Meta Approved Dozens of Ads With AI-Generated Child Abuse Imagery

Meta reviewed and approved dozens of paid ads that contained AI-generated child sexual abuse imagery, according to a Wired investigation built on research from the Tech Transparency Project. The ads ran across Facebook, Instagram, Threads, and Messenger, and several linked out to 'nudify' apps that digitally undress subjects in photos.

TTP found the ads sitting in Meta's own public ad library, posted between November 2025 and August 2026. TTP director Katie Paul told Wired that 'these ads made no effort to mask the images or hide what they were promoting,' which is the uncomfortable part of the story. This isn't a case of clever evasion beating a filter, it's an ad-review pipeline waving through material that a human reviewer would flag in a second.

Meta's response, as reported by Engadget, is that 'the majority of these ads had minimal reach and many were disabled before WIRED shared them,' and that many predate newer AI detection tools the company has since deployed. Take that framing as the company's own. Paul's counter, per Wired, is that the findings contradict Meta's public claims to prohibit deepfake-nude ads and suggest ad revenue is being prioritized over enforcement.

Why this matters beyond the obvious moral floor: brand-safety adjacency on Meta's surfaces is now a live concern for anyone buying Facebook or Instagram inventory this quarter. If the review system approves CSAM ads at all, even at low reach, every advertiser is one bad placement away from a screenshot going viral. Regulators watching the 'nudify' app category, which parallel reporting has tied in part to a Beijing-based advertiser called GatherOne running thousands of undressing-app ads, now have a sharper story to point at.

The honest caveat is that the reporting does not quantify 'dozens' precisely, does not say how many impressions the ads actually served, and does not tell us whether the accounts behind them were referred to law enforcement. Meta's claim that new detection tech has closed the gap is exactly that, a claim. The people who stand to benefit, if anyone, are the trust-and-safety vendors and rival platforms that can credibly say they caught this earlier, and the regulators who have been asking why ad libraries keep surfacing violative content the platform's own review approved.