outsource the implementation of disinformation campaigns to paid companies (Naciones Unidas México 2021).
Once campaigns gain virality, audiences share content with their networks on messaging apps such as Facebook Messenger, Telegram or WhatsApp (AFP 2021). This move to largely uncensored chat platforms is necessary to circumvent the increasing number of restrictions that public social media platforms have introduced to counter disinformation (Aguirre 2021). It is primarily at this moment that content—made by bots and reaching a small audience through posts on public platforms—is transformed into discussions between real online communities. A study from the 2018 election period found that cross-platform transmission patterns were gaining traction compared with the spread of disinformation on public platforms, stressing the importance of one-on-one chat communications (Glowacki et al. 2018). In parallel to private messaging chats, viral political topics on public social media platforms gain legitimacy and attention in public discussions outside online forums due to artificially inflated engagement numbers. This aligns with the public statements issued by prominent political figures and party representatives, such as Delgado’s open letter to the owners of social media platforms following disinformation campaigns against the Morena party.
Generative AI
AI was at the centre of political controversy in Mexico’s 2024 elections due to several instances of its use in promoting disinformation and conspiracy theories. Following several parties’ announcements of pre-candidates for the mayoral race in Mexico City, a faked audio recording of the then-current Mayor, Martí Batres, of the Morena party surfaced on social media. On the recording, a person purported to be Batres could be heard making conspiratorial remarks about plans to sabotage the campaign of the Morena party’s candidate, Omar García Harfuch (Ramis 2023). The audio clip went viral, provoking Batres to issue an official statement disputing its veracity on X, which was met with both scepticism and support from audiences (Batres 2023; Nación321 2023a). Alongside Batres’s statement, Mexico’s Director General of Digital Government, Eduardo Clark, issued a public explanation of how deepfakes are created, providing an example that he created himself in 15 minutes using software that costs only MXN 200 (USD 10) a month (Nación321 2023b).
In another case, a deepfake of presidential candidate Sheinbaum encouraged viewers to take part in a fraudulent investment scheme, which Sheinbaum later denounced in a post on X (Lagos 2024a). Beyond this prominent case, reports indicate that in at least three additional cases images of politicians were exploited to commit cybercrimes, including through phishing and identify-theft schemes (Del Pozo and Arroyo 2025). Many of these cases portrayed politicians inviting users to invest in fake financial platforms.
Julio César Bonilla, the Citizen Commissioner of Information at the CDMX, an autonomous constitutional body with the mandate to guarantee accountability and transparency, made a statement urging voters to be vigilant of how AI could be used to manipulate information (Montes 2024). Candidates also employed generative AI in campaigns for promotional purposes. One prominent example was iXóchitl, an AI-generated spokesperson identical to the presidential candidate Gálvez herself. iXóchitl was used as a deepfake stand-in for Gálvez in campaign videos and images with her ‘endorsement, supervision and approval’ (Gamboa 2023; Del Pozo and Arroyo 2025).
Gender-based political violence generated by AI contributed substantially to information pollution in the 2024 electoral cycle in Mexico. Given the high number of female candidates competing for key political positions, several tactics were employed to discredit their candidatures, including the production of non-consensual pornography, which was reported in at least two cases, one with the Citizens’ Movement candidate for mayor of Monterrey and another involving the Labor Party candidate in Tamaulipas (Méndez 2024).
However, keeping the above concerns in mind, it is important that the actual impact of AI systems on the elections is not overblown. Studies examining instances where AI-generated content was used to influence the elections found that such cases received little media attention or social media engagement and had no correlation or measurable impact on election outcomes (Del Pozo and Arroyo 2025).
Countermeasures
Certeza INE and access to electoral information
In 2018 the INE launched a campaign (called Certeza INE) to combat the spread of disinformation on social media during electoral processes. The campaign was relaunched ahead of the 2024 elections with five main strategies: (a) verifying information; (b) ensuring bilateral cooperation between stakeholders; (c) cooperating with the media; (d) building media literacy; and (e) providing transparency and verified electoral information (INE 2024a). The campaign’s online platform served as a fact-checking hub that utilized multidisciplinary expertise from bilateral partnerships to debunk false or misleading information circulating online. Representatives from the INE also travelled on missions throughout the country to provide local training for citizens in preparation for the election (INE 2024c). During these missions the INE gathered information on existing disinformation narratives and generated responses to debunk them on the web platform, thereby proactively curbing the spread of disinformation (INE n.d.a). Aside from the Certeza campaign, the INE also published electoral information such as poll locations and lists of registered political actors on its main website (INE n.d.b, n.d.c).
Chatbot to verify electoral information
As part of the Certeza campaign, the INE launched a virtual assistant intended to support constituents in finding correct and trustworthy information on the 2024 elections. The Inés chatbot, developed in partnership with WhatsApp, was connected to a phone number like a regular contact on the app. The bot provided users with a channel for submitting videos, photos or text. This material was subsequently transmitted to several fact-checking teams that responded to queries which were categorized, analysed and verified with the assistance of a collaborative intelligent platform. In addition to fact-checking, users could ask questions and receive verified information about the election from a network of independent actors, including journalists (INE 2024b).
Government collaboration with owners of social media platforms
In February 2024 the INE announced its collaboration with the Mexican CSO Movilizatorio and the social media platform company Meta. The partnership involved the launch of a new edition of the I Am Digital (Soy Digital) educational programme, aiming to tackle electoral disinformation. The programme is improving media literacy through workshops for students, civil society, the public sector and academics on how to be cognizant of disinformation. Workshop guides include online toolkits published on social media to encourage critical thinking and provide online navigational skills (Vergara 2024).
TikTok also announced a broad commitment to combat the spread of disinformation on its social media platform during the Mexican elections (TikTok 2024). In addition to site-wide measures which affect elections globally, they also instated an election guide specifically customized for the Mexican election. The guide provided official data, logistical information on how to vote and access to helplines. To complement the guide, TikTok introduced hashtags available exclusively for pre-verified content on the elections. The information in the guide was sourced from the INE, which partnered with TikTok for its development. TikTok also established other strategic partnerships with electoral authorities, including the Electoral Tribunal of the Federal Judiciary (Valverde 2024). The INE also entered into similar partnerships with Google and YouTube, aimed at strengthening digital protocols, reporting systems and access to electoral information for general audiences (Lagos 2024b)
Independent fact-checking and media literacy initiatives
There are several independent and media-affiliated Mexican fact-checking initiatives currently active online. Verificado is one of most prominent organizations dedicated solely to debunking fake news and improving media literacy. Founded ahead of the 2018 election, it is an independent organization which relies on individual contributions from journalists that are uploaded to a central website (Duke Reporters’ Lab n.d.). This online hub features fact-checks of circulating disinformation, in-depth analyses of misleading narratives and guides to navigating online information spaces. Leading up to the 2024 elections, the site worked proactively by posting reliable information on electoral processes and articles that encouraged users to think critically when confronted with political information (Barbosa 2024a, 2024b; Orozco 2024b). It also carried out efforts to improve media literacy through courses and workshops on fact-checking, journalism and ethics. The workshops were mainly directed at fellow journalists, but they were also available for general audiences (Verificado n.d.).
In May 2024 Verificado hosted an electoral hackathon together with the civic initiative Cómo Vamos Nuevo León. The hackathon invited participants from the public to conduct analyses of sources of electoral information with guidance from journalists and experts. The analyses focused on how to find, identify and evaluate the credibility of political information (Orozco 2024a).
Conclusion
The 2024 elections in Mexico highlighted many concerning developments around information integrity. As one of the most online countries in Latin America, Mexico was heavily affected by coordinated bot networks, allegedly deployed by political actors themselves. In addition to bot networks and cross-platform amplification tactics, gender-based violence online was a dominant theme in Mexico’s elections: female political candidates faced derogatory remarks or harassment online, with instances of deepfakes deployed to discredit candidates or discourage them from running for office.
While Mexico’s INE introduced several measures to combat disinformation, such as a nationwide campaign to improve access to factual information and media literacy training, the Mexican case highlights how offline societal issues can amplify digital challenges. The country’s political culture has long been marked by violent rhetoric and the defamation of opposition candidates, dynamics that have only intensified online. Moreover, widespread violence against women in Mexico makes them especially vulnerable to online abuse. As such, the Mexican example underscores the importance of addressing information integrity through a comprehensive, whole-of-society approach.
Introduction
Mongolia’s 2024 parliamentary election was held on 28 June, following constitutional changes enacted in May 2023 that expanded the size of the State Great Khural (the country’s unicameral parliament) and introduced a mixed electoral system (Congressional Research Service 2024). These reforms contributed to significant shifts in parliamentary representation: the ruling Mongolian People’s Party (MPP) narrowly retained its majority but with a significantly reduced margin (Dierkes 2024). The opposition Democratic Party (DP) increased its share of seats, while smaller parties also gained representation thanks to the new electoral rules. Additionally, the election resulted in a record number of women, young people and individuals from non-traditional professional backgrounds entering parliament (AP 2024; Lkhaajav 2024; Rikkilä Tamang 2024).
The election took place against a backdrop of widespread political disillusionment, with 86.4 per cent of Mongolians believing that political parties do not represent the will of voters (Adiya 2024). This sentiment was particularly strong among young Mongolians, who often feel undervalued by the political establishment and face economic challenges (AP 2024; RTL Today 2024). In this context, the MPP sought to reform the electoral system to appeal to a more diverse electorate. Major campaign issues included corruption, economic hardship in rural areas, mining and environmental policy, and healthcare (Dierkes 2024; OSCE/ODIHR 2024). Despite voter frustration, Mongolia remains a multiparty democracy, with voter turnout at 69.3 per cent—only a small decline from 2020, demonstrating continued strong political engagement (Dierkes 2024).
Digital context
At the start of 2024, Mongolia had approximately 2.91 million Internet users, or 83.9 per cent of the population. At the same time, the country was home to 2.5 million active social media users, or about 72 per cent of Mongolians, with Meta’s Facebook being the most widely utilized application (Kemp 2024e). This number reflects a significant increase from previous years, highlighting the growing importance of digital platforms in the country (Banzragch, Mashlai and Mavag 2024). However, digital access remains highly uneven, with urban areas, particularly the capital, Ulaanbaatar, having higher Internet use rates compared with rural provinces (Battsengel et al. 2019). Although Internet usage is generally high across all age groups, individuals aged 15 to 49 reported the highest use rates, reaching 93.1 per cent in a three-month period in 2023 (NSOM, UNICEF and UNFPA 2024). Conversely, gender disparities are minimal, with men and women reporting very similar rates of Internet usage.
The media landscape in Mongolia is diverse, comprising state-controlled, private and independent outlets (Myagmar and Nielsen 2001; OSCE/ODIHR 2024). Traditional media, especially television and radio, maintain substantial influence, particularly in rural regions (Embassy of the Czech Republic in Ulaanbaatar n.d.). However, Mongolians tend to be mistrustful of most news sources due to concerns regarding the quality and accuracy of the content (RSF n.d.). In addition, the rise of online news consumption has begun to reduce Mongolians’ reliance on traditional news sources and introduced concerns regarding media ownership and potential governmental influence over major news organizations. These issues have been exacerbated by recent prosecutions of journalists under defamation laws, which have intensified fears regarding press freedom and media independence (International IDEA 2024c). Additionally, the Communications Regulatory Commission (CRC), a government-controlled body, has the authority to suspend broadcast licences, a power that has raised questions about media freedom and potential censorship (Naranjargal 2018; Boldkhuyag et al. 2021).
In recent years, the Mongolian Government has implemented several laws and regulations to address challenges arising from the use of social media platforms. Notably, the Law on Cybersecurity and the Law on Personal Data Protection, both enacted in 2021, aim to protect information security and personal data (American Chamber of Commerce in Mongolia 2023). While these laws provide a foundation for addressing digital crimes, concerns persist about their potential impact on freedom of expression and the operations of media outlets. Furthermore, the proliferation of misinformation, particularly on social media, has become a growing issue in Mongolia (Mantas 2021). Both domestic political entities and foreign actors have been implicated in digital manipulation efforts, aiming to influence public opinion and electoral outcomes (Tsedendamba 2024). The CRC has been criticized for its role in regulating online content, with widespread allegations of censorship, partiality and a lack of transparency (Naranjargal 2018). Regulations impose broad content restrictions, and the government-appointed CRC oversees broadcast licensing without public consultation, raising concerns about fairness and political influence in media access (United States Department of State n.d.).
Disinformation tactics, techniques and procedures
Coordinated networks of influence
Mongolia’s 2024 election revealed the extent to which political influence networks shape media coverage and public perception. The Organization for Security and Co-operation in Europe’s Office for Democratic Institutions and Human Rights (OSCE/ODIHR) found that major Mongolian broadcasters disproportionately covered the ruling MPP and the opposition DP, with a notable bias towards the former (OSCE/ODIHR 2024). Many television channels and online media outlets echoed paid political advertisements, blurring the lines between independent journalism and party-driven narratives. Conversely, critical perspectives were limited, with investigative and analytical reporting largely absent from election coverage. Online media followed a similar pattern, with most outlets dedicating between 23 per cent and 42 per cent of election-related reporting to the MPP in a positive light (OSCE/ODIHR 2024). This coordinated network of influence, where political parties directly shaped media narratives, significantly restricted voter access to diverse and impartial information, undermining the democratic process.
Foreign influence played a significant role in Mongolia’s 2024 parliamentary election, with both China and Russia engaging in coordinated efforts to shape political narratives (Tsedendamba 2024). Given Mongolia’s strategic location between these two global powers, the country’s information space remains vulnerable to external propaganda campaigns. Chinese state-backed media outlets promoted narratives emphasizing economic stability and regional cooperation, aligning with the broader objectives of Beijing’s Belt and Road Initiative. Meanwhile, Russian media focused on anti-Western sentiment, reinforcing narratives critical of democratic institutions and Western alliances. These influence campaigns operated through a combination of manipulating social media, spreading narratives via traditional news platforms and engaging directly with political actors. According to a Political Barometer Survey, most Mongolians still see the Russian Federation as the best partner for Mongolia, despite Moscow’s decision to invade Ukraine in early 2022 (Tsedendamba 2024). The influence of foreign propaganda and disinformation complicates efforts to ensure an informed electorate, raising concerns about national sovereignty and the resilience of Mongolia’s democratic institutions.
Cross-platform amplification
Cross-platform amplification played a crucial role in spreading Russian narratives within Mongolia’s information space during the 2024 election. Disinformation often originated from statements made by Russian officials, which were then amplified through multiple media outlets, including Mongolian state media (Tsedendamba 2024). For instance, after the president of Buryatia claimed in September 2023 that the United States was attempting to orchestrate a coup in Mongolia, this narrative was quickly disseminated by over 20 Russian media outlets and subsequently reposted by mainstream news sources in Mongolia. Similarly, fabricated claims that the USA was planning a Ukraine-style intervention in Mongolia appeared in Russian state media in November 2023, before being republished by Mongolian platforms like RT Mongolia and other local newspapers (Tsedendamba 2024).
These disinformation tactics relied on a coordinated ecosystem of pro-Kremlin media and Mongolian state outlets, ensuring that false narratives reached broad audiences (Tsedendamba 2024). The method of republishing articles across various platforms can help lend legitimacy to misleading content by creating an illusion of independent corroboration (Wilson and Starbird 2020). This cross-platform strategy can also make fact-checking efforts more difficult, as disinformation could be framed as coming from ‘multiple sources’, reinforcing its perceived credibility among the public. The rapid spread of narratives about US influence and supposed threats to Mongolia’s sovereignty highlights the vulnerability of the country’s media landscape to foreign manipulation and underscores the critical need for greater resilience against cross-platform disinformation campaigns.
Strategic use of disinformation to disrupt the information environment
In Mongolia, disinformation has been strategically deployed to manipulate public perception and create confusion around critical issues, weakening trust in institutions and credible information sources. The Mongolian Fact Checking Center (MFCC) has highlighted the difficulty of countering misinformation in a country where there is no direct equivalent for the term ‘misinformation’, making public awareness campaigns even more challenging (Mantas 2021). Political actors and foreign influence networks have taken advantage of this linguistic and institutional gap, spreading disinformation that blurs the lines between truth and falsehood. One key example was the wave of Covid-19 vaccine misinformation, where false narratives were used to erode public trust in vaccines (WHO 2023).
The phenomenon of misinformation translation further complicates efforts to address these issues. Research by Quelle et al. (2023) indicates that when misinformation is translated across languages, it often undergoes semantic drift, altering its meaning and sometimes even amplifying its persuasive power. In Mongolia, this challenge is particularly pronounced, as misleading narratives introduced in foreign languages may take on new connotations when adapted to the local context. The dissemination of disinformation can overwhelm fact-checking efforts in Mongolia, particularly since the country has a large nomadic population that is difficult to reach with verified information (Mantas 2021). Therefore, there is a need for strategies that not only combat disinformation but also account for the linguistic and cultural complexities that shape its transmission and reception.
