Meta's Oversight Board has again criticised the company's handling of AI-generated content, releasing two cases that demonstrate Meta's rules are "consistently and fundamentally inadequate" to address the growing problem. The cases involved deepfake videos used to harass individuals on the platform, highlighting a troubling pattern in which AI-generated content is weaponised against people—particularly women—to silence or ridicule them in public discourse.

In the first case, a Scottish politician's likeness was used in a synthetic video making hateful statements, which Meta initially refused to remove because it hadn't been flagged by "trusted partner" organisations and the person who posted it hadn't disclosed it was AI-generated. The second case involved manipulated footage of a menstrual health education volunteer that went viral across social media, which Meta also failed to label as AI-generated and only took limited action against after the Oversight Board intervened. The board says both cases exemplify how women engaging in public issues face disproportionate harassment through deepfakes and misinformation, whilst Meta's labelling system—currently consisting of small "AI Info" banners—remains inadequate and the company has made few substantive changes to its AI content policies.