Meta Fails to Stop Hundreds of Child Sexual Abuse Ads, Including Real Children, Despite AI Detection Claims

Editor’s note: This article contains descriptions of imagery depicting child sexual abuse. Reader discretion is strongly advised.

Early last month, Meta faced significant scrutiny after it was revealed that the company had deleted approximately 50 ads from Facebook, Instagram, and Threads that directly contained child sexual abuse material (CSAM). This revelation followed an initial report by WIRED, prompting Meta to claim that most of these abusive advertisements had been published before the rollout of new AI tools specifically designed to "better detect and block" harmful ads. However, recent findings suggest that Meta’s AI system has been far from effective, indicating a pervasive and alarming failure to safeguard its platforms.

Instead of a decline, researchers have uncovered hundreds more disturbing ads featuring graphic sexual scenes involving children in recent weeks. Shockingly, some of these newly discovered ads were identical to those Meta had previously removed, raising serious questions about the efficacy and responsiveness of their content moderation systems. The Tech Transparency Project (TTP), a non-profit organization, shared its detailed findings with WIRED, reporting that over 250 additional ads containing CSAM have run across Meta’s social media platforms since the beginning of August. Unlike the 53 ads initially identified by TTP researchers, a critical and deeply concerning aspect of these new ads is the identification of images of real children, not merely AI-generated content.

These new discoveries include profoundly disturbing content depicting a minor from a European royal family. An officially released photograph of the royal was transformed into a video showing a "graphic sex act." To protect the victim, the researchers have refrained from naming the royal and have confirmed they contacted officials in the respective country. WIRED independently verified that the source image indeed originated from an official website of a royal family, underscoring the real-world impact and vulnerability created by these malicious advertisements.

"What we’re talking about here is not simply AI-generated child sex abuse material," stated Katie Paul, the director of the TTP, highlighting the grave escalation of the issue. Paul emphasized that it appears some images were directly taken from the web before being integrated into these abusive ads. She added a stark observation regarding the financial aspect: "There’s a lot of harm that can come from this all while Meta is raking in money in the process." This points to a disturbing scenario where illicit content generates revenue for the platform, even indirectly, while victims suffer severe exploitation.

Cumulatively, TTP researchers have identified that Meta has published more than 350 such abusive video ads since the end of last year. A significant portion of these ads directed users to harmful "nudification" apps, many of which are linked to Chinese developers. In every instance of discovering a live ad, TTP researchers immediately reported it to US child safety authorities, specifically the National Center for Missing and Exploited Children’s (NCMEC) CyberTipline, and utilized Meta’s own reporting tools. WIRED, upon reviewing TTP’s detailed findings, independently confirmed hundreds of other related "nudification" ads featuring explicit adult content running on Meta’s platforms, illustrating the broader scope of the problem.

While Meta has since removed the child abuse ads, citing violations of its policies on child exploitation, adult exploitation, and nudity, their initial publication remains a critical point of concern. Meta’s advertising standards explicitly state that all ads running on its platforms undergo review. Yet, these hundreds of ads circumvented that review process. TTP researchers noted instances where it took Meta up to a week to review live ads they had reported for containing CSAM. During this critical delay, screenshots seen by WIRED confirm that these ads continued to accumulate hundreds of additional views, further exposing users to illegal content. Furthermore, TTP observed that when some child sexual abuse ads were initially removed, they did not include disclosures indicating a violation of Meta’s child sexual abuse policies. These crucial details, TTP reports, only appeared to be added after WIRED specifically questioned Meta about the discrepancy, suggesting a reactive rather than proactive approach to transparency and accountability.

The cache of video ads, published across Facebook, Instagram, Messenger, Threads, and Meta’s wider advertising network since late last year, follow a disturbingly consistent pattern. Many commence with an innocuous picture of a preteen or teenage girl – with only one instance identified featuring a boy. Overlaid text typically states it is "not just a photo" and that "there are no restriction" [sic] on what can be done with the image. As the seconds-long video ads progress, the children’s faces are morphed into graphic sexual videos, often depicting them performing sex acts. Clicking on these ads directs users to download AI-face swapping or video apps available on both Apple and Google’s app stores. The pervasive tagline, "This is the AI that men actually use," underscores the predatory nature and target audience of these applications.