Microtargeted content
Political parties around the world are increasingly crafting highly personalized digital advertising campaigns (International IDEA 2020). These campaigns utilize online political advertising strategies that enable political actors to segment voters based on interests, location and online behaviour. While microtargeting has been effective in engaging voters, concerns remain over transparency and potential voter manipulation. In Mongolia’s 2024 election, the role of social media in political campaigning was particularly pronounced, given the short, 18-day official campaign period, which advantaged incumbents with longer-standing visibility. Smaller parties and independent candidates relied more heavily on social media to reach voters, while larger, well-funded parties were able to supplement digital efforts with in-person events and door-to-door canvassing (OSCE/ODIHR 2024). However, despite legal provisions for monitoring online campaign content, enforcement mechanisms were weak, making it difficult to track which messages were delivered to which audiences. This lack of oversight raises further concerns about potential voter manipulation through digital advertising strategies, particularly in a highly competitive and time-constrained electoral environment.
Spending patterns on digital advertising further illustrate disparities in campaign reach and messaging strategies. According to data from the Meta Ad Library, individual candidates and political entities collectively spent tens of thousands of euros on social media advertising, with some candidates outspending entire parties. The MPP, for example, reported over EUR 120,000 (approximately USD 140,000) in social media expenditures, far surpassing the amounts disclosed by smaller parties such as the HUN Party and National Alliance. Meanwhile, the DP reported minimal spending, highlighting the uneven use of digital platforms across political groups (OSCE/ODIHR 2024). Despite the extensive use of social media, campaign messages specifically targeting women were largely absent, underscoring ongoing gender gaps in political outreach.
Countermeasures
Legislative and regulatory initiatives
Mongolia has implemented several legislative and regulatory measures aimed at addressing disinformation and improving media oversight. The Parliamentary Elections Law, the Cyber Security Law, the Law on Freedom of the Press, the Law on Personal Data Protection and the Law on Broadcasting are all central to Mongolia’s efforts to regulate digital content and increase platform accountability, but concerns abound regarding their application (RSF 2024; Open Government Partnership 2021; Volodya 2022; American Chamber of Commerce in Mongolia 2023). The Parliamentary Elections Law specifically regulates social media during elections, imposing restrictions on political advertising and mandating content removal for electoral violations (OSCE/ODIHR 2024). Critics argue that vaguely defined provisions on ‘false information’ could be used to suppress dissent rather than effectively combat disinformation (Boldkhuyag et al. 2021). Additionally, the CRC has broad authority over online content, but its enforcement has been criticized for lacking transparency and fairness (Naranjargal 2018).
In addition, regulatory gaps have allowed the continued flow of disinformation, particularly during the Covid-19 pandemic. Limited oversight of social media platforms contributed to the spread of false health information, highlighting the need for stronger digital governance (Tovuu 2020). During the pandemic, Mongolia began imposing fines—under the Law on Offences—on anyone accused of spreading disinformation about possible risks of disasters (Ankhtuya 2020). However, practical enforcement of this legislation was limited, and the true impact was a chilled environment for journalists, who resorted to self-censorship in the face of potential fines and sanctions (Tovuu 2020).
Digital literacy initiatives
Efforts to bridge the digital divide and enhance digital literacy in Mongolia are ongoing, with a particular focus on vulnerable demographics such as youth, women and girls, and rural populations. Programmes such as Digital Nation, launched in 2020, aim to improve public access to digital technology and government services, enhance online security and promote responsible social media use (NSOM, UNICEF and UNFPA 2024). However, digital literacy gaps remain, particularly among the elderly, who are often more susceptible to misinformation due to limited media literacy training (Baatar 2023).
To counter these challenges, media organizations and civil society groups have launched training sessions to help citizens identify misinformation and navigate digital content responsibly. For example, the MFCC launched an interactive fact-checking tipline chatbot, called Asuu, on Messenger in May 2024 to help the public verify questionable content (Meedan 2024). Meanwhile, international organizations like DW Akademie and UNICEF Mongolia have partnered with local institutions to provide fact-checking education and digital literacy training, particularly during election periods (Tsolmonbaatar 2019; Battulga 2023). These initiatives emphasize the growing need to equip citizens with the tools to critically assess online content and prevent the spread of disinformation.
CSO engagement
CSOs in Mongolia play a critical role in combating disinformation, enhancing media literacy, and advocating for government transparency. Various CSOs, including the Nest Center for Journalism Innovation and Development, have worked to improve public access to accurate information through fact-checking initiatives and digital literacy programmes (Meedan 2024). The MFCC is leading a coalition of more than a dozen groups, consisting of media outlets, academics and CSOs, which are dedicated to these initiatives. CSOs have also pushed for greater government transparency by monitoring social media and public discourse for signs of misinformation (Batdorj 2020).
Furthermore, Mongolian CSOs have played a role in addressing health-related disinformation, particularly during the Covid-19 pandemic. CSO-led initiatives, in collaboration with UNICEF Mongolia, focused on dispelling myths surrounding vaccines and providing accurate health information to combat vaccine hesitancy (Battulga 2023). These efforts underscore the critical role of CSOs in strengthening public trust in factual information and countering misleading narratives across Mongolia’s media landscape.
Conclusion
Mongolia’s 2024 parliamentary election underscored both the strengths and vulnerabilities of the country’s democratic system in the face of growing digital influence and disinformation. While the electoral reforms allowed for more diverse representation in parliament, the campaign period was marked by media bias, foreign influence, and the proliferation of misleading narratives across social and traditional media platforms. Despite government efforts to regulate digital content and combat disinformation, concerns remain about the balance between media oversight and freedom of expression. CSOs and media literacy initiatives have played a crucial role in mitigating the impact of false information, but continued investment in public education and transparent governance will be essential to safeguarding Mongolia’s democratic integrity. As digital media becomes increasingly central to political discourse, Mongolia must strengthen its resilience against disinformation while upholding its commitment to free and fair elections.
Introduction
The results of Pakistan’s 8 February 2024 general election have raised concerns about the country’s electoral stability (Mir and Salikuddin 2024). A high degree of polarization and political unrest led to Internet and mobile service restrictions in several regions around election day (Sharma 2024). Shortly after the vote, Pakistan’s Electoral Commission released the results, confirming that associates aligned with the controversial ex-Prime Minister Imran Khan, of the Pakistan Tehreek-e-Insaf (PTI) party, received the largest number of votes. Khan himself is barred from holding public office for 10 years following his 2022 removal from office through a vote of no confidence and subsequent indictment on several charges (Reuters 2024). Nevertheless, he still wields significant influence in shaping the country’s political discussions and democratic procedures.
As no party held a simple majority and there was little willingness to cooperate with PTI affiliates, the Pakistan Muslim League Nawaz and the Pakistan People’s Party formed a coalition government, appointing Shehbaz Sharif as Prime Minister (Seddon and Javed 2024). The exclusion of the PTI, despite its status as the largest party in parliament, caused high levels of distrust in the electoral system among Pakistani voters, resulting in political unrest and protests across the country. This unrest was further exacerbated by statements from PTI candidates supporting the narrative that the 2024 election was stolen or rigged and an admission from a government official that the electoral results were manipulated (Gul 2024; Malik 2024). Ultimately, the unrest and uncertainty surrounding the integrity of the election negatively impacted voter turnout, which dipped from the previous election’s 52.1 per cent to 47.6 per cent, despite an increase in the number of eligible voters (ANFREL 2024b).
Digital context
Internet usage is steadily increasing in Pakistan and currently stands at 45.7 per cent of the total population, falling slightly below the regional average of South Asia at 51.5 per cent (Kemp 2023b). Less than half of the population (38.9 per cent of the population aged 18 and above) are social media users (Kemp 2023b). There is a considerable gender gap in Internet usage, often called the ‘digital purdah’; placing globally, Pakistan ranks 79th on the Internet Inclusivity Index, with most major social media platforms showing at least two male users for every female user (Economist Impact n.d.; Kamran 2023; Kemp 2023b). Internet access is also severely impacted by a lack of affordability and digital infrastructure in rural areas, creating deep fissures in the digital divide between regions and socio-economic groups. Consequently, the Internet usage rate is significantly higher in urban areas (Freedom House 2021).
Notwithstanding the low levels of Internet usage, disinformation has a significant impact on the country’s digital information landscape. A study by the United States Institute of Peace documented a positive correlation between political knowledge and trust in disinformation as well as high levels of disbelief in genuine political information (Mir and Siddiqui 2022). In another survey, 60 per cent of respondents expressed feeling uncertain of their ability to consistently identify misinformation (Naeem and Rehmat 2022). Concerns about the credibility of information extend to high levels of authority, as political and state officials routinely use deceptive and oppressive tactics in campaigns and regulation (Shaheen 2022).
The country’s digital information landscape has become increasingly inflamed since Khan’s removal from office in 2022—a major political event that resulted in regular public unrest and Internet shutdowns (Mao 2023). The digital arena has taken centre stage in this conflict, as the PTI has constructed its image to a great extent around being an Internet-based and social media–savvy party (Jahangir 2020). Disinformation and other influence tactics were widely used in the 2024 election to slander political rivals, undermine democratic procedures and misdirect voters.
Disinformation tactics, techniques and procedures
Coordinated networks of influence
- Volunteers. Pakistan’s primary networks of influence are commonly dubbed ‘troll armies’, which consist of party supporters who are mobilized to promote their affiliated party and slander political opponents or independents. Among the country’s political parties, the PTI employs this tactic most frequently, engaging Khan’s supporters in so-called ‘virtual jalsas’, meaning ‘virtual rallies’. These rallies are often incited by deepfake speeches, purportedly by Khan, which are live-streamed on YouTube and encourage supporters to spread party messages on social media (Khan 2023). Khan’s supporters are encouraged to take part in what is estimated to be one of the largest networks of its kind on X (formerly Twitter), where virality is achieved through the coordination of trending hashtags (TRT World 2019; Hendrix and Jahangir 2024). A participant in political campaigns on X in Pakistan describes, in an interview, how virality is achieved through a three-step process—planning, research and execution. The initial step requires the participants to examine public discussions in order to identify relevant political topics with potential to go viral. The second step entails research into the topic and the development of a common language and graphical profile for the campaign among its founders. Lastly, participants coordinate to post simultaneously during time slots with the highest potential to generate the maximum amount of traction (TRT World 2019). The influence of campaigns on X has resulted in frequent throttling (intentional slowing of the platform) and shutdowns, leading to a growing shift towards TikTok (Hendrix and Jahangir 2024). The members of troll armies may intentionally or unintentionally spread unverified and misleading information—further exacerbating the spread of misinformation.
- Employees and bots. The Oxford Internet Institute has documented—in addition to volunteer groups—the existence of paid ‘cyber troops’, who are employed full-time to promote parties and slander their rivals on social media. These troops are almost always artificially boosted by coordinated networks of bots masquerading as human users (Bradshaw et al. 2021). Actors that spread disinformation motivated by financial incentives often take part in so-called hashtag mills, where virality can be purchased. As of 2020 it was estimated that nearly 95 per cent of Pakistan’s trending political campaigns were artificially boosted to give the appearance of grassroots support (Jahangir and Popalzai 2020). These bot-driven campaigns also advocate for hostile and disruptive behaviour, such as inciting violence against political opponents (Bradshaw et al. 2021). In recent years, this type of coordination has increased, particularly affecting journalists and politicians, with women and other marginalized groups being especially targeted for trolling and character assassination (Dawn 2023). Generative AI and photo manipulation have been used to doctor images of female and transgender politicians, aiming to sexualize them without their consent and publicly humiliate them (A. Khalid 2023).
Misleading ‘fact-checking’ as an influence tool
The PTI’s volunteer network was previously coordinated to ‘debunk’ what the PTI alleged to be fake news. Volunteer flaggers would tag pieces of alleged disinformation to which PTI officials would publish responses that provided ‘authentic details’. The responses, however, were generally worded as offensive, personal attacks. Other Pakistani parties are increasingly taking similar measures in their social media posts (Jahangir 2020). In employing such tactics, political actors undermine independent initiatives to counter disinformation under the guise of credibility, by abusing the audience’s trust of fact-checking methods.
Cross-platform amplification
Actors aiming to spread disinformation in Pakistan have utilized connections between social media platforms to reach different audiences and to amplify the spread of content. These aims are achieved in two ways: (a) directly, by cross-posting the same content onto several platforms; or (b) indirectly, by starting trending hashtag campaigns whose virality enables them to appear separately in different groups or pages. A key tactic of these campaigns is the reiteration of material, achieved by reposting the same content many times (Jahangir 2020). The indirect method is also bolstered by social media platforms’ built-in algorithms that boost sensationalist or reactionary content to drive engagement (Khan 2024).
Microtargeting
Targeted political advertising, which is rising in popularity among political parties in Pakistan, is generally conducted through group pages on social media platforms such as Facebook and closed group chats on WhatsApp (Jahangir 2020). Experts have expressed concern that closed discussions in group chats may result in the filtering of information to align with party agendas and thereby shut out participants from diverse political discussions. This filtering of information is particularly concerning, as users in Pakistani WhatsApp groups rarely attempt to verify information or confront actors that spread disinformation (Haroon et al. 2021). Research on Pakistani message groups has shown that the transmission of misinformation between platforms enhances its pervasiveness, as false narratives that reached WhatsApp groups remained in circulation for longer than other categories of information, such as news or religious information (Naeem and Rehmat 2022). Another microtargeting tactic is the trading of databases that contain voters’ contact information. Software enables buyers to access millions of phone numbers categorized by identifiable traits such as gender or district. Vendors often offer additional services that enable buyers to send bulk messages to select demographics on WhatsApp (Jahangir 2018).
AI-generated content
The PTI is an early adopter of generative technology, incorporating extensive use of deepfakes into its campaign messaging. The most controversial application of AI has been the generation of videos simulating speeches held by the imprisoned Khan, based on scripts that he wrote himself. These applications are often overt, with notes acknowledging the use of AI in text attached to the deepfake (Hendrix and Jahangir 2024). However, not all uses of AI are disclosed, and the PTI has come under fire for not being transparent regarding other AI-generated material used in campaigning (Bahl 2023). Other actors have also covertly employed AI to misdirect voters by falsely suggesting that candidates have stepped down from running, impacting voters’ choices (Ali 2024).
One wave of deepfakes targeting the PTI circulated on social media just one day before the general election. These deepfakes portrayed prominent members of the party, Khan included, encouraging voters to support the party by boycotting the election. Members of the PTI responded by releasing statements on X that denounced the false boycotts and accused the Pakistan Muslim League Nawaz of spreading the deepfakes to influence voter choices through deception (Marchant de Abreu 2024).
One current obstacle limiting the application of deepfakes is the lack of Urdu integration into generative AI models, which impacts the believability of content (Jahangir 2024). Regardless, AI use in Pakistan exemplifies how deepfakes are used to undermine democratic processes and entrench political rifts (A. Khalid 2023).
Countermeasures
Civic initiatives
Independent journalists and CSOs have played an integral part in challenging disinformation in Pakistan’s political landscape, including through educational programmes to improve the digital literacy of journalists, enabling them to disarm and prevent the spread of disinformation. Other initiatives have been aimed at enhancing media and digital literacy among the general public by raising their awareness and improving their understanding of how disinformation is created and disseminated. Even so, the Pakistani civil liberties organization Freedom Network and the journalist coalition DigiMAP have stated that efforts have seen limited success due to conceptual, technical and financial challenges, in part owing to the absence of a joint strategy and low levels of collaboration between organizations and sectors. Another hurdle is that many efforts to counter disinformation directed at the public are provided in English, limiting their reach due to language barriers (Freedom Network 2022).
In a joint publication, DigiMAP and Freedom Network outlined recommendations for an effective response to disinformation in four areas (Naeem and Rehmat 2022):
- Fact-checking. Organizations need to develop consistent best practices regarding operations and technical tools. They also need to secure financing—remaining open to intersectoral collaboration.
- Journalist training. Training should focus on building conceptual clarity and familiarity with disinformation trends. It should also provide long-term support and follow-up instead of being conducted in brief bursts.
- Media and information literacy. Educational responses should be conducted in a dual format—in part geared towards students and teachers and in part through regular public awareness campaigns.
- Coalition building. Responses should develop a multistakeholder approach that displays consensus and involves exchanges of expertise.
Other members of civil society and academics have called on the authorities to take action, as they are largely responsible for securing adequate resources to target disinformation and raise the public’s awareness of AI-related issues. The situation is especially acute in Pakistan due to the country’s lack of digital infrastructure as well as the low level of awareness and resources available at the governmental level. In responding to this call to action, it is paramount for the government to review Pakistan’s digital resilience and reinforce it by cooperating at an intragovernmental and civil society level (A. Khalid 2023).
National AI policy
In May 2023 Pakistan’s Ministry of Information Technology and Telecommunication introduced a policy draft outlining directives for fostering AI development and adoption. The document identifies the ethical challenges involved in the use of generative AI to create misleading text, images and video and promises the publication of regulatory guidelines as part of a future AI regulation that will ‘address the possible spread of disinformation, data privacy breaches, or fake news’ (Islamic Republic of Pakistan 2022). The draft has received criticism for its piecemeal approach, esoteric wording and lack of concrete action (Z. Khalid 2023).
Regulation tackling disinformation
The Pakistani authorities have come under scrutiny for overregulation and overcriminalization in their efforts to combat disinformation, resulting in severe impacts on civil liberties (Human Rights Watch 2022; Shaheen 2022; RSF 2023). Human rights organizations have criticized the government for taking ‘draconian’ actions in efforts to combat fake news and misinformation. The government passed a 2022 revision of the Pakistan Electronic Crime Act that criminalized online ‘defamation’ of the authorities, imperilling voices of dissent, such as journalists and the political opposition, a group that already faces severe censorship and persecution in the country (Human Rights Watch 2022). The regulation is one of several other restrictions on freedom of speech and privacy online that have been passed in recent years (RSF 2023).
