Meta Allowed 350+ Child Abuse Ads Despite New AI Detection Tools

Meta's advertising systems approved and displayed more than 350 ads containing child sexual abuse material across Facebook, Instagram, and Threads since late 2025, according to new research from the nonprofit Tech Transparency Project shared with WIRED.

The ads promoted AI "nudification" apps and included graphic sexual content depicting minors. In several cases, researchers identified images of real children—including a member of a European royal family whose official photograph was transformed into explicit video content.

The findings come just weeks after Meta removed approximately 50 similar ads following earlier WIRED reporting. At that time, the company attributed most violations to ads posted before it deployed new AI detection tools designed to block harmful content. The latest research suggests those systems remain inadequate.

Ads featured real children and ran for days

Unlike the initial batch of ads, which used AI-generated imagery, TTP researchers confirmed that some of the newly discovered ads contained photos of real minors taken from the internet. The organization identified four victims: three U.S.-based teenagers with social media influencer accounts, one stock photo model labeled as "pre-teen" on commercial sites, and the European royal family member.

The ads followed a consistent pattern: they displayed an innocuous photo of a child with text claiming "there are no restriction" on what could be done with the image. When the video played, the child's face was morphed into explicit sexual content. Clicking led users to download face-swapping apps from Apple and Google's app stores.

Meta's ad library data shows the content reached more than 29,000 accounts in the European Union and 7,000 in the United Kingdom. Additional impressions occurred in the United States, Australia, and India, though Meta doesn't publish granular metrics for those markets.

Detection failures and delayed responses

All ads were supposed to undergo Meta's standard review process before publication. Yet hundreds cleared initial screening. When TTP researchers reported live ads through Meta's own tools, some remained active for up to a week, accumulating hundreds of additional views.

One ad that had reached four European accounts when reported was seen by more than 330 users by the time Meta removed it days later. When ads were eventually taken down, disclosure labels indicating child exploitation policy violations were only added after WIRED inquired about their absence.

Meta spokesperson Tracy Clayton said the company has "zero tolerance" for such content and noted that "many" of the flagged ads had already been removed, with most receiving fewer than 200 impressions. The company earned under $5,000 from the violating ads, according to Clayton, who added that Meta reports child exploitation content to the National Center for Missing and Exploited Children.

Many of the ads originated from Chinese companies acting as ad resellers and linked to apps with apparent Chinese developer connections. The apps themselves—nearly 50 across Apple's App Store and Google's Play Store—often disguised their true functionality during app store review processes, only revealing explicit content after approval.

Why it matters

This represents a fundamental failure of automated content moderation at scale. Meta generated billions in quarterly revenue while its systems repeatedly approved ads exploiting real children, some of whom were identifiable public figures. The company's claim that new AI tools would "better detect and block" such content has proven demonstrably false. For enterprise technology leaders, this underscores the limitations of relying solely on automated systems for high-stakes content decisions—and the reputational and legal risks when those systems fail. Multiple U.S. state attorneys general and Australia's eSafety regulator have now opened investigations.

Regulatory scrutiny intensifies

Virginia Senator Mark Warner, who sent Meta CEO Mark Zuckerberg a letter following the initial reporting, said it's "disappointing that Meta continues to run ads featuring child sexual abuse material" after being alerted to the problem. Attorneys general in Michigan and Florida confirmed they are investigating, while Australian regulators requested information from Meta at senior levels.

The violations occurred as Meta faces up to $16.7 billion in settlements over lawsuits alleging its platforms harm children. The company is also appealing a New Mexico court decision that would require improved CSAM reporting processes including human review.

Details of Meta's advertising failures were first reported by WIRED, based on research conducted by the Tech Transparency Project.