The researchers were able to confirm the real-world identities of four minors whose images were exploited in these ads. Three of these individuals are from the United States. Katie Paul elaborated, "There were several photos from the same teen who has a social media influencer account on one of Meta’s platforms." Another instance involved a different teen influencer with a public account. Paul further explained, "One was a stock photo model that was a real person. It was actually labeled as pre-teen on the stock photo websites." The fourth confirmed identity was the aforementioned member of a European royal family. All findings related to these ads were reported to the NCMEC’s CyberTipline, and TTP is consulting on appropriate methods to disclose their discoveries to those impacted.

The global reach of these abusive ads is extensive. Data from Meta’s ad library, which archives both live and removed ads, shows that these harmful advertisements were collectively shown to more than 29,000 accounts across the European Union, in addition to approximately 7,000 accounts in the United Kingdom. However, the true reach is likely far greater, as more than 250 of these ads appeared in the United States, 120 in Australia, and 80 in India, countries where Meta does not provide granular ad performance data. This suggests that thousands more individuals, including potentially vulnerable minors, may have been exposed to this content.

In response to these grave findings, Meta spokesperson Tracy Clayton stated, "We don’t tolerate nudify apps or any kind of child exploitation, whether real or AI-generated." The company highlighted its past actions, noting it has previously removed hundreds of thousands of ads for violating its policies on nudification. Clayton’s statement further claimed that "many" of the ads identified by TTP researchers had already been removed, and most had fewer than 200 impressions, with Meta reportedly receiving "under $5,000" in ad payments.

In an email to WIRED, Meta also indicated that it reports child exploitation content to NCMEC and is reporting these specific ads to the organization. However, Meta attempted to minimize the findings, asserting that the AI-generated videos primarily featured adults, and clips of minors were either brief or blurred, making detection more challenging. Clayton added, "The criminals behind this activity know we’ve built strong defenses, including systems that block violating ads as well as ongoing monitoring in case we initially miss something." This explanation, however, rings hollow given the sheer volume of detected ads and the documented delays in their removal.

TTP’s discovery of hundreds of child abuse material ads coincides with Meta facing significant legal challenges. The company is set to pay up to $16.7 billion to settle lawsuits alleging its social networks have harmed children. Furthermore, Meta is currently appealing a New Mexico court decision that, if enforced, would mandate improvements to its CSAM reporting processes and require human reviews within the state. These ongoing legal battles underscore the systemic nature of the challenges Meta faces in protecting minors on its platforms.

The renewed reporting by WIRED has prompted reactions from multiple lawmakers and regulators. Virginia Senator Mark Warner expressed his disappointment, stating, "It is disappointing that Meta continues to run ads featuring child sexual abuse material and nonconsensual intimate images after it was brought to their attention." Warner had previously sent a letter to Meta CEO Mark Zuckerberg following WIRED’s initial reporting, which, according to Warner’s office, has yet to receive a reply. He urged, "Meta needs to follow their own advertising standards, heed the call, and put an end to these illegal and inappropriate ads." (Meta has since informed WIRED it will reply to the letter).

Similarly, a spokesperson for Michigan’s attorney general confirmed their office "is aware of reports concerning AI-generated child sexual abuse material appearing in advertising on Meta. This is extremely troubling, and we are actively looking into the matter." Florida Attorney General James Uthmeier also publicly announced on X that his office is "investigating." Internationally, Australia’s eSafety internet regulator conveyed its "highly concerned" stance regarding the reports and is assessing Meta’s compliance with its rules, requesting "information on this matter from the company at a senior level."

Lacking Protections

Over the past six years, the online landscape has seen a proliferation of dozens of AI nudification or "undress" apps and websites. These technologies have been widely used to harm countless women and girls, generating millions in revenue annually for those behind them. As this dangerous ecosystem of nonconsensual intimate imagery (NCII) continues its unchecked growth, child protection organizations have issued urgent warnings that the technology is "increasingly" being exploited to create CSAM. Emma Hardy, director of communications at the Internet Watch Foundation, a UK-based nonprofit dedicated to removing child abuse content, emphasized this danger: "AI apps are making it far too easy to create realistic criminal imagery of child sexual abuse, creating the new risk of everyday photos of children being taken and transformed into sexual content." Hardy’s statement concluded with a direct call to action for platforms: "Platforms and companies must make sure there is nowhere these apps can be advertised and spread."