Government monitoring and mitigation of disinformation
Pakistan’s Ministry of Information and Broadcasting hosts a website dedicated to bringing attention to and verifying disinformation (Islamic Republic of Pakistan n.d.). The website directs visitors to an X account which, at the time of the writing of this report, had published its latest verification post in 2022. While previously in office, the PTI also launched an official X account to highlight circulating misinformation. However, media practitioners have underscored the risk that the tool might be misused to target political opponents and voices of dissent (Shaheen 2022).
Government removal of content
The 2021 Removal and Blocking of Unlawful Online Content (Procedure, Oversight and Safeguards) Rules give Pakistan’s online censorship authority (the Pakistan Telecommunication Authority, PTA) a mandate to remove content from social media that is deemed indecent, blasphemous or false (what constitutes ‘false information’ is not specified). It also gives the PTA the authority to remove or block content that could disrupt public order. Compliance with the regulation requires social media companies to remove content that is flagged by the PTA within 12 to 48 hours of receipt of a removal notice; failure to take the necessary action could result in a complete block of the platform in question (Daudpota 2021). The Pakistani Digital Rights Foundation (2020) has denounced the legislation as a threat to citizens’ fundamental rights to free expression and privacy.
Internet shutdowns
Social media platforms are fundamental communication channels for the PTI. Accordingly, platform blocks, Internet shutdowns and throttling have been frequently utilized to censor and deplatform the party. These measures had a significant impact following the 2024 election cycle, since the PTI conducted a contestation campaign through social media. General protests provoked by the accusations of electoral rigging led to nationwide disruptions to X and a shutdown of the PTI’s website (Hussain 2024). These actions followed a trend of previous shutdowns coinciding with planned campaign events, such as the virtual rallies held by the PTI. The shutdowns forced actors affiliated with the PTI to post their communications on alternate, open-source sites, where they published repositories of electoral data to prove the alleged rigging (Hendrix and Jahangir 2024). In addition, these shutdowns were widely criticized for their impact on the public’s ability to access and navigate electoral information, culminating in the online campaign #KeepItOn (AccessNow 2024).
Conclusion
Internet spaces were established as a primary battleground in Pakistan’s highly contested election. Online political discourse is not new to the country, as social media has played a central role in establishing the PTI as the country’s largest party. With Khan’s imprisonment, however, the PTI’s use of deepfakes presents a unique example of how modern technologies can provide alternative means of communicating with the electorate when political actors are faced with censorship. The power of social media in shaping public perceptions is evident from the government’s attempts to control public opinion through censorship and Internet shutdowns. These efforts exemplify how misdirection and disinformation stem not only from an overloaded information landscape but also from the restriction of access to online resources. These issues are exacerbated when combined, which can be observed in the government’s use of both Internet restrictions and deceptive tactics, such as inauthentic fact-checking, coordinated networks of influence or cross-platform amplification.
Introduction
South Africa’s 2024 general election was held on 29 May, resulting in the seventh consecutive victory for the incumbent African National Congress (ANC) (Haddad 2024). However, the ANC lost its parliamentary majority for the first time in the post-apartheid era, necessitating the formation of a coalition with smaller parties. This result, while presaged by polling, was nevertheless met with consternation by financial markets, which feared the radical promises of certain parties, including a commitment by the Economic Freedom Fighters to expropriate land without compensation and a proposal by Jacob Zuma’s uMkhonto weSizwe party (MK) to scrap the constitution itself (Towriss 2024). Central campaign issues included corruption, inequality, service delivery and unemployment, in particular youth unemployment (Ndzendze and Hlabisa 2024). Approximately two weeks following the election, the ANC formed a coalition with the liberal opposition Democratic Alliance and several smaller parties (Danaher 2024). Consequently, incumbent President Cyril Ramaphosa, in office since 2018, was re-elected to a second term (Imray and Magome 2024).
While voter turnout in South Africa had been on the decline for years, the election saw a turnout of only 58.64 per cent—a record low in the country’s 30-year democratic history (Haddad 2024). Potential reasons for this decline include increased political apathy and a decline in voter confidence in the political process (Cleveland and Alexander 2024). This situation is particularly pronounced among youth, who are dissatisfied with and uninformed about the electoral process; moreover, they are uninterested in and disillusioned with politics in general (Gumede 2024).
Digital context
At the start of 2024 there were 45.34 million Internet users in South Africa, standing at 74.7 per cent of the population and placing it in line with the Southern African regional average of 73.1 per cent (Kemp 2024a, 2024f). In 2024, 42.8 per cent of South Africa’s population used social media, equating to 26 million people, or 57.3 per cent of Internet users overall (Kemp 2024f). Meta-owned WhatsApp (93.9 per cent), Facebook (88.4 per cent) and Instagram (71.8 per cent), along with TikTok (73.6 per cent), are the most widely utilized social media apps in the country, constituting important conduits for younger generations to receive political information (McInnes 2024). While gender differences are small, disparities by age are much greater, with those between the ages of 18 and 34 accounting for the majority of social media users in South Africa (Cowling 2024). While narrowing in recent years, a significant rural–urban digital divide is also present, as rural dwellers are less likely to access the Internet and more reliant on mobile phones to do so (Wyrzykowski 2023).
The South African media landscape is broadly characterized by the decline of traditional print media and the simultaneous rise of online news sources. In 2024 two newspapers with long histories in the country ceased operations, and the Audit Bureau of Circulations reported a 7.4 per cent decline in the circulation of its print members; at the same time, several news sites celebrated new subscriber records (Roper 2024). However, television remains the most popular news source, followed by radio and social media (European Union Election Expert Mission 2024). Polling results from the 2024 Digital News Report indicate that, while interest among South Africans in consuming political news dropped from 81 per cent in 2021 to 70 per cent in 2024, concerns about misinformation remain significantly above the global average, at 82 per cent of the population (Newman et al. 2024). Social media remains popular for news consumption, with 82 per cent of respondents consuming short-form online news weekly, and 57 per cent indicating that they trust most of the news that they see most of the time (Newman et al. 2024). Therefore, while interest in political news and print media continues to decline, South Africans continue to engage with news on social media, and while people are concerned about misinformation, they are not in the habit of engaging critically with what they see online.
In response to the perceived lack of accountability by the major digital media platforms, the Competition Commission of South Africa initiated a media and digital platforms market inquiry with the stated intent of investigating potential violations of the Competition Act No. 89 of 1998 (Roper 2024). In the same period, the Film and Publication Board of South Africa proposed new regulations that would hold Internet service providers and social media companies accountable for propagating misinformation (Digital Watch 2024).
While South Africa’s education system does emphasize news literacy, there is little focus on accuracy, and only one province—the Western Cape—has recently introduced more comprehensive misinformation literacy curricula (Cunliffe-Jones 2021).
Disinformation tactics, techniques and procedures
AI-generated content
During the 2024 South African general election, deepfakes and other AI-generated content spread widely on social media platforms. On 9 March 2024 the daughter of former President Zuma shared a video on X in which then-former US President Donald Trump could be seen endorsing the MK party and urging Black voters to reject the ANC (van Damme 2024). The video was likely produced with the app Parrot AI, and while many users rejected its authenticity, its level of sophistication—particularly in the high quality of its audio—nevertheless generated significant uncertainty and confusion.
Coordinated networks of influence
WhatsApp is used extensively in South Africa for sharing political information (Dionne and Liu 2024). Voters are active in groups on the app and share content with one another, magnifying the reach of fake news. On this and other social media apps, influencers—who often have large followings—were paid to share and amplify misinformation during the election, and bot accounts were created and deployed for the same purpose (Allen and le Roux 2024). It is concerning that the main source of information manipulation was found to be political parties, with common narratives including attempts to undermine the credibility of the Electoral Commission and xenophobic rhetoric about migrants (European Union Election Expert Mission 2024).
A network of pro-Russian accounts was uncovered, which propagated fake news stories especially through the use of bot accounts on X (ADDO 2024; Sguazzin 2024). These influencers, trusted messengers and foreign actors largely supported the Economic Freedom Fighters and the MK party in the 2024 election and deployed strategies such as ‘rage baiting’1 and the use of authentic material taken out of context to maintain interest and build user engagement (Allen and le Roux 2024). When President Ramaphosa gave a speech warning that those who were threatening unrest prior to the election would be arrested, a coordinated effort among 13 Facebook accounts amplified this response, driving views, likes and interactions (ADDO 2024). While limited in its overall reach, this action by pro-MK actors contributed to the public discourse by fuelling indignation and outrage at the President’s response.
A significant strategy deployed by purveyors of misinformation was the use of hashtags to fuel polarizing content, an especially concerning technique in South Africa’s post-apartheid context. Through the use of such hashtags as #VoteMK2024 and #2024IsOur1994, accounts were able to magnify the reach of social media posts (ADDO 2024). Sentiment analysis shows that 88.6 per cent of posts on X were negative in their tone and outlook, as accounts frequently coupled hashtags with incendiary narratives that drove user engagement; for example, the hashtag #apartheid was used in posts alleging that the current government was a guardian of the apartheid-era regime (ADDO 2024).
Microtargeted content
By utilizing advanced data analytics, political parties and other organizations are able to tailor advertisements based on demographics and user browsing histories. In 2024 the Democratic Alliance spent far more than any other party on Meta ads, and it primarily targeted people based on their location (Kreuser and Gitonga Theuri 2024). Numerous other non-party-affiliated organizations also engaged in advertising based on age, gender, geography or other factors, producing content with subtle partisan messages. While only 18 per cent of political ads were found to have used microtargeting, they received disproportionately higher visibility than non-targeted ads, demonstrating the significance of its use even when used in a minority of political ads (Kreuser and Gitonga Theuri 2024).
Cross-platform amplification
Major disinformation narratives, such as the notion that the election or the vote-counting process was rigged, regularly began on one social media site before being cross-posted on other sites as well (Allen and le Roux 2024). Cross-posting is an effective disinformation strategy because the various social media apps differ in their reach and algorithms. For instance, while Facebook tends to reinforce viewers’ existing beliefs with content that aligns with their views, TikTok’s algorithm makes it easier to push content from accounts that the user is not following (Allen and le Roux 2024).
In February 2024 an MK Youth League leader gave an incendiary speech in which he promised that MK members would shut down South Africa if the party did not obtain a two-thirds majority in parliament. When 14 Facebook posts amplifying this speech failed to gain traction, 8 video clips were posted to TikTok, where they generated a combined 265,095 views (ADDO 2024).
Manipulating political narratives and framing though social media
Strategic framing of key election-related events propelled misinformation and sowed distrust in electoral institutions. Online complaints submitted to the Real411 platform, which tracks disinformation, indicated that major false narratives focused on incorrect information regarding the election process, parties and candidates, and the election results, as well as information intended to undermine the credibility of the Electoral Commission of South Africa (IEC) (European Union Election Expert Mission 2024). For example, a common theme online during the 2024 election was that the IEC was biased against particular candidates—especially Zuma and other MK party candidates—and would therefore not treat them fairly (Bird and Smith 2024). This narrative led to a wave of content impugning the credibility and integrity of the IEC and its officers, including allegations of vote rigging by the IEC and fraudulent claims that IEC commissioners had been arrested for such activities (Allen and le Roux 2024).
After a campaign of coordinated disinformation that alleged that the election had been rigged and that 9.3 million votes were ‘missing’, the MK party filed a legal challenge with the Electoral Court (Allen and le Roux 2024). While the claim was quickly dismissed, this real-world action built upon the pre-existing social media narratives disputing the integrity of South Africa’s elections.
Countermeasures
Voluntary initiatives to secure commitments and monitor disinformation
In July 2023 the IEC partnered with Google, Meta and TikTok as well as the non-profit organization Media Monitoring Africa to fight disinformation ahead of South Africa’s 2024 elections. These actors signed a Framework of Cooperation that secured commitments to, among other things, act in good faith, respect the data privacy of users and take certain measures against disinformation (IEC 2023). Under the terms of this agreement, an independent three-member commission was established to review reported cases of misinformation and to make recommendations to the IEC, which could in turn request that malicious content be removed. However, the agreement is non-binding, and the platform X is not party to the agreement (Ndlovu 2024). In addition, while feed-based apps like Facebook and X are able to take some action against the spread of misinformation—such as instituting community notes and user reporting—WhatsApp is not. The latter application is a closed platform that operates through private, encrypted messaging, making it impossible to track the flow of misinformation (Dionne and Liu 2024).
In the lead-up to the 2024 election, Meta established a South Africa–specific elections operations centre to identify threats in real time (Idé Siddo 2024). This initiative brought together experts from both Meta and South Africa to identify ways to better identify and combat misinformation. An election resource centre was also established to provide users with information on how to secure their accounts and identify trusted sources of information (Idé Siddo 2024). Furthermore, both Meta and TikTok engaged with independent CSOs, such as Media Monitoring Africa and Code for Africa, on best practices to combat misinformation and uphold free and fair elections (Deenik 2024). A review of social media platforms’ efforts revealed that more fact-checkers and monitors who are fluent in South Africa’s various official languages are needed to enhance early detection of misinformation, as well as more efforts to counter cross-platform misinformation (Ngubane n.d.; European Union Election Expert Mission 2024).
The IEC created the Real411 platform in the context of its Framework of Cooperation with social media platforms as a one-stop complaints portal, administered by the non-governmental organization Media Monitoring Africa, for the public to report online disinformation (European Union Election Expert Mission 2024). This innovative tool both acts as an early-warning system and provides specific information on which news items are most closely tied to false content, providing fact-checkers and digital literacy educators with key data on the most helpful approaches (Bird and Smith 2024). The platform received approximately 200 complaints, 47 per cent of which were deemed information manipulation (European Union Election Expert Mission 2024).
AI-powered tools to combat disinformation
Fact-checking organizations, such as Africa Check, deployed AI-powered tools to counter false claims made during the election campaign (Ndhlovu 2024). In collaboration with the Google News Initiative, Africa Check provided tools such as an alerts function, which flagged fake news on social media; a search function, which allowed journalists to more easily find factual claims; and a transcription function for furnishing accurate records of campaign speeches (Ndlovu 2024).
Legislative and regulatory measures
The Cybercrimes Act of 2020 provides that anyone found to have knowingly disseminated false or defamatory content may be held civilly liable for defamation (Grange 2024). However, this measure has had limited success in combating the misinformation that is rife on social media platforms, which are not subject to a clear regulatory framework during election periods (European Union Election Expert Mission 2024). In addition, the South African Film and Publication Board intends to target misinformation by holding social media companies liable for fake news content (Digital Watch 2024). The board further plans to collaborate with the South African Artificial Intelligence Association to develop regulations on the use of AI on social media apps (Illidge 2024). While the information regulator issued a guidance note for the use of personal data by political parties for the 2019 election, no such guidance was released ahead of the 2024 election, contributing to legal and regulatory gaps in party political communications online (European Union Election Expert Mission 2024).
Digital literacy initiatives
A March 2022 report by Africa Check stated that, while media literacy is part of the basic curriculum in South African schools, there is insufficient focus on how to fact-check and verify the media. However, the Western Cape Education Department, in partnership with Google, has introduced a more comprehensive media literacy module about online safety (Wasserman, Madrid-Morales and Rabin 2022). Non-governmental organizations, such as Media Monitoring Africa, have created their own digital literacy initiatives to engage the public on ways to better spot fake news. One such example is Web Rangers, a digital literacy project established in 2016 to enable young people to gain critical knowledge about online safety, and the project expanded its tools for parents in 2024 (Media Monitoring Africa 2024).
In addition, South Africa has developed Digital Skills Framework One (DSFOne), a national digital skills competency framework that aligns with the European Union’s Digital Competence Framework for Citizens, which defines digital skills across basic, intermediate and advanced levels. To address the low level of digital inclusion, the government has developed the 2020 Digital and Future Skills Strategy, which promotes work readiness, social participation and targeted support for youth, women and persons with disabilities. Key initiatives include the Digital Skills Forum, public–private partnerships, workplace-based learning, and policy coordination through the Presidential Youth Employment Initiative to bridge digital skills gaps and enhance economic opportunities (Domingo et al. 2024).
Conclusion
The 2024 South African general election was marked by significant electoral disinformation, driven by the widespread use of social media, AI-generated content and coordinated influence campaigns. The digital landscape, characterized by high Internet and social media penetration, played a crucial role in amplifying misinformation, particularly through WhatsApp groups, bot networks and influencer endorsements. Tactics such as deepfakes, microtargeting and cross-platform amplification contributed to voter confusion and distrust in electoral institutions, with claims of vote rigging and institutional bias fuelling online disinformation and institutional mistrust. In response, countermeasures were implemented, including cooperative agreements between the Electoral Commission of South Africa and major social media platforms and AI-powered fact-checking tools. However, gaps remain, particularly in establishing more comprehensive detection and monitoring strategies and addressing the regulatory challenges posed by misinformation. The 2024 election underscored the urgent need for more robust policies, greater transparency on the part of digital platforms and enhanced public education efforts to safeguard electoral integrity in the digital age.
