In June 2025, Meta announced it would be “building new technology to detect ads for nudify apps.” More than a year later, the company continues to allow thousands of ads for nonconsensual deepfake nude generators to run on Facebook and Instagram.
Since February, at least 11,336 ads have appeared on Meta’s platforms for websites that promise the ability to undress photos of real people and place them in sexually explicit videos using AI. The company eventually removed 85.7% of the ads, leaving 1,616 without any moderation action.
These ads reached millions of users, pushing traffic to the underlying apps and websites.
As part of a moderation pivot announced in January 2025, Meta CEO Mark Zuckerberg promised to focus on proactive enforcement of “illegal and high-severity violations.” Meta has since said that a shift to more AI-based systems “can catch more of these violations more accurately.”
Yet at least 5,909 of the ads in our sample promoted domains or publishers that Indicator previously documented in our reporting on this topic and shared previously with Meta. The company apparently failed to use this information in reviewing new ads.
Today’s audit raises significant questions regarding Meta’s ability to enforce its policies, and its oversight of ad partners based in China. In a July report, the Tech Transparency Project found that thousands of ads for nudifiers were placed via one of Meta’s largest Chinese ad partners. (At least 40 ads in our sample appeared to be connected to that company, based on name similarity.)
Katie Paul, executive director of the Tech Transparency Project, told Indicator that “many of these ads appear to be getting a pass when they are run by Chinese ad resellers that generate billions in revenue for Meta.” Paul says that Meta “has the technology to stop these ads before they upload, but doing so would mean they miss out on ad revenue for tens of thousands of ads. Even if Meta removes the ads later, they have already run, reached users, and padded the company's revenue.”
Meta told Indicator on Wednesday evening that our findings represented another instance of criminal actors adapting their tactics to circumvent moderation and that they would review and action the ads in our sample.
Get the full story — and the data
Members get access to all of the assets identified in our articles.
UpgradeUpgrade now to access:
- The full story
- The spreadsheet with all of the actors we identified in this report
- A walkthrough of how we conducted our investigation