As WIRED previously reported, some of the ads initially discovered by TTP researchers were traced back to Chinese companies operating as ad "resellers" in a country where Meta’s platforms are officially banned. TTP’s analysis further indicates that many of the apps promoted by these ads also appear to have strong Chinese connections. This is not Meta’s first encounter with such entities, as the company has previously launched legal action against a Hong Kong-based company linked to a set of nudifier platforms.

The abusive ads directed users to almost 50 different applications available across both Google’s Play Store and Apple’s App Store. Out of the 350 ads, approximately 300 promoted AI image or video generator apps, with the majority linked to China-based developers. By the time WIRED began its deeper investigation, some of these apps had already been removed from the stores. Among the apps designed to generate explicit imagery of adults, eight identical applications were found on Apple’s App Store, several linked to Chinese developer names. Upon opening, these apps displayed explicit AI-generated content, including a menu of sexual scenarios, disturbingly featuring one titled "threatening knife," depicting a knife held to a woman’s throat. WIRED independently identified another seven identical apps sharing the same design, explicit content, and developer contact details.

Apple spokesperson Adam Dema addressed these findings, stating, "The App Store was designed to be a safe and trusted place for users, and we have zero tolerance for developers who deliberately attempt to evade our review process." Dema explained that "The developers mentioned in this report initially submitted apps that are compliant with the App Review Guidelines, only to later surface nudify capabilities, which are against our rules." The company confirmed that 17 of the 20 apps included in the TTP research had already been removed, and action had been taken against the remaining three, along with the seven additional apps WIRED identified.

Google spokesperson Dan Jackson reiterated the company’s policy: "We do not allow apps that contain sexual content or the ability to nudify." Jackson added, "We continually take proactive steps to detect and remove apps with harmful content," confirming that all apps cited in the report had been removed. "When violations are reported to us, we investigate and take 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," he stated.

David Pham, a former Meta employee who investigated nonconsensual imagery and abuse on its platforms, noted the increasing prevalence of Chinese developers in creating nudify apps. Pham recounted observing similar ads to those discovered by TTP during his tenure at Meta, explaining the constant cat-and-mouse game played by those uploading the content to bypass safety systems. "Every time the platform or investigators would build a new detection model or a new way to have some type of enforcement, they quickly—within a day, maybe five days—built a new way to avoid detection," Pham detailed. These tactics included changing domains, obscuring the identity of posters, and subtly altering or blurring imagery to evade detection. Despite these challenges, Pham, who left Meta last year, firmly believes the company possesses the necessary funding and resources to effectively tackle these uploads. "I would like to live in a world where we’re not getting CSAM or NCII material sent directly to our phones," he asserted.

Meta’s current protections against ads containing CSAM and NCII clearly demonstrate significant deficiencies. A recent report by Indicator Media revealed that over 11,000 "undress" ads have appeared on Meta’s platforms since February of this year, with many reportedly posted by "repeat offenders." Independent analysis by WIRED further uncovered hundreds of ads containing explicit content with language strikingly similar to those identified by TTP researchers. (Meta claims it has "incorporated" the language highlighted by WIRED into its moderation systems, suggesting an ongoing effort to improve detection.)

One particularly egregious example from TTP’s research involved an ad featuring one of the teenagers whose real-world identity was discovered. This ad was initially reported to Meta when it had only reached four accounts in Europe. However, by the time it was finally removed days later, it had been seen by more than 330 users. Katie Paul from TTP articulated the profound frustration: "This is not something that it should take a small nonprofit doing research to stumble across and then have to investigate and report. We’re talking about one of the most well-resourced—technologically and financially—companies on the planet." The continued prevalence of such deeply harmful content on Meta’s platforms, despite its vast resources and public commitments, underscores a critical and ongoing failure in protecting its users, especially children.