Abubakar, I., ‘Cyber Security Authority educates journalists on misinformation and disinformation’, 3News, 10 October 2024, <https://3news.com/news/cyber-security-authority-educates-journalists-on-misinformation-and-disinformation>, accessed 8 March 2025
Access Now, ‘#KeepItOn in Pakistan: Ensure open and secure Internet for all during the 2024 elections’, 6 February 2024, <https://www.accessnow.org/press-release/keepiton-pakistan-2024-elections>, accessed 25 February 2025
Adiya, A., ‘What voters think: 2024 Mongolia elections’, Mongolia Weekly, 22 June 2024, <https://www.mongoliaweekly.org/post/what-voters-think-2024-mongolia-elections>, accessed 17 February 2025
Africa Check, ‘MFWA, Dubawa and Ghana Fact establish Ghana Fact-Checking Coalition ahead of December elections’, 19 November 2024, <https://africacheck.org/fact-checks/blog/mfwa-dubawa-and-ghana-fact-establish-ghana-fact-checking-coalition-ahead-december>, accessed 8 March 2025
African Digital Democracy Observatory (ADDO), ‘Unravel South Africa’s election chaos and Zuma’s influence’, Medium, 9 May 2024, <https://disinfo.africa/unravel-south-africas-election-chaos-and-zuma-s-influence-b9e9880ff3d3>, accessed 11 February 2025
African Union Election Observer Mission, ‘Preliminary Statement of the African Union Election Observation Mission to the Republic of Ghana’, 9 December 2024,, <https://www.peaceau.org/en/article/preliminary-statement-of-the-african-union-election-observation-mission-to-the-republic-of-ghana-9th-december-2024>, accessed 3 March 2025
Agence France-Presse (AFP), ‘Fake news on the rise ahead of Mexico elections’, France 24, 14 May 2021, <https://www.france24.com/en/live-news/20210524-fake-news-on-the-rise-ahead-of-mexico-elections>, accessed 15 May 2025
—, ‘Fake experts drive disinformation before Bangladesh polls’, France 24, 7 September 2023, <https://www.france24.com/en/live-news/20230907-fake-experts-drive-disinformation-before-bangladesh-polls>, accessed 5 March 2025
Aguilar, A., ‘A Sheinbaum le preocupan más los bots que las víctimas, critica Xóchitl Gálvez’ [Sheinbaum is more concerned about bots than victims, criticizes Xóchitl Gálvez], Animal Político, 21 February 2024, <https://animalpolitico.com/elecciones-2024/presidencia/sheinbaum-preocupan-mas-bots-victimas-xochitl-galvez>, accessed 15 May 2025
Aguirre, S., ‘¡Ay, no! La desinformación ya llegó a Telegram: Este es el “modus operandi”’ [Oh no! Disinformation has already reached Telegram: This is the modus operandi], Animal Político, 24 March 2021, <https://www.animalpolitico.com/tendencias/actualidad/canales-desinformacion-telegram-como-funcionan>, accessed 17 May 2025
Ali, F., ‘How deepfakes, AI and fake news impacted Pakistan’s elections 2024’, Samaa TV, 12 February 2024, <https://www.samaa.tv/208739652-how-deepfakes-ai-and-fake-news-impacted-pakistan-s-elections-2024>, accessed 21 February 2025
Allen, K. and le Roux, J., ‘Under the Influence? Online Mis/Disinformation in South Africa’s May 2024 Election’, Institute for Security Studies, Southern Africa Report No. 61, December 2024, <https://issafrica.s3.amazonaws.com/uploads/pages/1734073957062-sar-61.pdf>, accessed 11 February 2025
AISur, ‘Los Caminos Regulatorios para la IA en América Latina: Recopilación de Estudios de Caso en Brasil, México, Perú y Colombia’ [Regulatory Pathways for AI in Latin America: Compilation of Case studies from Brazil, Mexico, Peru and Colombia], May 2024, <https://www.alsur.lat/sites/default/files/2024-10/ALSUR%20-%20IA%20en%20Latam%20%5BESP%5D-1.pdf>, accessed 8 April 2025
American Chamber of Commerce in Mongolia, ‘AmCham Position Paper: Effective Internet Governance’, October 2023, <https://amcham.mn/storage/advocacies/October2023/gTbAbFf75DRA0JGz3yTX.pdf>, accessed 16 February 2025
AMLO, ‘Internet para Todos reduce brecha digital; México, uno de los países más conectados en América Latina’ [Internet for All reduces the digital divide; Mexico is one of the most connected countries in Latin America], 4 July 2024, <https://lopezobrador.org.mx/2024/07/04/internet-para-todos-reduce-brecha-digital-mexico-uno-de-los-paises-mas-conectados-en-america-latina>, accessed 21 June 2025
Amnesty International, ‘Bangladesh: Government must remove draconian provisions from the Draft Cyber Security Act’, 31 August 2023, <https://www.amnesty.org/en/latest/news/2023/08/bangladesh-government-must-remove-draconian-provisions-from-the-draft-cyber-security-act>, accessed 6 March 2025
Animal Político, ‘Morena se queja en el INE contra Xóchitl Gálvez por presunto uso de bots financiado desde el extranjero contra AMLO’ [Morena files complaint with INE against Xóchitl Gálvez for alleged use of foreign-funded bots targeting AMLO], 22 February 2024, <https://www.animalpolitico.com/elecciones-2024/presidencia/morena-queja-ine-xochitl-galvez-bots-financiado-extranjero>, accessed 15 May 2025
Ankhtuya, B., ‘Mongolia punishes fake news spreaders by criminal law’, News.mn, 30 March 2020, <https://news.mn/en/791544>, accessed 19 February 2025
ANTARA, ‘Ministry, Bawaslu form task force to monitor online campaigns’, 6 August 2023a, <https://en.antaranews.com/news/290349/ministry-bawaslu-form-task-force-to-monitor-online-campaigns>, accessed 19 February 2025
—, ‘Govt provides AI training to people through digital literacy movement’, 7 December 2023b, <https://en.antaranews.com/news/300489/govt-provides-ai-training-to-people-through-digital-literacy-movement>, accessed 20 February 2025
—, ‘Govt to prepare special AI regulation’, 22 December 2023c, <https://en.antaranews.com/news/301644/govt-to-prepare-special-ai-regulation>, accessed 19 February 2025
—, ‘Taking comprehensive measures to tackle election hoaxes: Ministry’, 17 January 2024a, <https://en.antaranews.com/news/303282/taking-comprehensive-measures-to-tackle-election-hoaxes-ministry>, accessed 19 February 2025
—, ‘Ministry coordinates with social media platforms ahead of elections’, 31 January 2024b, <https://en.antaranews.com/news/304554/ministry-coordinates-with-social-media-platforms-ahead-of-elections>, accessed 19 February 2025
Asian Network for Free Elections (ANFREL), Beyond Power: Upholding the Rule of Law amid Democratic Adversities: Final Report of the ANFREL International Expert Election Observation Mission to the 2024 Indonesian General Elections (Bangkok: ANFREL, 2024a), <https://anfrel.org/2024-indonesian-general-elections-beyond-power-upholding-the-rule-of-law-amid-democratic-adversities>, accessed 7 April 2025
—, ‘2024 Pakistani general election: FAFEN analysis of voter turnout’, 14 February 2024b, <https://anfrel.org/2024-pakistani-general-election-fafen-analysis-of-voter-turnout>, accessed 13 March 2025
Associated Press (AP), ‘Post-communist generation is hoping for a new era of democracy in Mongolia’, NY1, 5 July 2024, <https://ny1.com/nyc/all-boroughs/ap-top-news/2024/07/05/post-communist-generation-is-hoping-for-a-new-era-of-democracy-in-mongolia>, accessed 17 February 2025
Baatar, D., ‘Mongolia: Making the news media more transparent and viable’, DW Akademie, 19 May 2023, <https://akademie.dw.com/en/mongolia-making-the-news-media-more-transparent-and-viable/a-65579930>, accessed 19 February 2025
Bahl, V., ‘Pakistan political turmoil: These “PTI protest” images aren’t real’, France 24, 17 May 2023, <https://www.france24.com/en/tv-shows/truth-or-fake/20230517-pakistan-political-turmoil-these-pti-protest-images-aren-t-real>, accessed 14 February 2024
Baksh, S. W., Mungra, D., Ahasan, N., Pariwala, V., Jain, R., Surve, A., Banerjee, P., Mustofi Eron, S., Mehta, S. and Rashid Diya, S., ‘How Facebook Has Become a Political Battleground in Bangladesh’, Tech Global Institute, 19 December 2023, <https://infolab.techglobalinstitute.com/how-facebook-has-become-a-political-battleground-in-bangladesh>, accessed 4 March 2025
Bangladesh, People’s Republic of, Digital Security Act 2018, 8 October 2018, <https://natlex.ilo.org/dyn/natlex2/natlex2/files/download/110029/BGD110029.pdf>, accessed 17 April 2025
Bangladesh Awami League, ‘Awami League plans big social media campaign for upcoming election’, 15 October 2023, <https://albd.org/articles/news/41092/Awami-League-plans-big-social-media-campaign-for-upcoming-election>, accessed 4 March 2025
Bangladesh Telecommunication Regulatory Commission, ‘Regulation for Digital, Social Media and OTT Platforms’, October 2021, <https://btrc.gov.bd/sites/default/files/files/btrc.portal.gov.bd/notices/0031100b_c62f_46eb_9ce8_317e53ac881b/2022-02-06-04-33-68c9c154e5319e6e9179af538b3e47cb.pdf>, accessed 15 June 2025
Banzragch, O., Mashlai, B. and Mavag, C., ‘Mongolia Comprehensive National-Level Report on the Implementation of the Beijing Declaration and Platform of Action (2019-2024)’, Ministry of Labour and Social Protection, Government of Mongolia, August 2024, <https://www.unwomen.org/sites/default/files/2024-09/b30_report_mongolia_en.pdf>, accessed 16 February 2025
Barbosa, M., ‘Cómo identificar (y combatir) el discurso de odio en las elecciones 2024’ [How to identify (and combat) hate speech in the 2024 elections], Verificado, 4 March 2024a, <https://verificado.com.mx/como-identificar-y-combatir-el-discurso-de-odio-en-las-elecciones-2024>, accessed 16 May 2025
—, ‘El peligro de las fake media en las elecciones: qué son y cómo operan’ [The danger of fake media in elections: What they are and how they operate], Verificado, 16 April 2024b, <https://verificado.com.mx/fake-media-elecciones-que-son-y-como-operan>, accessed 16 May 2025
Batdorj, B., ‘Social media and public discourse in Mongolia’, Friedrich-Ebert-Stiftung, 27 May 2020, <https://asia.fes.de/news/social-media-and-public-discourse-in-mongolia.html>, accessed 18 February 2025
Batres, M. (@martibatres), ‘Está circulando un audio en el que se pretende atribuirme expresiones relacionadas ...’ [An audio [recording] is circulating in which statements are falsely attributed to me regarding ...], X, 1 November 2023, <https://x.com/martibatres/status/1719541442408792305>, accessed 17 May 2025
Battsengel, B., Sukhee, T., Ariunsan, G., Togtokhbayar, A. and Purevdorj, B., ‘Mongolia in the Digital Age: The Digital Readiness Assessment’, Access Solutions LLC, November 2019, <https://pathwayscommission.bsg.ox.ac.uk/sites/default/files/2019-11/Mongolia_Digital_Readiness_Assessment.pdf>, accessed 16 February 2025
Battulga, B., ‘Real-life success: Overcoming COVID-19 through healthy life style and vaccination: How exercise and vaccination can help us win COVID-19’, UNICEF Mongolia, 1 February 2023, <https://www.unicef.org/mongolia/stories/real-life-success-overcoming-covid-19-through-healthy-life-style-and-vaccination>, accessed 19 February 2025
Belinda, Y., ‘Is K-Pop Indonesian political parties’ latest tool to entice young voters?’, The Jakarta Post, 21 March 2023, <https://www.thejakartapost.com/culture/2023/03/21/is-k-pop-indonesian-political-parties-latest-tool-to-entice-young-voters.html>, accessed 13 February 2025
Bird, W. and Smith, T., ‘Real411 is ready to process your complaints about election misinformation and disinformation’, Daily Maverick, 5 May 2024, <https://www.dailymaverick.co.za/article/2024-05-05-2024-electionstackling-misinformation-disinformation>, accessed 11 February 2025
Boakye, R. J. Y., Bchir, M. and Missoh, A., ‘What Ghana’s presidential election means for the country and West Africa’, United States Institute of Peace, 19 December 2024, <https://www.usip.org/publications/2024/12/what-ghanas-presidential-election-means-country-and-west-africa>, accessed 3 March 2025
Bodine, A., ‘Semiliterate Ghanaians learn media and information literacy’, DW Akademie, 2 December 2022, <https://akademie.dw.com/en/mil-for-illiterate-ghanaians/a-63955851>, accessed 8 March 2025
Boldkhuyag, P., Batmunkh, O., Tserensodnom, L., Ikhagvasuren, G. and Munkhtugs, O., ‘Media Freedom Report 2020’, Globe International Center, 3 May 2021, <https://www.gic.mn/public/docs/freedom_report/media_freedom_report_2020_en.pdf>, accessed 16 February 2025
Bradshaw, S., Campbell-Smith, U., Henle, A., Perini, A., Shalev, S., Bailey, H. and Howard, P. N., ‘Country Case Studies. Industrialized Disinformation: 2020 Global Inventory of Organized Social Media Manipulation’, Oxford Internet Institute, 2021, <https://demtech.oii.ox.ac.uk/wp-content/uploads/sites/12/2021/03/Case-Studies_FINAL.pdf>, accessed 14 February 2025
Bratajaya, Y. and Tan, D., ‘Criminalization of disinformation threatens freedom of expression’, The Jakarta Post, 27 November 2023, <https://www.thejakartapost.com/opinion/2023/11/27/criminalization-of-disinformation-threatens-freedom-of-expression.html>, accessed 13 February 2025
Breuer, A., ‘Information Integrity and Information Pollution: Vulnerabilities and Impact on Social Cohesion and Democracy in Mexico’, German Institute of Development and Sustainability, IDOS Discussion Paper 2/2024, <https://www.idos-research.de/uploads/media/DP_2.2024.pdf>, accessed 8 April 2025
Broderick, R. and Arrendondo, Í., ‘Meet the 29-year-old trying to become the king of Mexican fake news’, Buzzfeed News, 28 June 2018, <https://www.buzzfeednews.com/article/ryanhatesthis/meet-the-29-year-old-trying-to-become-the-king-of-mexican>, accessed 17 May 2025
Business Standard, The, ‘“Digital Security Act succeeds in creating culture of fear”’, 14 January 2023, <https://www.tbsnews.net/bangladesh/digital-security-act-succeeds-creating-culture-fear-speakers-567734>, accessed 7 March 2025
Calderón, M., ‘Political Violence in Mexico’s 2024 Election: Gender-Based Political Violence against Women’, The Wilson Center, March 2024, <https://www.wilsoncenter.org/sites/default/files/media/uploads/documents/Gender-Based%20Political%20Violence%20Against%20Women.pdf>, accessed 8 April 2025
Castro, C., ‘Bangladesh elections: Censorship, AI deepfakes, and social media polarization’, TechRadar, 9 January 2024, <https://www.techradar.com/computing/cyber-security/bangladesh-elections-censorship-ai-deepfakes-and-social-media-polarization>, accessed 4 March 2025
Cerdeira, P., ‘¿Cómo funcionan los ‘Bots’ y cómo saber quiénes están detrás de ellos?’ [How do bots work, and how can we find out who’s behind them?], MVS Noticias, 22 February 2024, <https://mvsnoticias.com/entrevistas/2024/2/22/como-funcionan-los-bots-como-saber-quienes-estan-detras-de-ellos-627813.html>, accessed 15 May 2025
Civicus Monitor, ‘Bangladesh: Criminalisation of activists and crackdown on protests continue following one-sided elections’, 12 April 2024, <https://monitor.civicus.org/explore/bangladesh-criminalisation-of-activists-and-crackdown-on-protests-continue-following-one-sided-elections>, accessed 7 April 2025
Cleveland, G. and Alexander, W., ‘ANC loses majority after 30 years’, Friedrich Naumann Foundation for Freedom, 3 June 2024, <https://www.freiheit.org/liberal-workshop-south-africa/anc-loses-majority-after-30-years>, accessed 11 February 2025
Congressional Research Service, ‘Parliamentary Elections in Mongolia’, 9 July 2024, <https://crsreports.congress.gov/product/pdf/IN/IN12390>, accessed 15 February 2025
Cowling, N., ‘Distribution of social media users in South Africa as of January 2024, by age group and gender’, Statista, 23 February 2024, <https://www.statista.com/statistics/1100988/age-distribution-of-social-media-users-south-africa>, accessed 11 February 2025
Cunliffe-Jones, P., ‘Misinformation Policy in Sub-Saharan Africa: From Laws and Regulations to Media Literacy’, Communication and Media Research Institute, CAMRI Policy Briefs and Reports No. 7, 2021, <https://library.oapen.org/bitstream/handle/20.500.12657/50175/9781914386053.pdf>, accessed 11 February 2025
Daily Star, The, ‘Registration of online portals of newspapers, TV channels: TIB warns of govt control over media’, 1 September 2020, <https://www.thedailystar.net/registration-online-versions-newspapers-tv-news-portals-tib-warns-govt-control-over-media-1954193>, accessed 6 March 2025
