Meta’s ad library contained more than 50 image and video ads with AI-generated child sexual abuse imagery. Some of the ads reportedly ran across Facebook, Instagram, Messenger, or Threads as recently as this week.
The finding matters because paid ads pass through a commercial review pipeline, not just ordinary user posting. If abusive synthetic images can reach that system, platforms face pressure to improve detection before content is distributed or monetized.
AI-generated abuse material creates a difficult moderation problem because it can be produced at scale and may not match known-hash databases built for previously identified illegal images. Platforms therefore need classifiers, human review, reporting systems, and advertiser enforcement to work together.
The report does not show that every detection failure has the same cause. It does show that generative AI is making platform safety problems harder, especially when harmful content can be packaged as advertising.