May 29, 2024

AI deepfakes and election misinformation

A general election has just been called in the UK, and many other elections are planned around the world in the coming year. At the same time, rapid developments in generative AI models have led to the rise of increasingly convincing digitally altered/generated content or ‘deepfakes’. In this major election year, there have been concerns over the abilities of sophisticated AI-generated deepfakes having the potential to cause widespread misinformation and even influence election outcomes.

But what is the technology underpinning these deepfakes and can technology also provide the solution to mitigating against them?

And while there has been a lot of speculation over the influence of deepfakes and generative AI, what is the evidence on whether they have a real-world impact on elections? And what can be done in the short and long term to reduce the likelihood of AI threats, and in particular, for the UK general election? A new briefing paper from the Centre for Emerging Technology and Security (CETaS) at The Alan Turing Institute has analysed the current evidence to try and answer these questions.

Journalists came to this SMC briefing to hear both a background on all things deepfakes from some AI experts, as well as from the lead author of the briefing paper from The Alan Turing Institute on the findings and their implications.

Speakers included: 

Prof Philip Torr, Professor of Engineering Science, Five AI/Royal Academy of Engineering Research Chair in Computer Vision and Machine Learning, University of Oxford

Dr Christian Schroeder de Witt, Post Doctoral Researcher, AI & Security Research, University of Oxford

Dr Björn Ross, Assistant Professor in Computational Social Science at the University of Edinburgh School of Informatics

Sam Stockwell, Research Associate, Centre for Emerging Technology and Security (CETaS) at The Alan Turing Institute