Meta's AI content rules 'fundamentally inadequate,' oversight board rules

The Oversight Board, Meta's independent body for content decisions, has published two new rulings stating that Meta's rules for AI-generated content are "consistently and fundamentally inadequate" to address the rapid growth of such content on its platforms. Both cases involved AI-generated videos, including one that the board said should have been removed for violating Meta's hateful conduct rules and labeled prominently as AI-generated.

In the first case, an AI-generated video depicted a Scottish politician making offensive statements. The board said the video was synthetic but realistic. Meta told the board it did not remove the video when reported because it had not been flagged by its "trusted partner" organizations and did not appear to interfere with voting or other processes. The post was also not labeled as AI-generated. After the board's decision, Meta agreed to remove the video.

The second case involved an AI-manipulated video of a volunteer who had appeared in a TV interview about menstrual health education. According to the board, accounts across social media platforms digitally altered the footage to ridicule her, and the videos went viral. Meta automatically closed a report and an appeal about the post, which was never labeled as AI-generated. Meta removed harassing comments only after the board became involved, and later said it would remove the post.

The board said the cases show a broader pattern in which women engaging publicly on issues are disproportionately subjected to harassment and misinformation. It has recommended that Meta prioritize likely AI content for review when reported. Meta has 60 days to respond to the recommendations.