Meta Faces Escalating Scrutiny After Hundreds of AI Abuse Ads Reach Users Across Its Platforms

Meta hosted more than 300 advertisements containing suspected child sexual abuse material this year, according to the Tech Transparency Project, with ads appearing across Instagram, Facebook, Threads and Messenger. An earlier investigation identified more than 50 ads, indicating that the issue continued or expanded after Meta removed an initial batch. The advertisements promoted services that generated or altered sexualized images of children, including tools marketed to “nudify” real people.

The ads collectively reached more than 29,000 people, according to reporting cited by Ars Technica, although reach varied by ad and region. Some material reportedly used images of real children as inputs, raising separate concerns about privacy, consent and the creation of fabricated abuse imagery from identifiable victims.

The central failure involved Meta’s advertising systems, which approved paid promotions that violated the company’s own rules on child exploitation and non-consensual intimate imagery. Researchers located the advertisements in Meta’s public advertising library, then flagged examples that remained active before the company removed them.

One Facebook page presenting itself as a Church of Jesus Christ of Latter-day Saints organization reportedly ran a CSAM advertisement and was disabled only after the Tech Transparency Project raised it with Meta. The episode has drawn attention to the gap between automated screening, human review and post-publication enforcement. Meta has removed ads and pages identified by investigators, but the continuing discoveries indicate that detection did not stop the material before distribution.

The case involves both AI-generated abuse imagery and the misuse of photographs of real children. Hacker News AI reported that some of the 350 advertisements examined by investigators used images of actual children, including a member of a European royal family, to create sexualized material. That expands the issue from synthetic content to the exploitation of recognizable victims and the spread of images without consent.

The promoted services also targeted photos of young girls posted to Instagram. Their advertising model connected ordinary social-media images with tools designed to create fake nude images, exposing minors to sexual abuse and harassment even when the original photographs were nonsexual. Child-safety groups, including 5Rights Foundation, have documented Instagram accounts with large audiences sharing real and AI-generated CSAM, with some accounts reaching 150,000 followers.

Government pressure is building on several fronts. India directed Meta to immediately disable advertisements and content on Instagram that promoted or facilitated child sexual abuse material, after a BBC investigation found paid promotions of such material on the service. The government also sought an explanation for how the ads passed platform controls.

In the United States, Sen. Mark Warner asked Meta CEO Mark Zuckerberg to explain how ads featuring CSAM and non-consensual intimate images passed review, how long they ran and whether similar advertisements remain active. Warner’s letter frames the matter as a broader accountability issue, including the company’s responsibility to prevent paid distribution of illegal content. The parallel actions in India and Washington raise the cost of treating the problem as an isolated moderation error.

Meta’s stated rules prohibit child sexual abuse material and non-consensual intimate imagery, yet the advertising system still accepted promotions for services associated with both categories. That contradiction gives investigators a clear test: whether Meta can prevent repeat violations rather than remove individual ads after outside groups identify them.

The financial structure matters as well. Paid distribution gives advertisers a direct route to audiences and gives Meta an economic incentive to process ads at scale. Katie Paul of the Tech Transparency Project told WIRED that harm can continue while Meta earns advertising revenue. The findings also put pressure on the company’s claims about safety work around generative AI, because the same systems that enable synthetic abuse are being marketed through Meta’s own commercial infrastructure.

The immediate next steps are removals, government inquiries and further testing of Meta’s ad review systems. India’s directive requires action on the platform, while Warner is seeking detailed answers from Meta about enforcement, monitoring and continued exposure. Child-safety organizations are also pursuing legal and public pressure after reporting that Meta ignored or delayed responses to flagged material.

For Meta, the issue reaches beyond content moderation. The company must show that its advertising library, automated classifiers, page verification and human escalation procedures work together against rapidly changing AI services. Investigators will continue checking whether removed ads reappear under new pages, whether similar promotions run in additional countries and whether enforcement covers both generated abuse imagery and tools that transform real children’s photos.