—, ‘BTRC blocks BSS while weeding illegal news sites out’, 29 September 2021, <https://www.thedailystar.net/news/bangladesh/crime-justice/news/btrc-botches-attempt-comply-hc-order-2186196>, accessed 7 March 2025
Danaher, C., ‘South Africa to form multi-party coalition government in historic “new chapter,” DA party says’, CNN, 14 June 2024, <https://www.cnn.com/2024/06/14/africa/south-africa-governing-alliance-emerges-intl/index.html>, accessed 11 February 2025
Darko, S. A., ‘Ghana’s law on publication of false news is vague and easily abused’, Africa at LSE, 25 February 2022, <https://blogs.lse.ac.uk/africaatlse/2022/02/25/ghana-law-publication-of-fake-news-vague-easily-abused-free-speech>, accessed 8 March 2025
Daudpota, F., ‘Understanding Pakistan’s 2021 social media rules: Suppression of constitutional right to freedom of expression through prior restraint censorship of online content’, last revised 29 November 2021, <http://dx.doi.org/10.2139/ssrn.3597721>
Dawn, ‘Troll armies’, 16 August 2023, <https://www.dawn.com/news/1770418>, accessed 20 February 2025
Deenik, J.-A., ‘Disinformation, big tech, and the 2024 elections’, Legal Resources Centre, 22 October 2024, <https://lrc.org.za/disinformation-big-tech-and-the-2024-elections>, accessed 11 February 2025
Delgado, M. (@mario_delgado), ‘Compartimos las cartas dirigidas a las plataformas @X, @Meta y @TikTokLatAM ...’ [We are sharing the letters sent to the platforms @X, @Meta, and @TikTokLatAM ...], X, 7 March 2024, <https://twitter.com/mario_delgado/status/1765846710897483946>, accessed 17 May 2025
Del Pozo, C. M. and Arroyo, D. R., ‘Generative AI and Its Influence on Mexico’s 2024 Elections’, Friedrich Naumann Foundation for Freedom, February 2025, <https://www.freiheit.org/publikation/generative-ai-and-its-influence-mexicos-2024-elections-0>, accessed 8 April 2025
Dewantara, J. A., Syamsuri, S., Wandira, A., Afandi, A., Hartati, O., Cahya, N., Sapitri, P. and Nurgiansah, T. H., ‘The role of buzzers in social media in guiding public opinion regarding political choices’, Journal of Etika Demokrasi, 7/3 (2022), pp. 681–94, <https://doi.org/10.26618/jed.v7i3.8103>
DFRLab, ‘#ElectionWatch: Down ballot bots in Mexico’, 7 June 2018, <https://medium.com/dfrlab/electionwatch-down-ballot-bots-in-mexico-e1bee023291d>, accessed 15 May 2025
Dierkes, J., ‘Mongolia’s MPP maintains power with strategic gains’, East Asia Forum, 23 July 2024, <https://doi.org/10.59425/eabc.1721772000>
Digital Literacy Center, [n.d.], <https://digitalliteracy.gov.bd>, accessed 6 March 2025
Digital Rights Foundation, ‘Removal and Blocking of Unlawful Online Content (Procedure, Oversight and Safeguards Rules, 2020: Legal Analysis)’, 30 November 2020, <https://digitalrightsfoundation.pk/wp-content/uploads/2020/12/Removal-and-Blocking-of-Unlawful-Online-Content-Procedure-Oversight-and-Safeguards-Rules-2020_-Legal-Analysis.pdf>, accessed 21 June 2025
Digital Watch, ‘South Africa’s proposed misinformation regulations spark debate’, 10 April 2024, <https://wp.dig.watch/updates/south-africas-proposed-misinformation-regulations-spark-debate>, accessed 11 February 2025
Digitally Right and International Media Support, ‘Bangladesh: A Mapping of Innovative Media’, December 2021, <https://www.mediasupport.org/wp-content/uploads/2022/01/BD-Media-Mapping-Report-Short.pdf>, accessed 15 June 2025
Dionne, K. Y. and Liu, S., ‘Understanding the fake news about South Africa’s elections’, Good Authority, 27 May 2024, <https://goodauthority.org/news/disinformation-misinformation-fake-news-about-south-africas-2024-elections>, accessed 11 February 2025
Djajadiningrat, V. P., ‘Indonesia digital policy against distribution of information disorder: A better policy for sustainable society’, ES3 Web of Conferences, 426 (2023), <https://doi.org/10.1051/e3sconf/202342602029>
Domingo, E., Muscat, S., Arnold, S., Salzinger, M. and Chetty, P., ‘The Geopolitics of Digital Literacy and Skills Cooperation with Africa’, European Centre for Development Policy Management, Discussion Paper No. 369, June 2024, <https://ecdpm.org/application/files/8717/1921/9245/The-geopolitics-digital-literacy-skills-coorperation-with-Africa-ECDPM-Discussion-Paper-369-2024.pdf>, accessed 11 February 2025
Duke Reporters’ Lab, Fact-checking sites around the world, [n.d.], <https://reporterslab.org/fact-checking>, accessed 16 May 2024
Echeverría, M., ‘Antecedent conditions of disinformation in Mexico: A theoretical framework’, Proceedings of the International Conference on Communication and Applied Technologies (2023), pp. 332–43, <https://doi.org/10.2991/978-94-6463-254-5_33>
Economist, The, ‘Tiktok is a key battleground in Indonesia’s election’, 1 February 2024, <https://www.economist.com/asia/2024/02/01/tiktok-is-a-key-battleground-in-indonesias-election>, accessed 15 February 2025
Economist Impact, Inclusive Internet Index, [n.d.], <https://impact.economist.com/projects/inclusive-internet-index/2022>, accessed 23 February 2025
Eddymurthy, I. A. and Sihombing, A. K., ‘Legal update: Indonesia takes big step forward with new revision of electronic information and transactions law’, Lexology, 17 January 2024, <https://www.lexology.com/library/detail.aspx?g=ed376f51-5ed6-4610-8cef-150c61651bbf>, accessed 7 April 2025
Electoral Commission of Ghana, 2024 General Election Results, 2024, <https://ec.gov.gh/2024-election-results>, accessed 20 June 2025
Electoral Commission of South Africa (IEC), ‘Electoral Commission partners with social media giants to combat disinformation in 2024 national and provincial elections’, 3 July 2023, <https://www.elections.org.za/pw/News-And-Media/News-List/News/News-Article/Electoral-Commission-partners-with-social-media-giants-to-combat-disinformation-in-2024-National-and-Provincial-Elections?a=AISDGvpz75ps1usOfX7oimHCQG6/AToNAzCQK374oSg=>, accessed 11 February 2025
Embassy of the Czech Republic in Ulaanbaatar, Media in Mongolia, [n.d.], <https://mzv.gov.cz/ulaanbaatar/en/media_and_presentation/media_in_mongolia/index.html>, accessed 16 February 2025
European Union Election Expert Mission,’ South Africa, National and Provincial Elections, 29 May 2024, Final Report’, June 2024, <https://www.eods.eu/library/EU%20EEM%20ZAF%202024%20FR.pdf>, accessed 11 February 2025
Fakultas Hukum Universitas Indonesia, ‘Negative campaigns and black campaigns: What’s the difference?’, 10 October 2024, <https://law.ui.ac.id/negative-campaigns-and-black-campaigns-whats-the-difference>, accessed 7 April 2025
Fikri, C., ‘Jokowi’s approval rating hits 76% in LSI survey’, Jakarta Globe, 10 December 2023, <https://jakartaglobe.id/news/jokowis-approval-rating-hits-76-in-lsi-survey>, accessed 12 February 2025
Firdaus, A., ‘Fake Suharto video fuels debate on AI use in Indonesian election campaign’, Benar News, 12 January 2024, <https://www.benarnews.org/english/news/indonesian/suharto-deepfake-used-in-election-campaign-01122024135217.html>, accessed 13 February 2025
Freedom House, ‘Freedom on the Net 2021: Pakistan’, 2021, <https://freedomhouse.org/country/pakistan/freedom-net/2021>, accessed 21 March 2025
—, ‘Freedom on the Net 2023: Bangladesh’, 2023a, <https://freedomhouse.org/country/bangladesh/freedom-net/2023>, accessed 4 March 2025
—, ‘Freedom on the Net 2023: Indonesia’, 2023b, <https://freedomhouse.org/country/indonesia/freedom-net/2023>, accessed 7 April 2025
—, ‘Freedom on the Net 2023: Mexico’, 2023c, <https://freedomhouse.org/country/mexico/freedom-net/2023>, accessed 15 May 2025
Freedom Network, ‘Countering disinformation efforts need urgent boost in Pakistan’, 13 December 2022, <https://www.fnpk.org/countering-disinformation-efforts-need-urgent-boost-in-pakistan>, accessed 19 February 2025
Gamboa, V., ‘Nace “iXóchitl”, vocera especial de la precandidate, creada con inteligencia artificial’ [Introducing ‘iXóchitl’: An AI-generated spokesperson for the presidential pre-candidate], El Universal, 17 December 2023, <https://www.eluniversal.com.mx/elecciones/nace-ixochitl-vocera-especial-de-la-precandidata-creada-con-inteligencia-artificial>, accessed 15 May 2025
Ghana, Republic of, Ministry of Communication, Digital Technology and Innovations, ‘Ghana IGF 2024: Empowering Sustainable Digital Transformation for Inclusive Growth’, 6 September 2024a, <https://moc.gov.gh/2024/09/06/ghana-igf-2024-empowering-sustainable-digital-transformation-for-inclusive-growth>, accessed 6 March 2025
—, Ministry of Communications and Digitalisation, ‘Ghana Digital Economy Policy and Strategy 2024’ (2024b), <https://nita.gov.gh/theevooc/2024/12/Ghana-Digital-Economy-Policy-Strategy-Document.pdf>, accessed 6 March 2025
Ghana Cyber Security Authority, About Us, [n.d.], <https://www.csa.gov.gh>, accessed 6 March 2025
Ghana Fact-Checking Coalition, ‘Countering Electoral Disinformation: Lessons from Ghana’s 2024 Elections’, January 2025, <https://www.fact-checkghana.com/wp-content/uploads/2025/02/GFC-Report-Brochure-Final-View-1.pdf>, accessed 6 March 2025
Ghana News Agency, ‘Elections 2024: UNESCO trains journalists in countering hate speech, misinformation on social media’, BusinessGhana, 30 June 2024, <https://www.businessghana.com/site/news/general/309430/Elections-2024%3A-UNESCO-trains-journalists-in-countering-hate-speech%2C-misinformation-on-social-media>, accessed 8 March 2025
Glowacki, M., Narayanan, V., Maynard, S., Hirsch, G., Kollanyi, B., Neudert, L.-M., Howard, P., Lederer, T. and Barash, V., ‘News and Political Information Consumption in Mexico: Mapping the 2018 Mexican Presidential Election on Twitter and Facebook’, Oxford Internet Institute, 29 June 2018, <https://demtech.oii.ox.ac.uk/wp-content/uploads/sites/12/2018/06/Mexico2018.pdf>, accessed 21 June 2025
Grange, H., ‘Combatting fake news and disinformation is a civil duty’, Good Governance Africa, 19 January 2024, <https://gga.org/combatting-fake-news-and-disinformation-is-a-civil-duty>, accessed 11 February 2025
Gul, A., ‘Official resigns, admits tampering in Pakistan’s controversy-marred vote’, Voice of America, 17 February 2024, <https://www.voanews.com/a/official-resigns-admits-tampering-in-pakistan-s-controversy-marred-vote-/7491919.html>, accessed 20 February 2025
Gumede, N., ‘Trust deficit drives youth voter apathy’, Africa Practice, 13 March 2024, <https://africapractice.com/trust-deficit-drives-youth-voter-apathy>, accessed 11 February 2025
Gutiérrez Aranda, V., ‘Identifician “bots” de difamación contra Sheinbaum; ya se prepara denuncia’ [Defamation bots targeting Sheinbaum identified, complaint already being prepared], Tribuna, 20 February 2024, <https://www.tribuna.com.mx/mexico/2024/2/20/identifican-bots-de-difamacion-contra-sheinbaum-ya-se-prepara-denuncia-361247.html>, accessed 17 May 2025
Haddad, M., ‘South Africa elections final results: What happens next?’, Al Jazeera, 2 June 2024, <https://www.aljazeera.com/news/2024/6/2/south-africa-elections-results-what-happens-next>, accessed 11 February 2025
Harbowo, N., ‘Waiting for Constitutional Court judges to become “funnels” of justice’, Kompas, 19 April 2024, <https://www.kompas.id/baca/english/2024/04/19/en-menantikan-hakim-jadi-corong-keadilan>, accessed 7 April 2025
Harmadi, S. H. B, ‘2024 voter demographics’, Kompas, 28 November 2023, <https://www.kompas.id/baca/english/2023/11/27/en-demografi-pemilih-2024>, accessed 13 February 2025
Haroon, D., Arif, H., Tariq, A. T., Nawaz, F., Qazi, I. A. and Mustafa, M., ‘Investigating misinformation dissemination on social media in Pakistan’, arXiv, 17 June 2021, <https://arxiv.org/abs/2106.09338>
Haskins, C., ‘Phony X accounts are meddling in Ghana’s election’, Rest of World, 15 November 2024, <https://restofworld.org/2024/ghana-election-ai-bots-x-twitter>, accessed 6 March 2025
Hendrix, J. and Jahangir, R., ‘Pakistan and the intersection of tech and elections’, Tech Policy Press, 24 February 2024, <https://www.techpolicy.press/pakistan-and-the-intersection-of-tech-elections>, accessed 24 February 2025
Human Rights Watch, ‘Bangladesh: Repeal abusive law used in crackdown on critics’, 1 July 2020, <https://www.hrw.org/news/2020/07/01/bangladesh-repeal-abusive-law-used-crackdown-critics>, accessed 6 March 2025
—, ‘Pakistan: Repeal amendment to draconian cyber law’, 28 February 2022, <https://www.hrw.org/news/2022/02/28/pakistan-repeal-amendment-draconian-cyber-law>, accessed 19 February 2025
Hussain, A., ‘X faces restrictions in Pakistan amid protests over alleged vote rigging’, 19 February 2024, Al Jazeera, <https://www.aljazeera.com/news/2024/2/19/x-faces-restrictions-in-pakistan-amid-protests-over-alleged-vote-rigging>, accessed 21 February 2025
Iannone, A., ‘Social media and the 2024 election: Right down to the buzzer’, Indonesia at Melbourne, 14 February 2024, <https://indonesiaatmelbourne.unimelb.edu.au/social-media-and-the-2024-election-right-down-to-the-buzzer>, accessed 14 February 2025
Idé Siddo, B., ‘How Meta is preparing for the 2024 South African elections’, Meta, 12 April 2024, <https://about.fb.com/news/2024/04/how-meta-is-preparing-for-the-2024-south-african-elections>, accessed 11 February 2025
Idris, I., ‘Indonesia’s misinformation program undermines more than it teaches’, Monash University, 28 February 2022, <https://lens.monash.edu/@politics-society/2022/02/28/1384482/indonesias-misinformation-program-undermines-more-than-it-teaches>, accessed 20 February 2025
Illidge, M., ‘Plan to regulate AI and misinformation in South Africa’, My Broadband, 11 May 2024, <https://mybroadband.co.za/news/internet/535839-plan-to-regulate-ai-and-misinformation-in-south-africa.html>, accessed 11 February 2025
Imperva, ‘2024 Bad Bot Report’, 2024, <https://www.imperva.com/resources/resource-library/reports/2024-bad-bot-report-report-ty?lang=EN&asset_id=6912&gated=1>, accessed 15 May 2025
Imray, G. and Magome, M., ‘South Africa’s President Ramaphosa is reelected for second term after a dramatic late coalition deal’, AP News, 15 June 2024, <https://apnews.com/article/south-africa-president-election-parliament-b71d8f3de7b3bb4be76f94e8d75f0daf>, accessed 11 February 2025
Indonesia Business Post, ‘Ministry Communication and Informatics prepares AI ethics guidelines guaranteeing security and rights’, 17 November 2023, <https://indonesiabusinesspost.com/insider/ministry-communication-and-informatics-prepares-ai-ethics-guidelines-guaranteeing-security-and-rights>, accessed 19 February 2025
Inforial, ‘2024 election: Bawaslu joins forces with communications ministry to tackle hoaxes’, The Jakarta Post, 16 October 2023, <https://www.thejakartapost.com/adv/2023/10/16/2024-election-bawaslu-joins-forces-with-communications-ministry-to-tackle-hoaxes.html>, accessed 12 February 2025
Information and Communication Technology Division, Government of the People’s Republic of Bangladesh, ‘National Strategy for Artificial Intelligence’, March 2020, <https://ird.gov.bd/sites/default/files/files/ird.portal.gov.bd/miscellaneous_info/46fc550e_bdbb_4ec4_b2ca_1da23e8ccab0/National%20Strategy%20for%20Artificial%20Intellgence%20-%20Bangladesh%20.pdf>, accessed 17 April 2025
InSight Crime, ‘#NarcoPresidenteAMLO hits 170 million mentions’, 12 February 2024, <https://insightcrime.org/the-organization/narcopresidenteamlo-hits-170-million-mentions>, accessed 17 May 2025
Instituto Nacional de Estadística y Geografía (INEGI), Encuesta Nacional sobre Disponibilidad y Uso de Tecnologías de la Información en los Hogares 2019 [National Survey on the Availability and Use of Information Technologies in Households 2019], <https://www.inegi.org.mx/rnm/index.php/catalog/556>, accessed 3 July 2025
Instituto Nacional Electoral (INE), Certeza INE 2024, [n.d.a], <https://centralelectoral.ine.mx/certeza>, accessed 16 May 2025
—, Ubica tu Casilla – Proceso Electoral 2023-2024 [Locate Your Polling Station: 2023–2024 Electoral Process], [n.d.b], <https://ubicatucasilla.ine.mx>, accessed 16 May 2025
