Meta Missed 332 AI Child Abuse Ads Targeting Real Children
Meta removed 332 AI child abuse ads from Facebook and Instagram after delays, as US and Australian regulators scrutinize its worldwide ad review failures.
Summary
On September 8, 2026, the Tech Transparency Project said Meta missed 332 AI-generated child sexual abuse material ads on Facebook and Instagram this year, after 53 were flagged in August and Meta announced stronger detection. Most promoted China-made AI apps, often nudify tools. Ads transformed real photos of a young European royal, a 14-year-old Instagram influencer and a preteen stock model into graphic abuse videos; unmatched victims were overwhelmingly preteen girls. The Justice Department says AI CSAM is as harmful as other CSAM. Reach exceeded 29,000 people in the EU and 6,800 in the UK; about 80 percent ran in the US, with some in Australia. Meta said average impressions were below 200 and total spending below $5,000.
Meta removed all flagged ads, but dozens ran for days; only about half of more than 300 disappeared from its archive within hours. One reported at launch rose from four impressions to 330. Identical ads received opposite rulings, reports took at least a week or no action, and many removals were not initially labeled child exploitation, indicating possible undercounting. Chinese partners GIMC, Meetsocial and BlueFocus carried identical content. Duplicate ads came from fake pages Kenny Maddie Regan, Lynn Nicole Brooks and Briflabremgles Nihendrimes; Divine Productions impersonated the Church of Jesus Christ of Latter-day Saints. Meta blamed brief or blurred clips.
The findings follow Meta’s $18 billion child-safety settlement. Campaign for Accountability executive director Michelle Kuppersmith questioned Meta’s entire review system, while TTP director Katie Paul warned delays prolong harm as Meta earns money. Michigan and Florida attorneys general are investigating, Senator Mark Warner demanded action in August, and Australia’s eSafety regulator requested senior-level information after banning social media for under-16s. Meta plans to answer Warner. Internet Watch Foundation communications director Emma Hardy urged platforms to block these apps from advertising.
Positives
- Meta removed all CSAM ads flagged by investigators and disabled the deceptive Divine Productions page after notification.
- The $18 billion settlement requires Meta to make platform changes intended to improve child safety.
- Michigan and Florida attorneys general are investigating the ads, while Senator Mark Warner has demanded Meta stop them.
- Australia’s eSafety regulator requested senior-level information from Meta following its social media ban for children under 16.
Risks & concerns
- 332 AI-generated CSAM ads escaped Meta’s detection this year, even after 53 ads were flagged in August and stronger systems were announced.
- Real photos of a European royal, a 14-year-old influencer and a preteen stock model were converted into graphic abuse videos.
- More than 29,000 EU users and 6,800 UK users were reached, while about 80 percent of the ads ran in the US.
- One promptly reported ad climbed from four impressions to 330 because Meta waited days to remove it.
- GIMC, Meetsocial and BlueFocus, three Meta advertising partners in China, carried ads containing identical CSAM.
- Inconsistent classifications, opposite rulings on identical ads and week-long reviews suggest Meta may be undercounting child exploitation.