AI's Influence on Elections Ignored at UN Summit, Experts Warn

The recent Global Dialogue on AI Governance held in Geneva has drawn attention for its failure to adequately address the intersection of artificial intelligence and electoral processes. While representatives from all 193 member states of the United Nations gathered to discuss the governance of AI, critical issues surrounding its impact on elections were largely overlooked, a gap that experts argue could jeopardize democratic integrity worldwide.

In the context of AI's role in elections, discussions often gravitate towards disinformation tactics like deepfakes. However, the real concern lies in how AI is integrated into the very systems that manage elections. The UN's preliminary report from its Independent International Scientific Panel on AI echoed this sentiment, focusing on information integrity rather than the structural risks posed by AI-enabled technologies in electoral management.

Take India as a cautionary example. The country's election commission initiated a program in 2015 that connected voter ID records to Aadhaar, a biometric database controlled by a government body rather than an independent electoral authority. This initiative utilized a machine learning algorithm to identify duplicate or deceased voters, resulting in the removal of approximately 5.5 million voters. The Supreme Court intervened, but the program was reinstated in 2021, demonstrating a troubling reliance on flawed algorithms that could disenfranchise voters.

This situation highlights a broader trend where electoral authority is increasingly ceded to private technology vendors, especially in regions like Africa and Latin America. Many election technologies are designed and operated by entities that elude national oversight, which raises serious concerns about data security and accountability. As AI tools become more prevalent in voter registration and verification processes, risks multiply, including the potential for sensitive data to be mishandled or for algorithmic errors to go unnoticed.

Moreover, the use of AI-driven chatbots for voter assistance can exacerbate existing disparities. Systems like OpenAI’s ChatGPT may inadvertently disseminate biased or inaccurate information, particularly in areas with high linguistic diversity and limited digital access. This further marginalizes already underrepresented communities and complicates the electoral landscape.

Critically, without stringent governance standards, election officials may find themselves responsible for outcomes influenced by technologies they cannot fully regulate or inspect. As one expert noted, the accountability for algorithmic decisions may fall upon institutions that lack the necessary control over the systems in question.

To address these concerns, the African Union has taken a proactive stance by establishing guidelines for electoral commissions regarding AI technologies. The AU has emphasized the importance of controlling the AI ecosystem to prevent corruption in voter registration, a move that underscores the need for a governance model that prioritizes electoral integrity.

As discussions around AI's impact on various sectors grow, the importance of addressing its influence on elections cannot be overstated. Ensuring citizens retain the power to elect or remove their leaders is fundamental to democratic accountability. By neglecting electoral issues in AI governance, the Global Dialogue risks undermining the very foundation of democracy. As the next meeting in New York approaches in 2027, it is imperative that electoral integrity becomes a focal point in the global AI governance agenda.

Source: restofworld.org