—, Actores Políticos [Political Actors], [n.d.c], <https://www.ine.mx/actores-politicos>, accessed 17 May 2025
—, ‘Estrategias’ [Strategies], 2024a, <https://centralelectoral.ine.mx/wp-content/uploads/2024/03/Estrategias_Certeza-2024.pdf>, accessed 16 May 2025
—, ‘Te presentamos a “Inés”, la asistente virtual del INE para combater la desinformación sobre las elecciones 2024’ [Meet ‘Inés’, the INE’s virtual assistant to combat misinformation about the 2024 elections], 26 April 2024b, <https://centralelectoral.ine.mx/2024/04/26/te-presentamos-a-ines-la-asistente-virtual-del-ine-para-combatir-la-desinfomacion-sobre-las-elecciones-2024>, accessed 17 May 2025
—, ‘SEs y CAEs recorren todo el país para capacitar a la ciudadanía que interarán Casillas el 2 de junio’ [Electoral supervisors and electoral assistant trainers are travelling across the country to train the citizens who will staff polling stations on 2 June], 12 May 2024c, <https://centralelectoral.ine.mx/2024/05/12/ses-y-caes-recorren-todo-el-pais-para-capacitar-a-la-ciudadania-que-integraran-casillas-el-2-de-junio>, accessed 16 May 2025
International IDEA, Voter Turnout Database, ‘Ghana’, [n.d.a], <https://www.idea.int/data-tools/data/question-country?question_id=9189&country=84&database_theme=293>, accessed 10 April 2025
—, Voter Turnout Database, ‘Indonesia’, [n.d.b], <https://www.idea.int/data-tools/data/country?country=104&database_theme=293>, accessed 20 June 2025
—, Voter Turnout Database, ‘Mexico’, [n.d.c] <https://www.idea.int/data-tools/data/question-country?question_id=9189&country=144&database_theme=293>, accessed 8 April 2025
—, Democracy Tracker, ‘Ghana’, [n.d.d], <https://www.idea.int/democracytracker/country/ghana>, accessed 3 March 2025
—, Online Political Advertising and Microtargeting: The Latest Legal, Ethical, Political and Technological Evolutions, Webinar Series Report, 15 and 18 June 2020 (Stockholm: International IDEA, 2020), <https://www.idea.int/sites/default/files/publications/online-political-advertising-and-microtargeting-the-latest-legal-ethical-political-and-technological-evolutions-en.pdf>, accessed 17 February 2025
—, Misión de Acompañamiento Técnico Electoral al Instituto Nacional Electoral de México, Proceso Electoral Federal y los Procesos Electorales Locales Concurrentes 2023–2024, ‘Informe Final’ [Electoral Technical Assistance Mission to the National Electoral Institute of Mexico, 2023–2024 Federal Electoral Process and Concurrent Local Elections, ‘Final Report’], 1 July 2024a, <https://repositoriodocumental.ine.mx/xmlui/handle/123456789/175918>, accessed 17 April 2025
—, Codes of Conduct on the Rise: Fair and Ethical Political Campaigning Online, Policy Brief (Stockholm: International IDEA, 2024b), <https://doi.org/10.31752/idea.2024.23>
—, Democracy Tracker, Mongolia – November 2024c, <https://www.idea.int/democracytracker/report/mongolia/november-2024>, accessed 2 March 2025
International Telecommunication Union (ITU), National Movement on Digital Literacy #Siberkreasi, [n.d.], <https://www.itu.int/net4/wsis/archive/stocktaking/Project/Details?projectId=1514925904>, accessed 16 February 2025
International Trade Administration (ITA), Ghana Digital Economy, 20 September 2024, <https://www.trade.gov/country-commercial-guides/ghana-digital-economy>, accessed 6 March 2025
Islam, Z. and Jamil Khan, M., ‘Spin doctors go into overdrive ahead of polls’, The Daily Star, 12 November 2023, <https://www.thedailystar.net/news/bangladesh/politics/news/spin-doctors-go-overdrive-ahead-polls-3467331>, accessed 5 March 2025
Ismaila, I., ‘#GhanaDecides2024: CJID deploys team to counter election misinformation’, HumAngle, 7 December 2024, <https://humanglemedia.com/ghanadecides2024-cjid-deploys-team-to-counter-election-misinformation>, accessed 8 March 2025
Jahangir, R., ‘How political parties manipulate cyberspace for electioneering’, Herald, 24 July 2018, <https://herald.dawn.com/news/1398599/how-political-parties-manipulate-cyberspace-for-electioneering>, accessed 20 February 2025
—, ‘Life in PTI’s social media bubble’, Dawn, 16 August 2020, <https://www.dawn.com/news/1574529>, accessed 20 February 2025
—, ‘Online rallies, AI script: New threats loom ahead of 2024 polls’, Dawn, 16 January 2024, <https://www.dawn.com/news/1806163/online-rallies-ai-script-new-threats-loom-ahead-of-2024-polls>, accessed 19 February 2025
Jahangir, R. and Popalzai, S., ‘Dawn investigations: ‘#Grabyourkeyboards’: Inside Pakistan’s hashtag mills’, Dawn, 14 January 2020, <https://www.dawn.com/news/1518967/grabyourkeyboards-inside-pakistans-hashtag-mills>, accessed 21 February 2025
Jaime, A., ‘Misinformation about Mexico’s historic election is spreading. Here’s what to know’, Teen Vogue, 6 June 2024, <https://www.teenvogue.com/story/misinformation-about-mexicos-historic-election-is-spreading-heres-what-to-know>, accessed 8 April 2025
Jeffrie, N., Bahia, K., Lindsey, D., Ouattara Boni, A-N., Sibthorpe, C. and Zagdanski, J., ‘The Mobile Gender Gap Report’, GSMA, May 2024, <https://www.gsma.com/r/wp-content/uploads/2024/05/The-Mobile-Gender-Gap-Report-2024.pdf>, accessed 7 April 2025
Jonathan, S., Sanusi, S. A., Akinbola, T. E. and Nwankwo, E. G., ‘Information Ecosystems ahead of Ghana’s 2024 Elections: Assessing Misinformation/Disinformation Threats and Patterns, Voter Competence and Media Readiness’, Centre for Journalism Innovation and Development, 2024, <https://thecjid.org/document/information-ecosystems-ahead-of-ghanas-2024-elections-assessing-misinformation-disinformation-threats-and-patterns-voter-competence-and-media-readiness>, accessed 6 March 2025
Joy News, ‘Our greatest fear for Election 2024 is misinformation – EC’, 13 November 2024, <https://www.myjoyonline.com/our-greatest-fear-for-election-2024-is-misinformation-ec>, accessed 8 March 2025
Juniarto, D., ‘AI in the 2024 Indonesian election: Is having ethical guidelines enough?’, The Jakarta Post, 12 February 2024, <https://www.thejakartapost.com/opinion/2024/02/10/ai-in-the-2024-indonesian-election-is-having-ethical-guidelines-enough.html>, accessed 20 February 2025
Kamran, S., ‘The thermometer broke! Digital purdah, class, and gender transgressions on Pakistani TikTok’, Social Media + Society, 9/2 (2023), <https://doi.org/10.1177/20563051231182359>
Karim, T. A., ‘The outcome of the election in Bangladesh and what lies ahead for its foreign policy’, The National Bureau of Asian Research, 11 January 2024, <https://www.nbr.org/publication/the-outcome-of-the-election-in-bangladesh-and-what-lies-ahead-for-its-foreign-policy>, accessed 4 March 2025
Kemp, S., ‘Digital 2023: Indonesia’, DataReportal, 9 February 2023a, <https://datareportal.com/reports/digital-2023-indonesia>, accessed 13 February 2025
—, ‘Digital 2023: Pakistan’, DataReportal, 13 February 2023b, <https://datareportal.com/reports/digital-2023-pakistan>, accessed 13 March 2024
—, ‘Internet Use in 2024’, DataReportal, 31 January 2024a, <https://datareportal.com/reports/digital-2024-deep-dive-the-state-of-internet-adoption>, accessed 11 February 2025
—, ‘Digital 2024: Bangladesh’, DataReportal, 23 February 2024b, <https://datareportal.com/reports/digital-2024-bangladesh>, accessed 4 March 2025
—, ‘Digital 2024: Ghana’, DataReportal, 23 February 2024c, <https://datareportal.com/reports/digital-2024-ghana>, accessed 6 March 2025
—, ‘Digital 2024: Mexico’, DataReportal, 23 February 2024d, <https://datareportal.com/reports/digital-2024-mexico>, accessed 15 May 2025
—, ‘Digital 2024: Mongolia’, DataReportal, 23 February 2024e, <https://datareportal.com/reports/digital-2024-mongolia>, accessed 16 February 2025
—, ‘Digital 2024: South Africa’, DataReportal, 23 February 2024f, <https://datareportal.com/reports/digital-2024-south-africa>, accessed 11 February 2025
Khalid, A., ‘Election illusions: Deepfakes and disinformation in Pakistan’, Digital Rights Foundation, 13 November 2023, <https://digitalrightsfoundation.pk/election-illusions-deepfakes-disinformation-in-pakistan>, accessed 21 February 2025
Khalid, Z. ‘Pakistan’s draft national AI strategy is a hodgepodge of technospeak’, Centre for Strategic and Contemporary Research, 21 June 2023, <https://cscr.pk/explore/themes/politics-governance/pakistans-draft-national-ai-policy-is-a-hodgepodge-of-technospeak>, accessed 11 June 2025
Khan, H., ‘The influence of fake news in Pakistan’s 2024 general election’, Stratheia, 24 January 2024, <https://stratheia.com/the-influence-of-fake-news-in-pakistans-2024-general-election>, accessed 20 February 2025
Khan, N., ‘Ex-PM Khan’s party says “millions” attended virtual rally despite social media disruption’, Arab News Pakistan, last updated 18 December 2023, <https://www.arabnews.pk/node/2427636/pakistan>, accessed 10 June 2025
Klepper, D., ‘Sexist tropes and misinformation swirl online as Mexico prepares to elect its first female leader’, AP News, 1 June 2024, <https://apnews.com/article/mexico-election-sheinbaum-facebook-lopez-obrador-79adddaf8300f30af51fddbbf3165216>, accessed 8 April 2025
Kportufe, F., ‘AI-fueled disinformation in Ghana’s 2024 elections – Ghana Fact-Checking Coalition’, Citi Newsroom, 29 January 2025, <https://citinewsroom.com/2025/01/ai-fueled-disinformation-in-ghanas-2024-elections-ghana-fact-checking-coalition>, accessed 6 March 2025
Kreuser, C. and Gitonga Theuri, J., ‘Contextualising Political Advertising Policy to Political Micro-targeting in the South Africa 2024 Elections’, African Internet Rights Alliance, November 2024, <https://aira.africa/wp-content/uploads/2024/11/Contextualising-Political-Advertising-Policy-to-Political-Micro-Targeting-in-South-Africa-2024-Elections.pdf>, accessed 11 February 2025
Kutanto, H., ‘Peran DJPPI Kominfo Dalam Mencegah Hoax Pemilu 2024’ [The Role of DJPPI Kominfo in Preventing Election Hoaxes 2024], Ministry of Communications and Informatics, 7 December 2023, https://djppi.komdigi.go.id/news/peran-djppi-kominfo-dalam-mencegah-hoax-pemilu-2024, accessed 19 February 2025
Kwafo, E. N. Y., ‘Penplusbytes hosts inaugural Youth against Disinformation Conference’, Modern Ghana, 24 October 2024, <https://www.modernghana.com/news/1351507/penplusbytes-hosts-inaugural-youth-against-disinfo.html>, accessed 8 March 2025
Lagos, A., ‘Claudia Sheinbaum, víctima de deepfake: su imagen fue utilizada en un video fraudulento generado por IA’ [Claudia Sheinbaum, victim of a deepfake: Her image was used in a fraudulent AI-generated video], Wired, 25 January 2024a, <https://es.wired.com/articulos/claudia-sheinbaum-victima-de-deepfake-su-imagen-fue-utilizada-en-un-video-fraudulento-generado-por-ia>, accessed 17 May 2025
—, ‘Transparencia México: Recursos para combatir la desinformación durante el proceso electoral [Transparency Mexico: Resources to combat disinformation during the electoral process], Wired, 12 March 2024b, <https://es.wired.com/articulos/transparencia-mexico-recursos-para-combatir-la-desinformacion-durante-el-proceso-electoral>, accessed 17 April 2025
Lamb, K., Potkin, F. and Teresia, A., ‘Generative AI may change elections this year. Indonesia shows how’, Reuters, 8 February 2024, <https://www.reuters.com/technology/generative-ai-faces-major-test-indonesia-holds-largest-election-since-boom-2024-02-08>, accessed 13 February 2025
LaMonica, M., ‘Ghana elections: Ethnicity is not allowed in the country’s politics, yet it still helps shape outcomes’, The Conversation, 1 August 2024, <https://theconversation.com/ghana-elections-ethnicity-is-not-allowed-in-the-countrys-politics-yet-it-still-helps-shape-outcomes-234818>, accessed 8 March 2025
Le Monde, ‘Indonesia: Prabowo Subianto wins presidency, official results confirm’, 20 March 2024, <https://www.lemonde.fr/en/international/article/2024/03/20/indonesia-prabowo-subianto-wins-presidency-official-results-confirm_6639261_4.html>, accessed 21 June 2025
Li, E. R., Shultz, B. and Jankowicz, N., ‘Deepfake Pornography Goes to Washington: Measuring the Prevalence of AI-Generated Non-consensual Intimate Imagery Targeting Congress’, American Sunlight Project, 11 December 2024, <https://static1.squarespace.com/static/6612cbdfd9a9ce56ef931004/t/67586997eaec5c6ae3bb5e24/1733847451191/ASP+DFP+Report.pdf>, accessed 17 April 2025
Lkhaajav, B., ‘Mongolia’s political parties showcase candidates for the 2024 parliamentary election’, The Diplomat, 22 May 2024, <https://thediplomat.com/2024/05/mongolias-political-parties-showcase-candidates-for-the-2024-parliamentary-election>, accessed 17 February 2025
Macdonald, G., ‘A Perilous Moment for Bangladesh’s Democracy’, United States Institute of Peace, 4 March 2024, <https://www.usip.org/publications/2024/03/perilous-moment-bangladeshs-democracy>, accessed 4 March 2025
Malik, M., ‘PTI, police face off in several cities amid protests against “rigging’’’, Dawn, 18 February 2024, <https://www.dawn.com/news/1815124/pti-police-face-off-in-several-cities-amid-protests-against-rigging>, accessed 19 February 2025
Mantas, H., ‘The Mongolian Fact Checking Center fights misinformation in a country where the word doesn’t exist’, Poynter, 10 August 2021, <https://www.poynter.org/fact-checking/2021/the-mongolian-fact-check-center-fights-misinformation-in-a-country-where-the-word-doesnt-exist>, accessed 16 February 2025
Mao, F., ‘Pakistan shut down the internet – but that didn’t stop the protests’, BBC, 12 May 2023, <https://www.bbc.com/news/world-asia-65541769>, accessed 22 February 2025
—, ‘Prabowo Subianto: The tainted ex-military chief who will be Indonesia’s new leader’, BBC, 15 February 2024, <https://www.bbc.com/news/world-asia-68237141>, accessed 15 February 2025
Marchant de Abreu, C., ‘Artificial intelligence and deepfakes take over Pakistan’s elections’, France 24, 9 February 2024, <https://www.france24.com/en/tv-shows/truth-or-fake/20240208-artificial-intelligence-and-deepfakes-takeover-pakistan-elections>, accessed 22 March 2025
Martínez Chacón, M., ‘Expertos ven una “avalancha” de noticias falsas en futura elección presidencial de México’ [Experts see an ‘avalanche’ of fake news in Mexico’s upcoming presidential election], Los Angeles Times, 22 August 2023a, <https://www.latimes.com/espanol/mexico/articulo/2023-08-22/expertos-ven-una-avalancha-de-noticias-falsas-en-futura-eleccion-presidencial-de-mexico>, accessed 15 May 2025
—, ‘Experts warn against wave of fake news ahead of Mexico’s 2024 presidential election’, Independent, 25 August 2023b, <https://www.independent.co.uk/news/world/americas/us-politics/andres-manuel-lopez-obrador-ap-mexico-experts-monterrey-b2399596.html>, accessed 15 May 2025
Masduki, ‘The politics of disinformation in Indonesia (2014–2019)’, in G. López-García, D. Palau-Sampio, B. Palomo, E. Campos-Domínguez and P. Masip (eds), Politics of Disinformation: The Influence of Fake News on the Public Sphere (John Wiley & Sons, 2021), <https://doi.org/10.1002/9781119743347.ch5>
Maulia, E., ‘Jokowi boosts policy support for Prabowo as Indonesia vote nears’, Nikkei Asia, 15 January 2024, <https://asia.nikkei.com/Politics/Indonesia-election/Jokowi-boosts-policy-support-for-Prabowo-as-Indonesia-vote-nears>, accessed 14 February 2025
McInnes, K., ‘South African digital & social media statistics 2024’, Meltwater,19 March 2024, <https://www.meltwater.com/en/blog/social-media-statistics-south-africa>, accessed 11 February 2025
Media Foundation for West Africa (MFWA), ‘Disinformation and elections: Here’s what we learnt during Ghana’s 2024 elections’, 19 February 2025, <https://mfwa.org/issues-in-focus/disinformation-and-elections-heres-what-we-learnt-during-ghanas-2024-elections>, accessed 3 March 2025
Media Monitoring Africa, ‘Digital literacy project Web Rangers launches WhatsApp channel for parents’, Press Council of South Africa, 5 September 2024, <https://presscouncil.org.za/2024/09/05/web-rangers-media-literacy-for-children>, accessed 9 February 2025
