AI news ·
Meta runs more than 300 paid ads containing child sexual abuse imagery
Meta ran over 300 paid Facebook and Instagram ads containing AI-generated child sexual abuse material, including manipulated photos of real minors. Some ads were placed by Meta's own authorized resellers in China, and the company removed about 150 only after investigators shared their findings.
A Tech Transparency Project investigation found that Meta ran more than 300 paid ads on Facebook and Instagram containing child sexual abuse material (CSAM) this year. The ads, which included digitally altered images of real children-among them a minor member of a European royal family-were approved by Meta's review systems and generated revenue for the company.
Most of the ads followed a single format: a clip of adult pornography followed by a photo of a child manipulated with AI to depict the child performing a sex act. TTP identified multiple photos of real children taken from social media posts and websites that were used in these ads. The investigation also found that several CSAM ads were placed by Meta's own authorized advertising reseller partners in China, including a state-controlled company.
Meta did not respond to a request for comment from TTP. Hours after TTP shared its findings with Meta, the company removed about 150 ads still visible in its Ad Library and revised violation notices for over a hundred others to reflect child exploitation policies rather than generic adult-content rules. The company told Wired and Bloomberg that most of the ads had fewer than 200 impressions and total ad spend was under $5,000. "We don't tolerate nudify apps or any kind of child exploitation, whether real or AI-generated," Meta said in a statement.
Real children's photos used in AI-manipulated ads
TTP researchers identified the CSAM ads through the Meta Ad Library, which retains ads run in the European Union and United Kingdom for one year. The 332 ads found included images of children ranging from toddlers to young teenagers. Many showed a still photo of a child that was digitally animated into a video of sexual abuse.
One ad featured a photo released last year by a European royal family of one of its youngest members. The image was manipulated to show the minor performing a graphic sex act. Another ad used a photo a 14-year-old influencer had posted to Instagram showing her new sports club uniform. The ad animated the image into a video depicting oral sex. A third set of ads used stock photos of a preteen girl in athletic wear, which morphed into a video of an adult man removing her shirt and molesting her.
The ads carried captions and voiceovers that were explicit about their purpose. One voiceover said, "There are no restrictions. All the characters are real people and there are many options. This is the AI men actually use." Another ad displayed the text, "Our AI transforms your hidden desires into hypnotic videos. Dark, seductive, unforgettable."
Chinese ad resellers placed CSAM content
TTP found that three of Meta's authorized ad reseller partners in China-GIMC, Meetsocial, and BlueFocus-ran ads containing the same CSAM content. These resellers are part of a system that channels Chinese advertising onto Facebook and Instagram despite the platforms being banned in China. Reuters has reported that Meta exempts these companies from immediate punishment for policy violations.
GIMC, a state-controlled Chinese business and one of Meta's 11 authorized resellers in China, placed an ad showing a girl in a sexual position with text reading "I can show you more." The ad directed users to a VPN app with more than 100,000 downloads. BlueFocus, which works closely with the Chinese Communist Party's main propaganda outlet and was blocked by U.S. national security officials from merging with an American firm in 2018, also placed a CSAM ad directing to a VPN app with over 500,000 downloads.
Meta has given all three reseller companies industry performance awards in recent years and pays them a 10% commission on ads purchased through their accounts. None of the three companies responded to TTP's requests for comment.
Nudify apps and fake accounts
Of the 332 CSAM ads identified, 298 directed users to AI image- and video-generator apps, with 253 pointing to apps from China-based developers. TTP verified that many of these apps were capable of "nudifying" people in photos or inserting them into sexual videos. Apple removed three nudify apps from its App Store after TTP shared its findings. Google suspended all flagged apps from its Play Store.
Apple said it has "zero tolerance for developers who deliberately attempt to evade our review process" and described the apps as engaging in "bait-and-switch behavior." A Google spokesperson said the company took "swift action, which in the case of these apps has included suspending hundreds of violating apps and restricting related search terms like 'nudify' on our store."
Many of the CSAM ads were linked to Facebook pages for what appeared to be fake people or entities-accounts with no content, few followers, and cloned profile images. One page listed a Chicago address and phone number that answered with an unintelligible voice recording. Another tried to pass itself off as being from the Church of Jesus Christ of Latter-day Saints. These accounts violate Meta's policies requiring authentic identities and prohibiting coordinated inauthentic behavior.
Enforcement gaps and reporting failures
TTP reported 129 active CSAM ads to Meta. The company took at least a week to review the reports. In 57% of cases, Meta responded that the ad did not violate its Advertising Standards. In 43% of cases, Meta said the ad had already been removed, though six remained visible in the Ad Library. One ad showing a real preteen girl reached an additional 330 users while Meta deliberated over the report.
The investigation also revealed that Meta was inconsistent with identical ads. The company removed one CSAM ad but told TTP that an identical copy did not violate its standards. TTP further found that archived CSAM ads in the Ad Library remain visible to anyone, and there is no mechanism to flag them to Meta. Most EU ads did not fulfill Digital Services Act transparency requirements, with disclosure fields left blank or filled with gibberish.
The ads pose potential legal questions for Meta. While Section 230 of the Communications Decency Act broadly protects platforms from liability over user-generated content, it does not extend to federal criminal violations, including laws covering sexual exploitation of children. The FBI has warned that CSAM created with AI-generated images is illegal, and the material at issue here appeared in paid advertisements approved and run by Meta.
Why this matters for legal, policy, and communications professionals
For legal and regulatory affairs teams, the investigation surfaces direct questions about platform liability when AI-generated CSAM appears in paid advertising rather than organic user posts. The distinction between Section 230 protections and federal criminal statutes on child exploitation is central here. Professionals working in AI regulatory compliance and content moderation policy will need to track whether this prompts legislative or enforcement action on ad-review obligations. For communications and marketing leaders, the finding that Meta's own authorized resellers placed CSAM ads complicates brand safety assessments and vendor due diligence for any organization advertising on Meta platforms. Policy professionals focused on AI public policy should note the cross-jurisdictional dimension: ads ran in the EU, UK, Australia, India, and the U.S., each with distinct transparency and child safety regulations now being tested by AI-manipulated content in paid media.