Experts have warned that the use of artificial intelligence (AI) to generate or enhance images in politics poses a significant threat to democratic processes, following a Labour MP's apology for sharing a manipulated image of Prime Minister Rishi Sunak. Karl Turner, MP for Hull East, posted an image on X (formerly Twitter) showing Sunak pouring a sub-standard pint at the Great British Beer Festival, with a woman looking on derisively. The image was altered from an original where Sunak's pint appeared acceptable and the woman had a neutral expression.
The Conservative Party criticised the post, with Deputy Prime Minister Oliver Dowden calling it 'unacceptable' and urging Labour leader Sir Keir Starmer to disown it. Dowden told LBC that MPs who shared the image should delete it, as it was 'clearly misleading'. The row has highlighted concerns about the potential for AI-manipulated content to disrupt the upcoming general election, expected next year.
Academics have called for urgent action. Professor Wendy Hall of the University of Southampton stated that 'the use of digital technologies including AI is a threat to our democratic processes' and should be prioritised on the AI risk register, especially with major elections in the UK and US approaching. Dr Shweta Singh of the University of Warwick emphasised the need for ethical principles to assure users of news trustworthiness, warning that without regulation, fair elections are impossible.
Professor Faten Ghosn of the University of Essex urged politicians to clearly mark any AI-manipulated content, citing US efforts to require disclosure in political adverts. Labour MP Darren Jones questioned how the public can identify deepfakes and challenged the Science Secretary on government action. The Science Department is consulting on its AI white paper, which advocates general principles rather than specific bans, though Sunak has recently called for 'guardrails' on AI development.
Major AI companies, including Amazon, Google, Meta, Microsoft, and OpenAI, have agreed to watermark AI-generated content. Microsoft President Brad Smith warned that governments must address AI-generated disinformation by early 2024 to protect elections. The row underscores growing political concern over AI regulation ahead of a fiercely contested election campaign.