Mediana, ‘Campaigns on social media: Buzzers and influencers are increasingly “fluid’’’, Kompas, 27 December 2023, <https://www.kompas.id/baca/english/2023/12/26/en-kampanye-di-media-sosial-sosok-pendengung-dan-pemengaruh-semakin-cair>, accessed 14 February 2025
Meedan, ‘Meedan and Nest Center launch Mongolia’s first fact-checking tipline’, 17 May 2024, <https://meedan.com/post/meedan-and-nest-center-launch-mongolias-first-fact-checking-tipline>, accessed 19 February 2025
Méndez, C., ‘Difunden “fotografía íntima” de Mariana Rodríguez creada con IA; alista denuncia’ [‘Intimate photograph’ of Mariana Rodríguez created with AI released; she is preparing a complaint], El Heraldo de México, 22 January 2024, <https://heraldodemexico.com.mx/elecciones/2024/1/22/difunden-fotografia-intima-de-mariana-rodriguez-creada-con-ia-alista-denuncia-571656.html>, accessed 17 April 2025
Mgwili-Sibanda, F., ‘TikTok reinforces commitment to platform integrity with key initiatives ahead of Ghana elections’, TikTok Newsroom, 26 November 2024, <https://newsroom.tiktok.com/en-africa/ghana-elections>, accessed 8 March 2025
Millan, C., Kao, J. and Saito, Y., ‘Mexico election results for presidential, congressional races’, Bloomberg News, 2 June 2024, <https://www.bloomberg.com/graphics/2024-mexico-election-results>, accessed 8 April 2025
Ministry of Communications and Informatics (Kominfo), Literasi Digital 2022 [Digital Literacy Index 2022], [n.d.], <https://data.komdigi.go.id/opendata/dataset/indeks-literasi-digital-indonesia>, accessed 15 February 2025
Mir, A. and Salikuddin, T., ‘Understanding Pakistan’s Election Results’, United States Institute of Peace, 13 February 2024, <https://www.usip.org/publications/2024/02/understanding-pakistans-election-results>, accessed 21 February 2025
Mir, A. and Siddiqui, N., ‘Losing Facts to Fiction: Nationalism, Misinformation and Conspiracy Theories in Pakistan’, United States Institute of Peace, 17 November 2022, <https://www.usip.org/publications/2022/11/losing-facts-fiction-nationalism-misinformation-and-conspiracy-theories>, accessed 21 February 2025
Modern Ghana, ‘Ghana Fact-Checking Coalition records over 100 false publications’, 12 December 2024, <https://www.modernghana.com/news/1365129/election2024-ghana-fact-checking-coalition-recor.html>, accessed 6 March 2025
Montes, R., ‘Manipulación y mentiras son riesgos de la inteligencia artificial en elecciones, advierte comisionado’ [Manipulation and misinformation are election risks posed by artificial intelligence, commissioner warns], Milenio, 3 March 2024, <https://www.milenio.com/politica/elecciones/comisionado-advierte-riesgos-inteligencia-artificial-elecciones>, accessed 15 May 2025
Muhtadi, B. and Muslim, K., ‘Assessing Prabowo-Gibran’s victory: An exit-poll aftermath analysis of the 2024 presidential election’, Fulcrum, 20 May 2024, <https://fulcrum.sg/assessing-prabowo-gibrans-victory-an-exit-poll-aftermath-analysis-of-the-2024-presidential-election>, accessed 7 April 2025
Myagmar, M. and Nielsen, P. E., ‘The Mongolian media landscape in transition’, Nordicom Review, 22/2 (2001), <https://doi.org/10.1515/nor-2017-0350>
Nación321, ‘Martí Batres acusa sabotaje a su imagen con inteligencia artificial’ [Martí Batres makes accusation of AI-driven sabotage of his image], 31 October 2023a, <https://www.nacion321.com/elecciones/marti-batres-acusa-sabotaje-a-su-imagen-con-inteligencia-artificial>, accessed 21 June 2025
—, ‘Matrí Batres y Eduardo Clark explican cómo se hace un audio con IA’ [Martí Batres and Eduardo Clark explain how an AI-generated audio clip is made], 1 November 2023b, <https://www.nacion321.com/centro/marti-batres-y-eduardo-clark-explican-como-se-hace-un-audio-con-ia>, accessed 15 May 2025
Naciones Unidas México, ‘La desinformación se ha convertido en una industria y puede causar más muertes que la COVID-19: experto’ [Disinformation has become an industry and could cause more deaths than COVID-19: expert], 16 July 2021, <https://coronavirus.onu.org.mx/la-desinformacion-se-ha-convertido-en-una-industria-y-puede-causar-mas-muertes-que-la-covid-19-experto>, accessed 17 May 2025
Naeem, W. and Rehmat, A., ‘Countering Disinformation in Pakistan: Lessons and Recommendations for Digital Journalism’, Freedom Network and DigiMAP, December 2022, <https://www.fnpk.org/wp-content/uploads/2022/12/Countering-Disinformation-Report-1.pdf>, accessed 15 July 2025
Naranjargal, K., ‘Internet freedom in Mongolia’, FOCUS, 90 (2018), <https://www.hurights.or.jp/archives/focus/section3/2018/03/internet-freedom-in-mongolia.html>, accessed 16 February 2025
National Statistics Office of Mongolia (NSOM), United Nations Children’s Fund (UNICEF) and United Nations Population Fund (UNFPA), ‘Mongolia Social Indicator Sample Survey-2023: Key Indicators Report’, July 2024, <https://mics.unicef.org/sites/mics/files/2024-08/Mongolia%202023%20SISS-MICS%20Key%20Indicators%20Report_English.pdf>, accessed 16 February 2025
Ndhlovu, L., ‘Combating disinformation in South Africa’s elections’, Development and Cooperation, 14 October 2024, <https://www.dandc.eu/en/article/during-recent-elections-south-africa-disinformation-spread-rapidly-discrediting-political>, accessed 11 February 2025
Ndlovu, N., ‘AI deepfakes and SA’s fight to protect the 2024 election’, Tech Central, 30 January 2024, <https://techcentral.co.za/ai-deepfakes-fight-protect-2024-election/238743>, accessed 11 February 2025
Ndzendze, B. and Hlabisa, N., ‘South Africa’s new coalition government: Implications for social, economic and foreign policy’, Africa Policy and Research Institute, 16 July 2024, <https://afripoli.org/south-africas-new-coalition-government-implications-for-social-economic-and-foreign-policy>, accessed 11 February 2025
Newman, N., Fletcher, R., Robertson, C. T., Arguedas, A. R., and Nielsen, R. K., ‘Reuters Institute Digital News Report 2024’, Reuters Institute for the Study of Journalism, 2024, <https://reutersinstitute.politics.ox.ac.uk/sites/default/files/2024-06/RISJ_DNR_2024_Digital_v10%20lr.pdf>, accessed 11 February 2025
Ngubane, S., ‘Countering disinformation in South Africa’s elections: The role of social media platforms’, Republic of South Africa, Government Communication and Information System, [n.d.], <https://www.gcis.gov.za/gcis.gov.za/countering-disinformation-in-south-africa-s-elections-the-role-of-social-media-platforms>, accessed 11 February 2025
Nurhayati-Wolff, H., ‘Breakdown of social media users by age and gender in Indonesia as of January 2021’, Statista, 30 January 2023, <https://www.statista.com/statistics/997297/indonesia-breakdown-social-media-users-age-gender>, accessed 30 May 2025
Open Government Partnership, Enabling Freedom of the Press (MN0055), 2021, <https://www.opengovpartnership.org/members/mongolia/commitments/mn0055>, accessed 19 February 2025
Organization for Security and Co-operation in Europe, Office for Democratic Institutions and Human Rights (OSCE/ODIHR), ‘Mongolia, Parliamentary Elections, 28 June 2024, ODIHR Election Observation Mission, Final Report’, 13 December 2024, <https://www.osce.org/files/f/documents/e/c/583375_2.pdf>, accessed 15 February 2025
Organization of American States, ‘Preliminary Report of the OAS Electoral Observation Mission in Mexico’, 4 June 2024, <https://www.oas.org/en/media_center/press_release.asp?sCodigo=S-009/24>, accessed 8 April 2025
Orozco, L., ‘Realizan primer Hackaton electoral, organizado por Como Vamos Nuevo León y Verificado’ [First electoral hackathon held, organized by Cómo Vamos Nuevo León and Verificado], Verificado, 10 May 2024a, <https://verificado.com.mx/primer-hackaton-como-vamos-nuevoleon-verificado>, accessed 16 May 2025
—, ‘Cómo ejercer tu derecho al voto: paso a paso’ [How to exercise your right to vote: Step by step], Verificado, 14 May 2024b, <https://verificado.com.mx/votar-paso-a-paso-explicador>, accessed 16 May 2025
—, ‘Elecciones 2024: Violencia política de género en aumento’ [Elections 2024: Gender-based political violence on the rise], Verificado, 22 May 2024c, <https://verificado.com.mx/elecciones-2024-violencia-politica-aumento>, accessed 8 April 2025
Osae-Kwapong, J., ‘Ghana’s election 2024: Technology and the changing nature of political competition’, Centre for International Governance Innovation, 5 December 2024a, <https://www.cigionline.org/articles/ghanas-election-2024-technology-and-the-changing-nature-of-political-competition>, accessed 6 March 2025
—, 'The 2024 presidential & parliamentary elections: Electoral competition and the future of Ghana’s democracy', Ghana Center for Democratic Development, 19 December 2024b, <https://cddgh.org/2024/12/19/the-2024-presidential-parliamentary-elections-electoral-competition-and-the-future-of-ghanas-democracy>, accessed 3 March 2025
Pakistan, Islamic Republic of, Ministry of Information and Broadcasting, Fact Checker MoIB, [n.d.], <https://moib.gov.pk/Pages/147/FactChecker>, accessed 22 June 2025
—, Ministry of Information Technology and Telecommunication, ‘Draft National Artificial Intelligence Policy’, 2022, <https://moitt.gov.pk/SiteImage/Misc/files/National%20AI%20Policy%20Consultation%20Draft%20V1.pdf>, accessed 17 April 2025
Parkin, B., ‘Deepfakes for $24 a month: How AI is disrupting Bangladesh’s election’, Financial Times, 13 December 2023, <https://www.ft.com/content/bd1bc5b4-f540-48f8-9cda-75c19e5ac69c>, accessed 5 March 2025
Parliament of Ghana, Cybersecurity Act, 2020 (Act 1038), <https://csdsafrica.org/wp-content/uploads/2021/08/Cybersecurity-Act-2020-Act-1038.pdf>, accessed 6 March 2025
Pellegrini, S. and Breda, T., ‘Five Key Takeaways from the 2024 Elections in Mexico’, Armed Conflict Location and Event Data (ACLED), 2 July 2024, <https://acleddata.com/2024/07/02/five-key-takeaways-from-the-2024-elections-in-mexico>, accessed 8 April 2025
Penplusbytes, ‘Understanding the Economics of Disinformation on Ghana’s Democracy’, July 2024, <https://penplusbytes.org/wp-content/uploads/2024/07/Understanding-the-economics-of-disinformation-on-Ghanas-Democracy-1.pdf>, accessed 3 March 2025
Prempeh, H. K., ‘Ghana’s December 2024 elections put the resilience of the country’s vaunted democracy to the test’, Africa Policy Research Institute, 29 November 2024, <https://afripoli.org/ghanas-december-2024-elections-put-the-resilience-of-the-countrys-vaunted-democracy-to-the-test>, accessed 8 March 2025
Prothom Alo, ‘Facebook shutdown on election day proposed at inter-ministerial meeting’, 21 November 2023, <https://en.prothomalo.com/bangladesh/zqdfv70bg5>, accessed 7 March 2025
Quelle, D., Cheng, C., Bovet, A. and Hale, S. A., ‘Lost in translation: Multilingual misinformation and its evolution’, arXiv preprint, 27 October 2023, <https://arxiv.org/abs/2310.18089>, accessed 3 March 2025
Rahman, S. A., ‘Why opposition parties in Bangladesh are boycotting next month’s elections’, Voice of America, 8 December 2023, <https://www.voanews.com/a/opposition-party-in-bangladesh-to-boycott-next-month-s-elections-/7390460.html>, accessed 4 March 2025
Rahman, S., Dip, S., Karim, F. and Naser, S., ‘Whispers amidst the Echo Chamber: Decoding Bangladesh’s 2024 Election Disinformation’, Center for Critical and Qualitative Studies and Earki, January 2024, <https://www.fact-watch.org/whisperers-amidst-the-echo-chamber-decoding-bangladeshs-2024-election-disinformation>, accessed 10 June 2025
Ramis, J., ‘Las voces generadas por IA se vuelven indistinguibles de las humanas (también en español)’ [AI-generated voices are becoming indistinguishable from human ones (including in Spanish)], Wired, 8 November 2023, <https://es.wired.com/articulos/voces-generadas-por-ia-indistinguibles-las-humanas-en-espanol>, accessed 15 May 2025
Ratcliffe, R., ‘From military leader to “harmless grandpa”: The rebranding of Indonesia’s Prabowo’, The Guardian, 9 January 2024, <https://www.theguardian.com/world/2024/jan/09/indonesia-election-prabowo-subianto-rebranding-kidnapping-accusations>, accessed 7 April 2025
Rayda, N., ‘Indonesia elections 2024: How AI has become a double-edged sword for candidates and election officials’, Channel News Asia, 4 February 2024, <https://www.channelnewsasia.com/asia/ai-disinformation-deepfakes-indonesia-elections-4091296>, accessed 14 February 2025
Reporters Without Borders (RSF), Mongolia, [n.d.], <https://rsf.org/en/country/mongolia>, accessed 10 June 2025
—, ‘Threats to press freedom in Pakistan from hastily adopted laws ahead of elections’, 11 August 2023, <https://rsf.org/en/threats-press-freedom-pakistan-hastily-adopted-laws-ahead-elections>, accessed 22 February 2025
—, ‘Mongolia: RSF presents its recommendations to strengthen ambitious press freedom bill’, 3 December 2024, <https://rsf.org/en/mongolia-rsf-presents-its-recommendations-strengthen-ambitious-press-freedom-bill>, accessed 19 February 2025
Reuters, ‘Imran Khan: What are the criminal cases against Pakistan former PM?’, 31 January 2024, <https://www.reuters.com/world/asia-pacific/main-criminal-cases-against-pakistans-imran-khan-2024-01-31>, accessed 21 February 2025
Rezwan, ‘The government of Bangladesh quietly passed the new Cyber Security Act 2023’, Global Voices, 19 September 2023, <https://globalvoices.org/2023/09/19/the-government-of-bangladesh-quietly-passed-the-new-cyber-security-act-2023>, accessed 6 March 2025
Riaz, A., ‘How Bangladesh’s Digital Security Act is creating a culture of fear’, Carnegie Endowment for International Peace, 9 December 2021, <https://carnegieendowment.org/2021/12/09/how-bangladesh-s-digital-security-act-is-creating-culture-of-fear-pub-85951>, accessed 6 March 2025
—, ‘What Bangladesh’s widely boycotted election reveals about its future’, Atlantic Council, 18 January 2024, <https://www.atlanticcouncil.org/blogs/new-atlanticist/what-bangladeshs-widely-boycotted-election-reveals-about-its-future>, accessed 4 March 2025
Rikkilä Tamang, L., ‘In Mongolia, electoral reform ushers in a record number of women to parliament’, International IDEA, 11 July 2024, <https://www.idea.int/news/mongolia-electoral-reform-ushers-record-number-women-parliament>, accessed 15 February 2025
Roper, C., ‘South Africa’, in N. Newman, R. Fletcher, C. T. Robertson, A. R. Arguedas and R. K. Nielsen, ‘Reuters Institute Digital News Report 2024’, Reuters Institute for the Study of Journalism, 2024, <https://reutersinstitute.politics.ox.ac.uk/sites/default/files/2024-06/RISJ_DNR_2024_Digital_v10%20lr.pdf>, accessed 11 February 2025
RTL Today, ‘Mongolians to vote in poll dominated by corruption worries’, 25 June 2024, <https://today.rtl.lu/news/world/a/2208008.html>, accessed 17 February 2025
Sam, A., Yusuf, R. I. and Rahayu, P. B., ‘Gemoy and Samsul: How the propaganda won the 2024 Indonesian election’, Informasi, 54/2 (2024), pp. 133–42, <http://doi.org/10.21831/informasi.v54i2.75464>
Seddon, S. and Javed, F., ‘Pakistan election: PMLN and PPP reach agreement on coalition government’, BBC, 20 February 2024, <https://www.bbc.com/news/world-asia-68354256>, accessed 21 February 2025
Sguazzin, A., ‘Pro-Russian X accounts turn to touting Zuma’s MK Party’, Tech Central, 2 April 2024, <https://techcentral.co.za/pro-russian-x-accounts-zuma-mk-party/242310>, accessed 11 February 2025
Shaheen, S., ‘Fake news, escalation, and polarization: Pakistan’s disinformation vulnerabilities’, South Asian Voices, 12 May 2022, <https://southasianvoices.org/fake-news-escalation-and-polarization-pakistans-disinformation-vulnerabilities>, accessed 23 February 2025
Sharma, S., ‘Internet and mobile phone services shut in Pakistan in “ominous start” to election day’, Independent, 8 February 2024, <https://www.independent.co.uk/asia/south-asia/pakistan-national-elections-2024-internet-shutdown-b2492636.html>, accessed 22 February 2025
Shukla, V., ‘Deepfakes and elections: The risk to women’s political participation’, Tech Policy Press, 29 February 2024, <https://www.techpolicy.press/deepfakes-and-elections-the-risk-to-womens-political-participation>, accessed 17 April 2025
