Detect Deepfakesby Resemble AI
Deepfake case study · Video

[Tech] - Meta ordered to remove UK deepfakes as oversight

Meta oversight board orders removal of deepfakes targeting a Scottish councillor and a Muslim activist, citing inadequate platform safeguards

Incident date
Sep 2026
Target
Labour party councillor in Scotland and a Muslim campaign volunteer
Updated Sep 18, 2026 · 1 min read

Meta’s independent oversight board has mandated the removal of deepfake content from Facebook, ruling that the company’s current approach to AI-generated imagery is fundamentally inadequate. The decision follows two specific cases involving a Scottish Labour councillor and a young Muslim campaign volunteer, both of whom were the subjects of malicious, AI-manipulated videos.

What happened

The board identified two primary incidents that highlighted failures in Meta’s content moderation policies. In the first case, an AI-generated video falsely depicted a Labour party councillor in Scotland making inflammatory remarks about refugees. The board noted that the audio was not fully synchronized with the councillor’s facial movements, a common indicator of AI manipulation. Despite this, Meta initially determined the content did not violate its policies and refused to apply an AI label. The oversight board concluded that the video violated rules against hateful conduct by attributing predatory sexual behavior to refugees as a group.

In the second incident, the board ordered the removal of an AI-generated video mocking a young Muslim woman engaged in a menstrual health education campaign. The video depicted the woman performing absurd exercises and eating junk food, which the board ruled was a breach of its bullying and harassment policy. These manipulated videos reached tens of millions of views across various platforms.

Following these investigations, the oversight board characterized the pattern as a systemic issue where women in public discourse are disproportionately targeted by harassment. The board issued nine policy recommendations to Meta, including the expansion of "high risk" AI labels, the implementation of click-through warning screens for deceptive content, and stricter penalties for accounts that repeatedly share deepfakes. The board emphasized that Meta must improve the transparency of its data regarding AI labels and ensure that algorithms demote content identified as high-risk to limit its spread.

Sources