Meta removes AI deepfake videos targeting women after

Meta has removed two AI-generated videos targeting women from its platforms following a ruling by the company's Oversight Board, which criticized existing safeguards against deepfakes as fundamentally inadequate. The board's intervention highlights a growing pattern of synthetic media being used to harass and silence women in public life.

What happened

The two removed videos targeted specific individuals: a Labour Party councillor in Scotland and a Muslim campaigner in Europe. The video featuring the Scottish councillor falsely depicted her making inflammatory remarks about refugees. Evidence of manipulation included a mismatch between the speaker’s audio and facial movements, and there was no record of the politician having made the comments shown.

The second video targeted a Muslim campaigner in Europe in an attempt to undermine her credibility as a health advocate. It depicted her eating unhealthy food while promoting mindful eating and performing exaggerated exercises.

The Oversight Board determined that both videos violated Meta’s policies regarding misinformation, hate speech, and harassment. The board emphasized that the current volume of synthetic media appearing on platforms like Facebook and Instagram exposes significant shortcomings in Meta's moderation approach. In response to the board's binding ruling, Meta agreed to comply with the removal orders within seven days. Oversight Board Co-chair Pamela San Martin has since called for stronger measures, including clearer labeling of deceptive AI content, tougher consequences for repeat offenders, and the removal of deepfakes used to harass private individuals.