Meta at the Center of Scandal — Researchers Found Ads for Nudify Products
In December 2025, we reported that Meta had removed 134 million fraudulent ads over the course of a year. But the story didn't end there — researchers recently discovered a large number of examples containing AI-generated child sexual abuse material (CSAM) in the ad library.
Today we're analyzing this case and breaking down how AI-generated content passed moderation — and whether anything is likely to change going forward.
CSAM Content in Meta Ads
Wired and BBC published reports almost simultaneously in early August, revealing that researchers had found over 50 ads containing CSAM content running across Facebook, Instagram, Messenger, and Threads. Some of the campaigns promoted apps with image generation and modification capabilities, and certain ads had already accumulated thousands of impressions.
The issue was uncovered by researchers at the Tech Transparency Project (TTP), who spotted ads for AI services offering so-called nudify content. They found ads that didn't merely promote such tools — they contained prohibited content outright. Some creatives led to AI apps capable of altering images and superimposing faces onto sexualized scenes.
The ads discovered by researchers were published between November 2025 and early August 2026. Many remained active for several days, and some were no longer serving at the time of discovery — yet remained accessible in the ad library for several months.
Researchers also documented instances of identical ads being reposted. Most tellingly, approximately 30 new creatives were identified after WIRED had already contacted Meta about the issue.
The scale of the campaigns was relatively small — most ads received only a handful of impressions. However, at least one of them, according to the Meta Ad Library, reached 2,563 accounts in the EU. Meta does not disclose reach statistics for the US and other countries, meaning the actual audience could have been significantly larger.
The most striking detail is that some of the ads made no attempt to conceal the actual product behind a neutral creative. Researchers found ads that explicitly used sexualized imagery of minors alongside copy designed to push users further down the funnel.
In one case, an ad led to an AI service where, after clicking through, the content was modified so that a child's face from the ad creative appeared in a video with explicit 18+ context.
Meta stated that it removed the identified ads for policy violations. The company also noted that some of the ads had minimal reach, and that many had been blocked prior to WIRED's inquiry. At the same time, Meta acknowledged that some of the violating ads slipped past the AI systems designed to detect and block policy violations at the point of ad upload.
Why Isn't Meta's AI Moderation Working?
Meta uses automated ad review before publication and may re-review creatives after a campaign goes live. The system analyzes images, video, text, and other components, while also factoring in advertiser behavior and associated accounts. When necessary, a human moderator is brought in to assist.
In June 2025, the company announced it was developing technology to detect ads for nudify products. The system was designed to look not only for obvious violations within creatives, but also for copies of previously removed ads, newly created accounts, and other signals characteristic of such advertisers.
Meta acknowledges that combating this type of abuse is an ongoing arms race. Advertisers rotate creatives, spin up new accounts and domains, and continuously attempt to evade detection. Blocking a single ad is therefore not enough — the system must identify related campaigns and new variations after every change.
The platform already has dedicated AI systems for detecting such content, but they offer no guarantee that a prohibited ad won't make it into the auction. And the faster generative tools produce new creative variations, the harder it becomes for the system to rely on known patterns to block them.
At the time of writing, the creatives are no longer available in the Ads Library
Conclusion
Meta has more than enough tools and resources to tackle ads that blatantly violate its ad network policies. But years of watching Facebook suggest that these tools work precisely to the extent required by the balance between advertising revenue growth and public backlash.
