Meta's automated systems approved AI-generated child abuse imagery in ads, researchers find

Over 50 ads containing AI-generated child sexual abuse imagery appeared across Facebook, Instagram, Threads, and Messenger (Nov 2025–Aug 2026), per TTP research reported by Wired. Current detection relies on matching known images, like matching a fingerprint — but AI-created images are new, so they slip through. Meta's backup filters also failed. Meta did not dispute the ads were approved and shown.