The latest edition of the International Political Expo (IPE 25) took place in Cape Town, South Africa, from 24-25 January 2025 under the theme, “Unlocking Voting Potential: The Art of Resourcing and Skilling for Winning Campaigns”.  A key pillar for discussions under this theme was the role of data and technology in campaigns. It was in this context that the policy paper, “The Pervasive Influence of A.I. on Global Political Campaigns in 2024” was officially launched..

The paper focuses on the pivotal role that Artificial Intelligence (AI) played in the 2024 elections around the world, by assessing AI-driven political campaigns and the rise of AI-induced disinformation. The paper which is a Future Shift Labs publication, utilises a series of case studies to discuss three key considerations of AI in electoral governance: 

(1) The integration of AI into campaign strategies and their impact on efficiency and effectiveness;
(2) Shaping of voter behaviour and;
(3) Ethical considerations and regulatory challenges.  

Overall, the paper notes that the 2024 elections revealed that AI has both transformative potential and high risks when it comes to impacting political campaigns and elections more broadly.

AI powers democracy… 

On enhancing the efficiency and effectiveness of political campaigns, the paper points to examples where AI has heightened voter targeting and personalised messaging and also improved the outreach capacity of campaigns through inclusive strategies such as enabling multilingual messaging. Illustratively, an AI-powered tool was used in the United States to translate campaign messages into multiple languages while in India and Indonesia, there were examples of AI being used to generate real-time, data-driven and personalised campaign messages. There were even examples of campaigns using AI-generated speeches, press releases and social media posts.

… but at what cost?

However, in contrast to these benefits, the paper identifies significant risks such as the rise of disinformation which has eroded public trust and the lack of accountability in the use of AI due to inadequate regulations. The embrace of AI in the formulation of communication materials and outreach strategies does raise the question of whether campaigns lose the key ingredient of authenticity in engaging with voters. The rise of deepfake technology as a campaign tool is discussed extensively in the paper, with an emphasis on how it has contributed to increased disinformation, fueled social polarisation and eroded public trust in conventional media which has been surpassed by the online platforms that are used to disseminate deepfake materials.

The paper noted that in Africa, deepfake videos were encountered during elections in South Africa and Zambia, while interestingly in Mauritius, the ruling government tried to overcome the scandal of leaked recordings that incriminated them for acts of illegal surveillance and government corruption by branding them as deepfakes. In another vulnerability for our democracies, the paper found that in regions with low-technology literacy, personalised videos and deepfakes generated from AI were found to be rather persuasive. This connects with the wider concerns presented by the social digital platforms that are used to disseminate that content and the risk of lopsided control of content algorithms that some political actors could possess, amplify and manipulate for their own ends. It is in this context that the question of regulation and ethics arises.

Innovation meets regulation

The paper reviews frameworks from the European Union (EU), the United States and Singapore, and based on the analysis, concludes that the ideal ecosystem for the use of AI in elections consists of data privacy safeguards, standard electoral laws and commitments to the ethical use of AI. Noting the dynamism of AI, constructive interpretation of these norms is identified as a key principle in order to address emerging threats and guard against over-regulation which could stifle innovation. 

Elements of such constructive interpretation include strategically reinterpreting existing laws to confront emerging challenges, imposing transparency obligations and general disclaimers for the use of AI and ensuring human oversight. While it has registered limited success, the cooperation of digital platforms is also considered an essential pillar of regulation efforts since it is these platforms that are used to amplify and widely disseminate AI generated content. There have also been examples of political actors and other stakeholders ascribing to voluntary codes of conduct as a good practice. However, the voluntary nature of these codes means that adherence is inconsistent.

Conclusion

In conclusion, the paper rightly calls for States to adopt a holistic approach to the reality of AI in elections. Legal regulations that are developed need to apply to all key stakeholders who include state actors, online platforms and AI companies and developers. Laws should be supplemented with other measures such as ethical guidelines, technological solutions that minimise adverse effects such as disinformation, the institution of cybersecurity measures and the enhancement of technology literacy that fosters greater public vigilance against emerging threats.

AHEAD Africa fills the gap 

AHEAD Africa embraces the notion of ensuring a holistic and responsive framework for AI in elections as it aligns with the objective of contributing to higher levels of electoral integrity in African Union member states. Furthermore, AHEAD Africa views this paper as a call to action in various respects. First, since the paper benefited only from a limited number of African case studies, it invites a deeper inquiry into the state of AI involvement in Africa’s elections. AHEAD Africa will answer this call through a series of research initiatives in the following areas:

(1) The prospects and challenges of using AI in Africa’s elections;

(2) Consolidating best practices of tracking digital hate speech, misinformation, and disinformation and;

(3) Safeguarding human rights and democratic values in the midst of digital civic space regulations and enforcement of cybersecurity. 

In addition to research, AHEAD Africa will cultivate dialogues among stakeholders through platforms such as Experiential Learning Seminars, Election Academies and Election Integrity Summits. These platforms will harness contemporary experiences and perspectives on the implementation of policies and practices that relate to AI and more broadly, technology in electoral governance. AHEAD Africa will also stay engaged with relevant policy dialogues at the regional and continental level, such as on the implementation of the African Charter on Democracy, Elections and Governance (ACDEG) and the African Union Convention on Cyber Security and Personal Data Protection. It is through the convergence of learning, dialogue and advocacy initiatives, that we hope to make a contribution which steers the use of AI in elections towards strengthening electoral integrity in Africa.