New research from The Alan Turing Institute's Centre for Emerging Technology and Security (CETaS), warns that regulators must urgently address AI-related threats, to preserve trust in the UK's democratic system ahead of the July general election.
While the research finds limited evidence of AI directly altering election outcomes, it highlights early signs of damage to the broader democratic system.
The study reveals that out of 112 national elections since January 2023 or forthcoming in 2024, only 19 had examples of AI-enabled interference. However, current ambiguous electoral laws could lead to misuse of AI in the upcoming election. Potential threats include fake campaign endorsements, deepfakes inciting online hate, and confusion over AI-generated content.
To address these concerns, the researchers propose several recommendations. They urge the Electoral Commission and Ofcom to create guidelines for political parties on AI use in campaigning and require clear marking of AI-generated election material. They also suggest publishing new guidance for media reporting on AI-generated content, including AI threat awareness in voter information, and coordinating simulation exercises to prepare for potential AI-related election disruptions.
Sam Stockwell, Research Associate and lead author, emphasised the urgency of regulatory action given the imminent election. Dr. Alexander Babuta, Director of CETaS, stressed the importance of helping the public distinguish fact from fiction, to maintain faith in the democratic process.
As the UK approaches its July general election, this research underscores the critical need for swift regulatory action, to address potential AI-related threats to the electoral process. While the direct impact of AI on election outcomes remains limited, proactive measures are essential to preserve public trust in democracy.