Meta's Oversight Board Calls for Clearer Rules on AI-Generated Sexual Content

Meta's Oversight Board has criticized the company's unclear rules regarding sexually explicit AI-generated depictions of real people. The board reviewed two cases involving famous women, finding both in violation of Meta's policies. It recommended the company update its guidelines and criticized its reliance on media coverage for policy enforcement.

Meta's Oversight Board has sharply criticized the tech giant for having unclear rules concerning sexually explicit AI-generated depictions of real people, calling for urgent reforms to prevent such imagery from proliferating on its platforms.

This critique emerged after the board, which operates independently despite being funded by Meta, examined two cases involving pornographic deepfakes of well-known women. The subjects, identified only as public figures from India and the United States due to privacy concerns, were depicted in images that violated Meta's "derogatory sexualized photoshop" rule. This rule categorizes such content as bullying and harassment, and the board asserted that Meta should have acted swiftly to remove the offending images.

In one case, Meta failed to act on a report within 48 hours, allowing the ticket to close automatically. Meta reversed its decision only after an appeal to the board. In another instance, automatic systems removed the AI-generated image without human intervention. The board urged Meta to update its rules to include all forms of AI editing and to improve its image database policies, criticizing its current reliance on media coverage as dangerous and insufficient to protect victims of deepfake pornography.

(With inputs from agencies.)