Ofcom has warned it will penalise online platforms that fail to tackle the rising number of illegal deepfakes as researchers develop a new process for detecting them using techniques normally deployed in astronomy.

The regulator commissioned research that showed that two in five people have seen at least one deepfake in the last six months that includes depictions of sexual content, politicians and scam adverts.

Concern over their proliferation on social media is ramping; in January, research was published revealing that more than 100 deepfake video ads impersonating Prime Minister Rishi Sunak had reached 400,000 people on Facebook as part of a concerted smear campaign ahead of the general election.

According to Ofcom, among adults who say they have seen deepfake content, one in seven (14%) say they have seen a sexual deepfake. Most of the available evidence indicates that the overwhelming majority of this content features women, many of whom suffer from anxiety, PTSD and suicidal ideation...

  • A review of the article within Nature "AI deepfake faces detected using astronomy methods" dated 22 July 2024 has the following comment by Brant Robertson :-

    However if you can calculate a metric that quantifies how realistic a deepfake image may appear, you can also train the AI Model to produce even better deepfakes by optimizing that metric.

    https://doi.org/10.1038/d41586-024-02364-y

    Peter Brooks

    Palm Bay