Artificial Intelligence (AI) has become the new buzzword in today’s technological ecosystem. Although not entirely new, it has gained prominence in recent years due to significant advancements, and it is envisioned to revolutionise various aspects of society, including democracy and, particularly, elections.
The year 2024 has been termed a “super election year” with over 4 billion people eligible to vote in countries across the world, including approximately 21 countries on the African continent. Undoubtedly, AI stands to play a pivotal role in many of these countries’ elections (especially in countries with high AI adoption) and in the future.
This piece, a part of Dataphyte’s ongoing series on technology and elections, provides insights on the opportunities AI offers various political actors — political parties and candidates, electoral bodies, civil societies, and citizens — to transform democratic governance in Africa. Besides, it sets out the risks associated with the use of AI in emerging democracies, like those in Africa.
Source: Dataphyte
To begin, what is AI? In simple terms, the Electoral Institute for Sustainable Democracy in Africa (EISA) describes AI as advanced computing systems that combine various technologies to process and analyse large amounts of data. These systems mimic human abilities, such as reasoning, problem-solving, and decision-making by following instructions developed through algorithms. AI is considered the next set of technological advancement, and just as with other technologies shaping society, it offers significant potential for good, but also poses huge risks which can undermine democratic processes, and elections in particular.
Campaigns for Political Parties and Candidates
Perhaps one of the most noteworthy opportunities AI offers is the potential to democratise and level the campaign playing field for political parties and their candidates. In many African democracies, election campaign expenditure is increasing significantly.
For instance, an article by Dan Paget noted that parliamentary campaign spending in Ghana rose by 90 percent in just four years, now standing at £68,215. In Uganda, the average parliamentary campaign costs £94,000, in Kenya, £140,000, and in Nigeria, it is £352,000. These figures indicate money plays a crucial role in determining who gets elected, giving better-financed candidates a substantial advantage.
In light of this, AI tools could lower financial barriers to entry, especially for first-time and under-funded candidates who may lack substantial financial resources. With AI’s ability to handle a campaign’s most mundane and time-consuming tasks, such as generating press releases or identifying and targeting supporters, candidates would have less need to hire expensive consultants, particularly when finances are limited. The result could be a more open and accessible democracy, allowing small, modest campaigns to compete more effectively with well-funded juggernauts.
Additionally, candidates and political parties can take advantage of AI’s precise audience targeting capabilities, which are crucial in political campaigns. By using microdata from commercial data brokers, who possess detailed information about people’s reading, viewing, purchasing, and political behaviours, political parties and their candidates can fine-tune their targeting efforts. This allows them to deliver tailored messages to specific demographics, thereby influencing their final decisions, particularly those of undecided voters. This approach is likely to strengthen democracy, as the hallmark of democracy is the active participation of its citizens.
Furthermore, given that election campaigns in Africa are sometimes marred by hate speech, as witnessed in recent elections in Nigeria, Kenya and South Africa AI can be effectively and transparently used by election administration officials, civil societies, and observers to monitor the content, volume, and trends in electoral hate speech. This can help identify and address the unfair targeting of specific groups, including marginalised communities, thereby preventing such practices from tilting the playing field and ensuring that all candidates compete fairly and openly.
Notwithstanding these positives, AI-powered technologies can also be used for manipulative microtargeting, spread of misinformation and disinformation, which is becoming increasingly common during elections. Evidence from various studies shows that across various social media platforms, bots, often disguising as ordinary human accounts, have been responsible for spreading misinformation and contributing to the already highly politically tense environment. These bots sometimes employed by political parties and their candidates are highly effective at attacking voters from opposing camps and spreading information that can discourage them from voting. A notable example is Kenya’s 2017 election, where the consulting firm Cambridge Analytica allegedly used AI to target voters with disinformation, potentially influencing people’s emotions and eventually the outcome of the said election. This misuse of AI not only distorts public opinion but also undermines the credibility of democratic institutions, threatening the integrity of elections.
Additionally, AI can equip illiberal non-state actors and autocracies with easy access to a range and affordable unmanned tools that political opponents can utilise to further divide the electorate, intensifying polarisation and internal destabilisation which is harmful to democracy. This is particularly concerning in countries like Nigeria, where religion and ethnicity are already exploited as political tools by unscrupulous politicians. This risk of AI-fueled informational chaos is likely to increase as more countries adopt these technologies now and in the future.
Furthermore, while these new AI systems can indeed benefit political parties by democratising and levelling the campaign playing field, especially for the under-resourced, they could inadvertently create threats. AI-generated advertisements have the potential to flood voters with an information overload, making it difficult to distinguish fact from fiction.
