Meta has removed hundreds of ads featuring child sexual abuse material (CSAM) from its platforms, but researchers found even more. Since August, 250 new ads have appeared, some containing graphic scenes of real children. One even used a photo of a royal family member, transforming it into a video of a sex act. Meta’s AI tools, meant to detect harmful ads, have proved ineffective, with ads lingering for weeks before removal. The company has made over $5,000 from these ads, despite policies against child exploitation.

The ads typically show young girls, with misleading text suggesting no restrictions on what viewers can do with the images. When clicked, they direct users to AI face-swapping apps. Meta claims to have removed thousands of such ads, but researchers found many still in circulation. The company says it reports such content to NCMEC, but critics argue more needs to be done to prevent these disturbing ads from reaching vulnerable children.

Among the images used, four minors were identified, three from the US, and one from India. Meta’s ad library shows the ads were viewed by over 29,000 accounts across the EU, with 7,000 in the UK. These ads, while technically removed, highlight the ongoing struggle to protect children online. As AI becomes more prevalent, the gap between detection and prevention widens, leaving us with more questions than answers.