Meta’s ad review system approved and ran paid ads containing AI-generated child sexual abuse material. Not user posts, not third-party content slipping through the feed, but actual paid advertisements that Meta reviewed, approved, and collected money for. According to Engadget, researchers at the Tech Transparency Project (TTP) identified more than 50 of these ads running across Facebook, Instagram, Threads, and Messenger between November 2025 and August 2026.
TTP found the ads through Meta’s own ad library, a first-party tool the company offers to catalog where ads run and who they target. The ads included images of children alongside sexually suggestive text, and several linked out to so-called “nudify” apps that digitally undress photos of people. “These ads made no effort to mask the images or hide what they were promoting,” TTP Director Katie Paul told Wired. The point she’s making is hard to argue with: this wasn’t sophisticated evasion. These ads were obvious, and they still got through.
Meta’s advertising standards explicitly ban content that “sexually exploits or endangers children,” and the company says it reports offending material to the National Center for Missing and Exploited Children. But the review process relies on automated tools, which is almost certainly where the breakdown happened. Meta acknowledged the ads in a statement to Wired, saying “the majority of these ads had minimal reach” and that many were taken down before Wired contacted the company. It also noted that some of the ads predate a newer AI detection system it recently launched to catch policy violations at upload.
That explanation has obvious problems. Automated moderation failing to catch explicit CSAM is a serious gap, not a minor technical edge case. And Meta is not a small company operating on thin margins. It generated over $160 billion in revenue in 2024. The infrastructure to review ads more carefully exists. So does the financial incentive to avoid exactly this kind of scandal.
This is also not an isolated incident. In July, the BBC reported that Instagram ran ads promoting child sexual abuse material in India. In April 2026, the Consumer Federation of America sued Meta over scam ads that similarly violated the platform’s own standards. A pattern is forming, and it points to a moderation operation that is structurally underbuilt relative to the scale of the ad business it supports.
For developers and founders building on Meta’s ad platform, or using its APIs, this matters beyond the obvious ethical dimension. Repeated failures at this level increase the likelihood of regulatory intervention, particularly in the EU where the Digital Services Act creates direct liability for platforms that fail to control illegal content. Advertisers also face reputational risk simply by appearing alongside this kind of content. And any AI tool or app that connects to Meta’s ecosystem inherits some of that exposure by association.
