Facebook told users that pages posting AI-generated videos of older coaches sexually touching young female athletes didn’t break its rules, then removed the pages only after a reporter contacted Meta, according to Futurism. Futurism flagged 10 such accounts through Facebook’s normal reporting tools, and a week later Facebook said eight didn’t violate its standards, even though Meta’s policies ban sexualized depictions of children, including AI-generated ones. Futurism also found that after it filed the reports, Facebook began recommending more of the same videos, some with over 100,000 views. Once Futurism contacted Meta’s press team, all the pages came down, and Meta said its own systems found 98% of child endangerment content last quarter before any user reported it.
