Meta ran dozens of paid ads containing explicit AI-generated child sexual abuse imagery over the past nine months, according to findings from the Tech Transparency Project shared with Wired. The watchdog says more than 50 image and video ads appeared across Facebook, Instagram, Messenger, and Threads, targeting people in the US, UK, and more than a dozen European countries. One ad reached 2,563 accounts.
Some ads linked to nudify apps that undress photos. In one case, a video ad displayed a child’s face before cutting to sexual content. The ads were published between November and early August, and many sat undetected in Meta’s ad library for months, researchers say. Several had already been removed once for policy violations and then reposted unchanged.
Meta removed the ads after Wired inquired and said most had minimal reach, with many disabled before the story ran. The company pointed to new AI detection systems and said it removed 36 million pieces of child exploitation content last year. Researchers found roughly 30 more offending ads hours before publication, some published after Meta was first asked about the issue.
TTP reported the findings to NCMEC’s CyberTipline. Watchdog director Katie Paul said the ads were reviewed, approved, and allowed to run by Meta while it collected the ad dollars, adding that the ads made no effort to mask what they promoted.
