Meta approved and ran more than 50 paid ads containing AI-generated child sexual abuse material across Facebook, Instagram, Messenger, and Threads over the last nine months, according to research shared with WIRED by the Tech Transparency Project. The ads targeted users in the United States, United Kingdom, and more than a dozen European countries, with at least one reaching 2,563 accounts across eight European nations. Many linked out to so-called nudify apps that use generative AI to strip or face-swap images into sexual content.
The ads ran between November of last year and the start of August, and some were still live this week. TTP researchers initially surfaced about two dozen listings in Meta's ad library, then found roughly 30 more in the hours ahead of publication — several of them posted after WIRED first contacted Meta about the earlier batch. Meta has since removed the ads for violating its policies on child sexual exploitation and adult nudity.
One video ad used a thumbnail of a child sitting on the floor overlaid with the text "Realizing Deep Fantasies with Generation AI. There is so much more than what is shown, use your imagination," according to TTP. Another repeatedly posted ad showed a young girl in a sexual position with the caption "I can show you more." TTP director Katie Paul says the ads were reviewed and approved through Meta's advertising pipeline, not posted as organic user content.
Key facts
- 01Tech Transparency Project researchers found more than 50 image and video ads containing AI-generated CSAM in Meta's ad library, published between November and early August.
- 02At least one ad reached 2,563 accounts across France, Germany, Ireland, Italy, the Netherlands, Spain, Sweden, and the United Kingdom.
- 03Roughly 30 additional violating ads surfaced hours before publication, many posted after Meta was first contacted about the initial two dozen.
- 04Several ads linked to MaskAI, a Chinese-developed nudify app that Apple has now pulled from the App Store.
- 05Meta says it removed 36 million pieces of child sexual exploitation content last year and shared thousands of violating URLs with the Tech Coalition's Lantern service.
Meta's own policies say every ad is reviewed before publication, primarily by automated tools. The company said in a statement that most of the ads had minimal reach, many were disabled before WIRED shared them, and that many predate new AI detection technology deployed at ad upload. Meta said it removed 36 million pieces of child sexual exploitation content last year and has taken legal action against nudify app developers.
The disclosure is the second time in recent weeks that paid ads tied to child sexual abuse material have surfaced on Meta's platforms. In early July, a BBC investigation documented Instagram ads in India promoting the sale of what appeared to be CSAM, with links to Telegram channels selling the illegal content. Meta responded then with a statement citing a "zero tolerance" approach to child sexual exploitation.
Where advertisers were disclosed in the ad library, they were largely linked to Chinese companies. TTP identified one advertiser as Meet Social, previously one of Meta's ad resellers handling China, where Meta's platforms are blocked. A 2019 New York Times profile said Meet Social published thousands of ads on Facebook per day and expected more than $1 billion in sales. Meet Social did not respond to a request for comment.
Several of the ads linked to MaskAI, an app that surfaced in Apple's App Store without visible nudity in the store listing. Once installed, WIRED's review found it contained only AI-generated pornography and offered face-swap features for placing real people into sexual situations. Apple removed MaskAI after being contacted, saying it has "always strictly prohibited" apps designed to generate, distribute, or consume pornography.
“This raises significant questions about how seriously Meta is taking child exploitation in paid advertisements if its own systems can identify identical ads but it does not remove those identical ads for CSAM material.”— Katie Paul, Director, Tech Transparency Project
Paul says the TTP tranche included ads that appeared identical to ones Meta had already removed for policy violations elsewhere, suggesting the company's matching systems are not carrying enforcement decisions across the ad library. She also notes that Meta's ad library has no in-product mechanism for users to report violating ads once they have stopped running.
AI nudification services have proliferated since around 2020 as diffusion models became cheap and accessible. The category has generated millions for operators, produced a documented flow of nonconsensual imagery of adults and minors, and adapted quickly around each enforcement action. The Revenge Porn Helpline in the UK, which Meta cited to WIRED, called nudify apps "a genuinely adversarial problem" because of how rapidly the ecosystem routes around any single point of enforcement. Sean Litton, outgoing CEO of the Tech Coalition, said Meta has shared thousands of violating URLs with the group's Lantern service.
The gap in this case is not detection of a novel jailbreak or an obscure model — it is paid ad review of images that pattern-match to categories Meta has publicly committed to blocking, on inventory that generates revenue. Automated ad approval optimized for throughput is being outpaced by an adversarial supply chain that is willing to iterate faster than reviewer models are retrained, and the ad-library persistence issue means once an ad clears review, it stays discoverable for months even after it stops serving.
For the broader AI industry, the episode sharpens a regulatory argument already forming in Brussels and Washington: that platforms hosting generative-AI-adjacent advertising should carry liability that mirrors the liability applied to the underlying model providers. Meta's ad business is one of the largest downstream distribution channels for nudify-app operators, and enforcement at the ad layer is where the economics of this category live or die. If the pattern of identical ads slipping through repeat review continues to be documented, expect the pressure to shift from voluntary coalitions like Lantern toward statutory ad-review requirements — a cost line that would fall on every large ad platform, not just Meta.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