Siberkreasi, ‘Wakil Menteri Komunikasi dan Informatika: Kolaborasi jadi Kunci Penting Wujudkan Masyarakat Indonesia yang Berdaya dan Cakap Digital’ [Deputy Minister of Communication and Informatics: Collaboration is the Key to Realizing an Empowered and Digitally Capable Indonesian Society], 12 August 2023, <https://gnld.siberkreasi.id/wakil-menteri-komunikasi-dan-informatika-kolaborasi-jadi-kunci-penting-wujudkan-masyarakat-indonesia-yang-berdaya-dan-cakap-digital>, accessed 20 February 2025
Sinanoglu, S. and Breuer, A., ‘Countering Information Pollution to Protect Democracy’, German Institute of Development and Sustainability, IDOS Policy Brief 29/2024, <https://www.idos-research.de/fileadmin/user_upload/pdfs/publikationen/Policy_Brief/2024/PB_29.2024.pdf>, accessed 8 April 2025
Spinelli, A., ‘Overcoming Bangladesh’s electoral integrity deficit: Time for political compromise and dialogue’, International IDEA, 5 February 2024, <https://www.idea.int/news/overcoming-bangladeshs-electoral-integrity-deficit-time-political-compromise-and-dialogue>, accessed 4 March 2025
Statista, ‘Leading sources of information for political campaigns in Indonesia as of December 2023 – by generation’, [n.d.], <https://www.statista.com/statistics/1451104/indonesia-source-of-information-for-political-campaigns-by-generation>, accessed 15 February 2025
—, ‘Social media in Ghana – Statistics & facts’, 28 June 2024a, <https://www.statista.com/topics/9778/social-media-in-ghana>, accessed 6 March 2025
—, ‘Internet usage in Mexico – Statistics & facts’, 22 November 2024b, <https://www.statista.com/topics/3477/internet-usage-in-mexico>, accessed 10 June 2025
—, ‘Internet usage in Latin America – Statistics & facts’, 22 November 2024c, <https://www.statista.com/topics/2432/internet-usage-in-latin-america>, accessed 8 April 2025
Tarigan, E., ‘Indonesian presidential hopefuls are trying social media, K-pop to win young voters. Will it work?’, AP News, 8 February 2024, <https://apnews.com/article/indonesia-presidential-election-kpop-young-voters-73227fb99b0ed70f704c9f0f759622ce>, accessed 15 February 2025
Tenggara Strategics, ‘Hoaxes and AI: Communications ministry dials up combating efforts ahead of 2024 election’, The Jakarta Post, 9 October 2023, <https://www.thejakartapost.com/front-row/2023/10/09/hoaxes-and-ai-communications-ministry-dials-up-combating-efforts-ahead-of-2024-election.html>, accessed 19 February 2025
Thomson Reuters Foundation, ‘Ghana parties stoke anti-LGBTQ+ sentiment ahead of election’, Voice of America, 30 November 2024, <https://www.voanews.com/a/ghana-parties-stoke-anti-lgbtq-sentiment-ahead-of-election/7881015.html>, accessed 8 March 2025
TikTok, Conoce los resultados de las iniciativas de TikTok para proteger la integridad de su comunidad durante el proceso electoral de México 2024 [Learn about the results of TikTok's initiatives to protect the integrity of its community during the Mexico 2024 electoral process], 29 August 2024, <https://newsroom.tiktok.com/es-latam/tiktok-integridad-electoral-mexico-elecciones-2024>, accessed 16 May 2025
Timmerman, A., ‘Indonesia will enforce laws on content moderation with tight response time and harsh fines, documents show’, Rest of World, 16 August 2022, <https://restofworld.org/2022/indonesia-social-media-regulations>, accessed 20 February 2025
Tovuu, B., ‘Mongolia: Coronavirus and the media’, European Journalism Observatory, 19 May 2020, <https://en.ejo.ch/ethics-quality/mongolia-coronavirus-and-the-media>, accessed 18 February 2025
Towriss, D., ‘How South Africa’s elections may reshape its democracy’, International IDEA, 15 May 2024, <https://www.idea.int/blog/how-south-africas-elections-may-reshape-its-democracy>, accessed 11 February 2025
Transparency International Bangladesh, ‘Directive to national newspapers and TV channels to re-register respective online portals an example of blueprint for ultimate control of the media; TIB expresses grave concern about people’s right to free flow of information and freedom of opinion’, 1 September 2020, <https://www.ti-bangladesh.org/articles/press-release/6126>, accessed 10 April 2025
TRT World, ‘The battle of narratives: Pakistan’s troll armies’, YouTube, 10 December 2019, <https://www.youtube.com/watch?v=f0v2VWlsRCM>, accessed 22 February 2025
Tsedendamba, B., ‘Caught between two giants: How Mongolia’s democracy can counter authoritarian propaganda’, Power 3.0, 25 June 2024, <https://www.power3point0.org/2024/06/25/caught-between-two-giants-how-mongolias-democracy-can-counter-authoritarian-propaganda>, accessed 16 February 2025
Tsolmonbaatar, T., ‘Mongolia: Truly Media provides a new fact-checking program’, DW Akademie, 30 October 2019, <https://akademie.dw.com/en/fact-check-mongolia-a-new-verification-program-with-truly-media/a-51054358>, accessed 19 February 2025
United Nations Educational, Scientific and Cultural Organization (UNESCO) and IPSOS, ‘Survey on the Impact of Online Disinformation and Hate Speech’, September 2023, <https://www.unesco.org/sites/default/files/medias/fichiers/2023/11/unesco_ipsos_survey.pdf>, accessed 8 April 2025
United Nations Ghana, ‘Countering hate speech and misinformation- UNESCO kickstarts regional trainings for journalist ahead of 2024 elections’, 21 June 2024, <https://ghana.un.org/en/272297-countering-hate-speech-and-misinformation-unesco-kickstarts-regional-trainings-journalist>, accessed 8 March 2025
United States Department of State, ‘Mongolia 2021 Human Rights Report’, [n.d.], <https://www.state.gov/reports/2021-country-reports-on-human-rights-practices/mongolia>, accessed 2 March 2025
Utama, V. R., ‘Will the “gemoy” tactic be effective in wooing Gen Z voters?’, The Jakarta Post, 6 December 2023, <https://www.thejakartapost.com/opinion/2023/12/06/will-the-gemoy-tactic-be-effective-in-wooing-gen-z-voters.html>, accessed 15 February 2024
Valverde, D., ‘TikTok measures against disinformation in Mexico’s 2024 elections’, Mexico Business News, 15 April 2024, <https://mexicobusiness.news/tech/news/tiktok-measures-against-disinformation-mexicos-2024-elections>, accessed 15 April 2025
van Damme, P., ‘Disinformation, governance and the South African election’, ISS African Futures, 19 March 2024, <https://futures.issafrica.org/blog/2024/Disinformation-governance-and-the-South-African-election>, accessed 11 February 2025
Vergara, R., ‘INE y Meta lanzan programa “Soy Digital” para combatir desinformación en las elecciones’ [INE and Meta launch the ‘I Am Digital’ programme to combat disinformation during the elections], Proceso, 9 February 2024, <https://www.proceso.com.mx/nacional/2024/2/9/ine-meta-lanzan-programa-soy-digital-para-combatir-desinformacion-en-las-elecciones-323731.html>, accessed 17 May 2025
Verificado, ‘Cursos y talleres [Courses and workshops], [n.d.], <https://verificado.com.mx/cursos-y-talleres>, accessed 16 May 2025
Verity, J., ‘Data Codebook for a Round 9 Afrobarometer Survey in Ghana’, Afrobarometer, April 2024, <https://www.afrobarometer.org/wp-content/uploads/2024/05/GHA_R9.Codebook_21May24.final_.pdf>, accessed 10 June 2025
Volodya, B., ‘Law on cyber security’, Mondaq, 26 April 2022, <https://www.mondaq.com/security/1186842/law-on-cyber-security>, accessed 19 February 2025
Wasserman, H., Madrid-Morales, D. and Rabin, F., ‘An Assessment of Media Literacy and Fact-Checking Training Needs in South African Schools and Universities’, Africa Check, March 2022, <https://africacheck.org/sites/default/files/media/documents/2022-05/An%20assessment%20of%20media%20literacy%20and%20fact-checking%20training%20needs%20in%20South%20African%20schools%20and%20universities.pdf>, accessed 11 February 2025
Watmough, S. P., ‘From political pariah to president: Prabowo Subianto and the perils of populism in Indonesia’, European Center for Populism Studies, 24 October 2024, <https://www.populismstudies.org/from-political-pariah-to-president-prabowo-subianto-and-the-perils-of-populism-in-indonesia>, accessed 7 April 2025
Wiguna, D. K. S., ‘Kominfo: 4,8 juta konten negatif diblokir sejak 2018’ [Kominfo: 4.8 million negative pieces of content blocked since 2018], ANTARA, 23 February 2024, <https://kalteng.antaranews.com/berita/682557/kominfo-48-juta-konten-negatif-diblokir-sejak-2018>, accessed 19 February 2025
Wilson, T. and Starbird, K., ‘Cross-platform disinformation campaigns: Lessons learned and next steps’, Harvard Kennedy School Misinformation Review, 14 January 2020, <https://doi.org/10.37016/mr-2020-002>
World Health Organization, ‘Mongolia COVID-19 vaccines post-introduction evaluation (cPIE), 19-30 September 2022’, 29 September 2023, https://www.who.int/mongolia/publications-detail/WPR-DDC-2023-003, accessed 2 July 2025
Wulandari, L., ‘Shaping perceptions: Three layers of social media’s role in the 2024 Indonesia elections’, East West Center, Asia Pacific Bulletin No. 708, 4 November 2024, <https://www.jstor.org/stable/resrep64404>, accessed 7 April 2025
Wyrzykowski, R., ‘Rural–urban digital divide still poses a challenge in South Africa’, Opensignal, 13 December 2023, <https://www.opensignal.com/2023/12/13/rural-urban-digital-divide-still-poses-a-challenge-in-south-africa>, accessed 11 February 2025
Yerkes, S. and Hogan, E., ‘Ghana’s presidential contest shows why democracy requires more than well-run elections’, Carnegie Endowment for International Peace, 19 December 2024, <https://carnegieendowment.org/emissary/2024/12/ghana-election-democracy-challenges-lessons?lang=en>, accessed 3 March 2025
Yesmin, F., ‘Bangladesh needs a comprehensive framework to address the risks of AI’, The Business Standard, 3 November 2023, <https://www.tbsnews.net/thoughts/bangladesh-needs-comprehensive-framework-address-risks-ai-732414>, accessed 7 March 2025
Algorithm: Within social networks, algorithms are rules, signals and data that govern a platform’s operation. Algorithms determine how content is recommended, ranked, displayed and filtered to users.
Artificial intelligence: Artificial intelligence (AI) is a machine-based system that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations or decisions that can influence physical or virtual environments. Different AI systems vary in their levels of autonomy and adaptiveness after deployment.
Chilling effect: A chilling effect occurs when individuals or groups avoid expressing themselves due to fear of violating a law or regulation.
Cross-platform amplification: Cross-platform amplification refers to the strategic process of distributing and boosting information, narratives and content across multiple social networks to increase its impact, reach and visibility.
Deepfake: A deepfake is a piece of content (such as an image, video or audio recording) that has been digitally altered, often using generative AI, to either portray someone as another person or manipulate their appearance out of context to spread false or misleading information.
Disinformation: Disinformation is incorrect or misleading information with the deliberate intent to deceive individuals, influence public opinion and obscure the truth.
Echo chamber: An echo chamber is an environment or situation in which individuals hear and receive opinions of only one type or opinions and information that align with their personal values and beliefs.
Electoral management body: An electoral management body is an organization or body that has the sole purpose of and is legally responsible for managing some or all of the elements that are essential for the conduct of elections and instruments of direct democracy.
Generative AI: Generative AI, a subset of artificial intelligence, refers to systems that not only process information but generate new content based on the data they were trained on.
Hate speech: Hate speech can be understood as all types of expression that incite, promote, spread or justify violence, hatred or discrimination against a person or group of persons, or that denigrates them, by reason of their real or attributed personal characteristics or status.
Information integrity: A policy brief published by the UN (2023) defines ‘information integrity’ as the accuracy, consistency and reliability of information. Information integrity is threatened by disinformation, misinformation and hate speech.
Information pollution: Information pollution can be seen as the opposite of information integrity; it refers to the spread of low-value, irrelevant, biased, false and harmful information resulting in the dilution or outright suppression of essential facts.
Misinformation: Misinformation is incorrect or misleading information.
- ‘Rage baiting’ is defined as the use of coordinated adversarial engagements on major issues, such as race or migration.
Abbreviations
AI | Artificial intelligence |
---|---|
AL | Awami League, Bangladesh |
ANC | African National Congress |
ASEAN | Association of Southeast Asian Nations |
Bawaslu | General Election Supervisory Agency, Indonesia |
BNP | Bangladesh Nationalist Party |
BTRC | Bangladesh Telecommunication Regulatory Commission |
CRC | Communications Regulatory Commission, Mongolia |
CSA | Cyber Security Act, Bangladesh |
CSO | Civil society organization |
DP | Democratic Party, Mongolia |
DSA | Digital Security Act, Bangladesh |
EMB | Electoral management body |
IEC | Electoral Commission of South Africa |
INE | National Electoral Institute, Mexico |
Kominfo | Ministry of Communication and Digital Affairs, Indonesia |
LGBTQIA+ | Lesbian, gay, bisexual, transgender, queer, intersex and asexual |
MFCC | Mongolian Fact Checking Center |
MK | uMkhonto weSizwe party, South Africa |
MPP | Mongolian People’s Party |
NDC | National Democratic Congress, Ghana |
NPP | New Patriotic Party, Ghana |
PTA | Pakistan Telecommunication Authority |
PTI | Pakistan Tehreek-e-Insaf |
UNESCO | United Nations Educational, Scientific and Cultural Organization |
This report was written by lead author Thijs Heinmaa and co-author Alisa Schaible. The case studies were drafted by Cecilia Hammar (Bangladesh, Indonesia and Pakistan) and Mircea Lazar (Ghana, Mexico, Mongolia and South Africa). Valuable feedback and inputs on the draft were provided by Julia Keutgen, Sam van der Staak, Alberto Fernández Gibaja, Peter Wolf, Marilyn Neven and Sebastian Becker.
Thanks also go to Lisa Hagman and Tahseen Zayouna in International IDEA’s Publications team for supporting in the production of this report.
© 2025 International Institute for Democracy and Electoral Assistance
International IDEA publications are independent of specific national or political interests. Views expressed in this publication do not necessarily represent the views of International IDEA, or those of its Board or Council members. References to the names of countries and regions in this publication do not represent the official position of International IDEA with regard to the legal status or policy of the entities mentioned.
This publication was produced with the financial support of the European Union. Its contents are the sole responsibility of International IDEA and do not necessarily reflect the views of the European Union.
With the exception of any third-party images and photos, the electronic version of this publication is available under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 (CC BY-NC-SA 4.0) licence. You are free to copy, distribute and transmit the publication as well as to remix and adapt it, provided it is only for non-commercial purposes, that you appropriately attribute the publication, and that you distribute it under an identical licence. For more information visit the Creative Commons website: <http://creativecommons.org/licenses/by-nc-sa/4.0>.
Design and layout: International IDEA
DOI: <https://doi.org/10.31752/idea.2025.46>
ISBN: 978-91-7671-979-4 (HTML)
ISBN: 978-91-7671-978-7 (PDF)