Meta accused of approving AI-generated child abuse advertisements

Meta has found itself at the center of a new controversy regarding its ad moderation and child safety policies. Researchers from the Tech Transparency Project (TTP) announced that they identified dozens of paid advertisements containing what they described as AI-generated child sexual abuse material during an examination of the company's public Ad Library.

According to the research, the advertisements in question were published between November 2025 and early August 2026. It was reported that the ads reached users in the US, the UK, and more than 12 European countries, with a single ad shown in Europe reaching more than 2,500 accounts.

TTP Director Katie Paul stated that the content was not leaked from outside the platform; rather, these were advertisements that had been reviewed and approved by Meta's systems and from which the company generated revenue. The researchers argued that some of the ads were so explicit that there was no need to hide what the images were promoting.

NEW ADS DETECTED EVEN AFTER WARNINGS

According to the investigation, a significant portion of the ads directed users to "nudifying" applications that depict individuals as if they were undressed without their consent. It was reported that the application named MaskAI was removed from the App Store after Wired notified Apple about the issue.

The researchers stated that they initially found about two dozen violating ads, and that they identified 30 more new ads in the hours after Wired questioned Meta about the matter. This strengthened the claim that the system continued to approve similar ads even after the company was warned.

It was also alleged that some of the ads matched exactly with content that Meta had previously removed for the same violation. It was reported that most of the advertising accounts had no or very few followers, and that the advertiser information included China-based firms and a former Meta advertising intermediary.

META: CONTENT IS NOT ACCURATE, ACCESS WAS LIMITED

In a statement, Meta indicated that it did not find the content in question to be accurate and argued that access was limited. The company also announced that it removed 36 million pieces of child abuse content last year.

However, the company's history in the field of child safety has led to the new allegations being discussed in a broader context. Sources recalled that Meta faced more than 50 thousand sexual content reports every month in 2017, and that a change in the way Messenger shared information with law enforcement in 2019 affected approximately 7.5 million child abuse reports.

Leaked internal research in 2021 had fueled debates regarding Instagram's impact on the mental health of young girls. In 2023, the Wall Street Journal reported that a joint study by Stanford and UMass Amherst alleged that Instagram algorithms suggested and connected networks of pedophile accounts to one another.

Meta's obligations to protect children were also brought to the fore in a lawsuit filed by the New Mexico Attorney General's Office. In the case, which was stated to have concluded in March 2026, a jury ordered Meta to pay 375 million dollars in damages on the grounds that the company endangered children and misled the public.

The latest allegations have brought back questions about how harmful AI-generated images can pass through advertising systems and whether the automated moderation mechanisms of major platforms are sufficient in terms of child safety.