Generative AI Poses New Threats to Election Integrity, CSDI Report Warns

Generative AI Poses New Threats to Election Integrity, CSDI Report Warns

The rise of generative AI (Gen AI) has introduced fresh challenges to democratic processes, particularly elections, according to a new report by the Centre for the Study of Democratic Institutions (CSDI), a Canadian think tank. The report, titled "Harmful Hallucinations: Generative AI and Election," highlights the increasing risks that AI technologies pose to election integrity and democratic norms.

Authored by Chris Tenove, Nishtha Gupta, Netheena Mathews, and others, the report delves into the impact of generative AI on elections, focusing on the potential for misinformation, manipulated media, and deepfakes. The researchers express concern that while generative AI is not entirely new, its growing accessibility and rapid advancements have made it easier to produce deceptive content at lower costs. "Generative AI technologies lower the cost of producing deceptive content, amplifying existing threats to democracy," said Chris Tenove, assistant director of CSDI.

The report notes that 2024 has been dubbed the year of “Deepfake Elections,” reflecting numerous incidents where generative AI has been employed in election campaigns in the US, India, and the European Union. The researchers categorize the risks associated with generative AI into three main areas: deception, harassment, and pollution of information environments. The report provides real-world examples illustrating how AI can mislead voters, harass political candidates, and overwhelm information channels with low-quality, inaccurate content.

One notable example cited in the report is a deepfake video of US President Joe Biden that circulated in New Hampshire earlier this year. The deepfake, spread through robocalls, falsely urged voters to reserve their votes for the general elections rather than participating in the primaries, a tactic aimed at suppressing voter turnout.

As elections are imminent in several countries, including the US, India, and Brazil, the CSDI's study aims to assess the tangible impacts of generative AI on electoral processes and explore potential solutions to mitigate these risks. "There is significant hype and concern around generative AI's potential impact on politics and elections. Our goal was to evaluate the real evidence of harmful uses and their consequences, and to identify strategies to address these threats," Tenove stated.