Most people cannot spot AI deepfakes in scam test, survey finds
Related: Public sceptical over new Grok safeguards as Ofcom launches investigation
Consumer group Which? has warned that realistic deepfakes are leaving the public “at the mercy of scammers” after an experiment revealed many adults could not identify AI-generated content.
A Which? survey of 514 adults found that seven in 10 failed to correctly distinguish between real and fake videos, with 8 per cent getting every one wrong and 21 per cent identifying only a single correct clip.
One convincing deepfake showed Boots staff promoting an unlicensed weight loss product, prompting the retailer to state these adverts are fake and urge customer vigilance.
Which? is urging the government to ensure regulator Ofcom can take decisive action against technology companies that fail to block fraudulent schemes, by implementing robust codes of practice under the Online Safety Act.
The government responded by stating scammers are criminals and platforms have a responsibility under the Online Safety Act to proactively tackle illegal fraudulent content, including deepfake scam ads.
