Assistant Professor Shweta Singh of Warwick Business School commented:

“Deepfakes are an increasingly big problem for trust, elections and democracy across Britain. With the election just around the corner, it’s more important than ever that voters know that what they’re looking at is real. We’ve recently seen voters and even pundits and journalists fooled by deepfakes, including a video purporting to show Nigel Farage playing ‘Minecraft’.

“We know that most voters can spot deepfakes when they see them, but in elections when results can be decided by small handfuls of votes we still need to do what we can to tackle the threat of misinformation through deepfakes.

“I believe all AI generated content should be watermarked as such, so that at least the vast majority of video made through AI programmes is clearly marked, making it much harder to pass off as real. Whilst these watermarks can sometimes be removed, we must place obstacles in the way of the people producing these deepfakes. Each method we have of distinguishing real from fake helps us in the fight against political disinformation.”