Meta ran more than 350 AI child-abuse ads

Tech Transparency Project researchers found that Meta platforms had carried more than 350 abusive AI video ads since the end of last year, including some that used images of real children, WIRED reports. The researchers said more than 250 additional ads appeared since August, and the ads reached at least 29,000 EU accounts and 7,000 UK accounts in Meta’s ad-library data; Meta says the ads were removed and it received under $5,000 in payments. Teams deploying generative-media moderation should file this as a pre-publication review and response-time failure, not an image-generation edge case; the archive’s Meta agent analysis shows why action-taking systems need explicit trust boundaries.