Full Breakdown
Meta Oversight Board Orders Removal of Deepfake Videos and Calls for Policy Overhaul
By Drooid · · How we work
Core Decision and Content at Issue
The Oversight Board, Meta’s quasi-independent content referee, ordered the removal of two AI-generated videos from Facebook. The other showed a young Muslim woman campaigning for menstrual-health education performing absurd exercises and eating junk food, a portrayal the board classified as bullying and harassment. Both posts had been left up after Meta initially declined to label them as AI-generated or to apply its “high-risk” warning.
Background and Board Authority
Created in 2020, the Oversight Board serves as a binding adjudicator for Meta’s content policies. Its decisions are intended to guide the platform’s moderation practices worldwide. In this case, the board determined that the videos violated Meta’s hateful-conduct rules by attributing criminal behavior to an entire group of refugees and by targeting a private individual with manipulated imagery.
Official Findings and Recommendations
The board described Meta’s existing safeguards as “consistently and fundamentally inadequate” to curb the rapid rise of deepfakes. It recommended nine policy changes, including:
- Expanding “high-risk” AI labels and ensuring algorithms demote such content.
- Introducing click-through warning screens before users can view AI-generated material.
- Redefining “unwanted manipulated imagery” to cover deepfakes of private individuals.
- Increasing penalties for repeat offenders and improving transparency around label application.
The board also noted that the woman-targeted videos had amassed “tens of millions of views” across Meta platforms, underscoring the scale of the problem.
Impact and Broader Concerns
Board co-chair Pamela San Martin emphasized that deepfakes are being used to harass and silence women in public discourse. The rulings highlight a pattern in which women who speak on social issues face disproportionate misinformation attacks, prompting calls for stronger platform safeguards.
Verbatim Quotes
- “From politicians to private citizens, AI-generated deepfakes are increasingly being used to harass and silence women from engaging in public discourse,” — Pamela San Martin, an oversight board co-chair