Meta Ads Contained AI-Generated CSAM; Researchers Found 50+ in Ad Library

Meta has been running paid ads containing AI-generated child sexual abuse material (CSAM) and images of minors with sexually suggestive text, according to a report from the Tech Transparency Project (TTP) shared with WIRED. Over 50 image and video ads were discovered in Meta's ad library, some reaching thousands of accounts across the US, UK, and Europe.

Key Findings

• Ads ran from November 2024 to August 2025, targeted men in multiple countries.
• One video ad used a thumbnail of a child with text: “Realizing Deep Fantasies with Generation AI. There is so much more than what is shown, use your imagination.” Clicking it showed adults in sexual acts then added the child's face.
• Another ad showed a young girl with legs spread and text “I can show you more.”
• Most ads had minimal reach, but one reached 2,563 accounts in Europe.
• Ads linked to “nudify” or undressing apps.
• TTP reported findings to NCMEC CyberTipline; NCMEC doesn't comment on reports.

Meta removed the ads after WIRED reached out. A spokesperson said they work aggressively to keep such content off the platform, noting many ads had minimal reach and some predate new AI detection tools. Meta claims to have removed over 36 million pieces of CSE content last year and is taking legal action against nudify app developers.

However, researchers found 30 more ads hours before publication, some live, after Meta had been alerted. This indicates the AI detection systems are not fully effective.

Implications for Developers

This incident highlights the ongoing challenge of moderating AI-generated content at scale. For developers building AI content moderation systems, the key takeaway is the need for robust detection of synthetic CSAM, which may differ from traditional imagery. The fact that these ads passed through Meta's review process suggests gaps in automated and human moderation.