More than 250 additional advertisements containing child sexual abuse material ran across Meta's platforms in the weeks after the company rolled out new AI tools designed to detect and block exactly that kind of content, according to research from the nonprofit Tech Transparency Project shared with WIRED and published Tuesday. The finding undercuts the company's claim that its AI-powered ad screening had largely closed the gap, and it comes as lawmakers say they plan to investigate.
The new figures extend an earlier disclosure. Last month, Meta deleted roughly 50 ads posted on Facebook, Instagram and Threads that directly included child sexual abuse material, following previous reporting by WIRED. At the time, the company said most of the abusive ads had been published before it deployed the new AI detection tools. Instead of the expected decline, researchers say they found hundreds more of the ads in recent weeks — including some identical to ones Meta had just removed. For continuous coverage of platform safety and the latest AI news, follow AI Buzz Wire.
The Scale of the Failure
In total, researchers at the Tech Transparency Project have documented more than 350 of the abusive video ads running on Meta's platforms since the end of last year, spanning Facebook, Instagram, Messenger, Threads and Meta's broader advertising network. Unlike the initial set of 53 ads the group identified, the newer batch included images of real children rather than purely synthetic ones. WIRED confirmed that one source image came from the official website of a European royal family; researchers say an officially released photograph of a minor member of the family was manipulated into a video depicting a sex act. The researchers are not naming the royal to protect the victim and say they contacted officials in the relevant country.
Katie Paul, the group's director, told WIRED that the problem is not limited to AI-generated imagery. "What we're talking about here is not simply AI-generated child sex abuse material," she said, noting that many images appear to have been taken from the open web before being used in the ads. "There's a lot of harm that can come from this all while Meta is raking in money in the process." The researchers identified the real-world identities of four minors whose images were used in the ads.
A Funnel Into Nudification Apps
The ads follow a consistent pattern, according to the reporting. Many open on an innocuous picture of a preteen or teenage girl — the researchers found only one featuring a boy — overlaid with text declaring the image is "not just a photo" and that there are "no restriction" on what can be done with it. When the short video plays, the child's face is morphed into an explicit video. Clicking through leads to AI face-swapping and "nudification" apps available in Apple's and Google's app stores, many linked to Chinese developers. WIRED separately identified hundreds of related nudification ads depicting adults on Meta's platforms.
Slow Reviews and Missing Disclosures
Meta's policies state that the company reviews all ads that run on its platforms, and the ads violated its rules on child exploitation, adult exploitation and nudity. In practice, the researchers reported, it sometimes took Meta up to a week to review live ads that had been reported through the company's own tools — during which the ads accumulated hundreds of additional views, according to screenshots reviewed by WIRED. When some of the ads were removed, their takedown records initially lacked any disclosure that they violated the company's child sexual abuse policies; those details appeared only after WIRED asked Meta about the discrepancy.
The researchers reported every ad they found to the National Center for Missing and Exploited Children's CyberTipline and through Meta's reporting tools. Meta's own ad library shows the ads were collectively displayed to more than 29,000 accounts across European Union countries and around 7,000 accounts in the United Kingdom, with more than 250 appearing in the United States, 120 in Australia and 80 in India — figures that likely understate total reach, since some countries lack granular ad performance data.
Meta Responds
Meta pushed back on the findings' significance. "We don't tolerate nudify apps or any kind of child exploitation, whether real or AI-generated," company spokesperson Tracy Clayton told WIRED, adding that the company has previously removed hundreds of thousands of ads for violating its nudification policies. Meta said many of the ads the researchers found had already been removed, that most received fewer than 200 impressions, and that it collected under $5,000 in ad payments. The company also said the AI-generated videos featured adults while the clips involving minors were brief or blurred, making detection more difficult, and that it reports child exploitation content to NCMEC.
Lawmakers and Regulators Move
Multiple lawmakers and regulators now say they are investigating or looking into the ads following WIRED's reporting. Virginia Senator Mark Warner, who sent Meta CEO Mark Zuckerberg a letter after the earlier reporting, said it is "disappointing that Meta continues to run ads featuring child sexual abuse material and nonconsensual intimate images after it was brought to their attention." Michigan's attorney general said the office is "actively looking into the matter," Florida's attorney general posted that his office is investigating, and Australia's eSafety regulator said it is "highly concerned" and has requested information from Meta at a senior level.
The scrutiny adds pressure to a company already paying up to $16.7 billion to settle lawsuits alleging its social networks harmed children, and appealing a New Mexico court decision that would require it to improve its CSAM reporting processes. The broader question for the AI industry is whether automated moderation can keep pace with abuse that is itself increasingly AI-assisted — and evidence that a detection system failed to catch hundreds of violating ads weeks after being cited as the fix will feed directly into those debates.
Stay Ahead of AI
Follow the latest AI developments and breaking artificial intelligence news as regulators and platforms wrestle with AI-powered harms.
Read more AI news →
