Meta approved and ran paid ads with AI-generated child sexual abuse material, researchers find

Meta sells ad review as a pre-publish quality filter to advertisers and to the public. Researchers now have a catalogued case where that filter cleared more than 50 paid ads containing AI-generated child sexual abuse material between November 2025 and August 2026, on the one category of content with no legitimate advertising use, while the company collected the ad dollars.

The Tech Transparency Project (TTP), a research group that monitors major tech platforms, identified the ads inside Meta's own first-party ad library, a public catalog that indexes paid campaigns running across the company's apps. The ads ran on Facebook, Instagram, Threads, and Messenger, and several linked out to "nudify" apps, which are services that generate stripped-down images of real people without their consent.

"Meta's ad review approved and ran these ads," TTP Director Katie Paul told Wired. "These ads made no effort to mask the images or hide what they were promoting. It's important to point out that this isn't content posted by third parties on Facebook or Instagram — these are ads that were reviewed, approved, and allowed to run by Meta, never encountering interference while the company collected the ad dollars."

Meta's advertising standards explicitly forbid ads that "contain content that sexually exploits or endangers children," and the company says it reports such material to the National Center for Missing and Exploited Children (NCMEC), the U.S. nonprofit that acts as the national clearinghouse for child exploitation reports. Sexually explicit adult ads are also banned from the ad system. The more than 50 ads Paul described cleared all three gates.

Meta catches ad violations by reviewing each ad before publication, using automated tools as part of that pipeline. The more than 50 ads documented by TTP, combining images of children with sexually suggestive text and other offensive content, cleared that gate. Researchers described the ads' reach as thousands of accounts before removal.

The ads are citable because they sit inside Meta's own ad library, which any researcher with a browser can query. TTP reproduced its count of more than 50 ads from that system directly. The ads were not pulled from a dark corner of the platform; they were public, paid, and indexed. TTP's count is an approximation rather than a precise total. The organization describes "more than 50" ads and notes the figure is reproducible from the library, not exhaustive.

Meta's on-record response, carried by Wired and Engadget, accepts the existence of the ads but reframes their significance. "Sexual exploitation is horrific, and we work aggressively to keep it off our platform," a Meta spokesperson told Wired. "The majority of these ads had minimal reach and many were disabled before WIRED shared them. Many of these ads also predate new AI technology we launched recently to better detect and block violating ads at upload — and we're constantly improving these systems. From removing over 36 million pieces of child sexual exploitation content last year to taking legal action against nudify app developers…"

The defense addresses three things: scale of effort, timeliness of detection technology, and downstream enforcement. The pre-publish review pipeline approved paid placements of the one category of content with no legitimate advertising use, and the catalogued artifact remained in Meta's own ad library for the duration of the campaign. Even under Meta's framing of minimal reach and many removals before Wired's outreach, the placement happened, the ad dollars were collected, and the ad stayed in the library.

TTP is running a separate investigation into a Chinese partner behind a deluge of nudify ads on Meta platforms, and several of the CSAM ads in the more-than-50 set linked to nudify-app destinations. The overlap suggests the catalogued ads are a slice of a larger advertising pipeline, one where nudify apps pay Meta for distribution, and where the same app category supplies the destination for the worst-of-the-worst ad creative.

The next test is whether the new detection tool Meta cites changes the catalogued result. Can the ads still be retrieved from the ad library today? Has the company disclosed what share of CSAM-flagged submissions in the new pipeline now arrive as ad-review rejects rather than post-publication removals? And is the nudify-app pipeline itself subject to the same pre-publish review, or only the creative it submits?