Electoral Bodies
Beyond political parties and candidates, AI also presents electoral bodies, such as the Nigerian Independent National Electoral Commission (INEC) and Kenya’s Independent Electoral and Boundaries Commission (IEBC), with opportunities to revolutionise voter registration and identification systems. Evidence from countries like South Africa and Eswatini shows that AI algorithms have been instrumental in improving voter register management and reducing administrative errors. In countries like Nigeria and Ghana, where digital technologies like the Bimodal Voter Accreditation System (BVAS) are already performing similar functions, AI tools can further enhance these processes.
Additionally, AI enhances election monitoring capabilities by analysing vast datasets to detect irregularities and potential fraud instances. For instance, AI-powered software can detect inconsistencies in voter turnout patterns, triggering prompt responses from electoral authorities to investigate and correct anomalies, thereby safeguarding the integrity of the electoral process.
Also, the use of AI can streamline vote counting and results transmission, speeding up the dissemination of election outcomes and reducing delays as often the case in Nigerian elections. AI-powered tabulation systems can markedly decrease the time needed for tallying votes and reporting results, thereby enhancing public trust in electoral transparency and efficiency.
While these indeed are exciting developments, they are not without some risks. In fact, some authors have attributed the low deployment of AI for elections by a number of electoral bodies in Africa partly to the risks. For instance, AI-based authentication of voters’ facial or fingerprint features could produce authentication errors, and algorithms may lead to bias against specific voters. Similarly, desperate political actors can exploit these relatively low-cost generative AI to impersonate election officials and clone election results management processes. This is a concern shared by several authors as well. These acts no doubt can compromise the integrity of elections with far reaching consequences for democracy. AI can enhance the capabilities of cyber attackers, making it easier to disrupt election infrastructure, including voting machines and voter management devices, including ballot counting systems.
Civil Societies and Citizen
For civil societies organisations (CSOs) and citizens, AI can also play a crucial role in facilitating their involvement further in the electoral and democratic process. For instance, AI can empower CSOs with enhanced capabilities in data collection and analysis, strengthening their efforts in ensuring transparency in the electoral process. For example, AI-driven platforms can enable CSOs to monitor and document instances of electoral irregularities more effectively, amplifying their voice in advocating for electoral transparency and accountability. Also, AI tools can facilitate real-time communication and mobilisation of supporters, empowering civil societies to mobilise grassroots movements and foster civic engagement in democratic processes.
On the part of citizens, AI will democratise access to political information and fact-checking tools which can be leveraged to factcheck claims by politicians especially during campaigns, empowering them to make informed decisions during elections. In a similar vein, AI-powered applications can significantly contribute to providing voters with real-time updates on candidate policies and election results, thereby fostering transparency and accountability in electoral processes. While the use of these tools may not fully resolve the transparency challenges evident in the electoral processes of many African countries, such as Nigeria, studies indicate that their effective implementation has the potential to enhance transparency in the electoral process.
In addition to elections and voting, other aspects of democratic governance stand to benefit from the AI revolution. For instance, AI can empower citizens to voice their opinions, organise others with similar viewpoints, and act on their priorities beyond the ballot box, such as through protests—a crucial form of political participation that drives democracy. These tools can also enhance citizens’ ability to engage directly with political processes via interactive platforms that facilitate public participation in policy debates and decision-making.
Nonetheless, there are also potential risks associated with AI adoption by CSOs and citizens in their involvement or efforts to enhance democratic processes and civic engagement. For CSOs, the digital divide that comes with the adoption of digital tools like this has a tendency to lead to reduced collaboration and marginalisation, particularly among smaller organisations with limited access to AI technologies and digital literacy. For instance, disparities in AI adoption can limit the capacity of rural advocacy groups to effectively use AI-driven tools for monitoring elections and reporting electoral malpractices at the grassroots level. Additionally, concerns about the authenticity and verification of AI-generated information are significant, especially in regions where misinformation campaigns threaten to undermine the credibility of civil society initiatives aimed at promoting democratic governance.
Similarly, citizens are vulnerable to AI-powered misinformation and manipulation, as shown by AI-generated deepfake videos disseminating false information during Nigeria’s recent elections. These deceptive practices not only mislead voters but can also erode trust in democratic institutions, threatening the legitimacy of electoral outcomes.
As this piece has demonstrated, the adoption of AI into African democratic processes offers opportunities that can strengthen the democratic processes and also risks that can weaken it. Therefore, as different political actors in Africa navigate this technological frontier, it is important to develop regulatory frameworks and ethical guidelines to maximise the benefits while mitigating the potential harms associated with its adoption to the electoral process and more broadly democratic governance on the continent.