Meta's Oversight Board Orders Removal of AI Deepfakes Targeting Women
Meta's Oversight Board ordered the company to remove two AI-generated deepfake videos—one depicting a Scottish Labour councillor making inflammatory comments about refugees, and another manipulating footage of a Muslim health advocate—citing "fundamentally inadequate" safeguards. The board found that Meta's policies and labeling practices failed to prevent deepfakes from being used to harass and silence women in public discourse.
What happened
Meta's Oversight Board on September 17, 2026 ordered Meta to remove two AI-generated deepfake videos from Facebook and Instagram. The first video, posted without AI-generated disclosure, depicted a Scottish Labour Party councillor making incendiary comments about refugees, with audio not synchronized to facial movements. The second video manipulated existing TV interview footage of a Muslim health advocate to ridicule her by falsely depicting her consuming junk food while giving dietary advice.
Reported impact
The Oversight Board found that the deepfakes were used to harass and silence women from engaging in public discourse. One woman reported the effect of the deepfake video as "quite traumatic." The board stated that AI-generated deepfakes are "increasingly being used to harass and silence women from engaging in public discourse," with women who engage publicly on issues disproportionately subjected to harassment and misinformation.
Index analysis
The Oversight Board determined Meta's moderation policies are "consistently and fundamentally inadequate" to address the proliferation of AI-generated content. The board found Meta did not remove the first video when reported because it had not been flagged by "trusted partner" organizations and did not appear to interfere with voting processes. In both cases, Meta failed to label the videos as AI-generated content despite the board's own policies requiring such labeling.
Governance controls
The Oversight Board's decisions on individual cases are binding on Meta. Meta said it welcomed the board's decision and committed to comply within seven days. The Oversight Board recommended Meta adopt stronger approaches to labeling deceptive content, stricter penalties for repeat offenders, and removal of deepfakes harassing private individuals.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.