Researchers find AI child abuse imagery in Meta ads on its platforms

In short: Researchers say Meta ran and approved dozens of paid ads that included AI-generated child sexual abuse imagery and sexual content involving minors.

What happened

Researchers at the Tech Transparency Project (TTP) say they found more than 50 image and video ads in Meta’s public ad library that contained explicit AI-generated child sexual abuse material, or CSAM (illegal sexual images involving children), plus other ads showing minors with sexually suggestive text.

The ads appeared across Facebook, Instagram, Messenger, and Threads, based on Meta’s ad library entries. TTP says the ads ran between November and early August, and some ran as recently as this week. Some ads targeted people in the United States, the United Kingdom, and more than a dozen European countries, and some targeted men.

TTP says some ads linked to “nudify” or “undressing” apps, which are tools that use AI to make fake nude images (like a photo editing filter, but used for sexual abuse). Wired reports that one app, MaskAI, was listed in Apple’s App Store and appeared to contain AI-generated pornography and face-swap features. Apple told Wired it removed MaskAI after being contacted.

After Wired contacted Meta, Meta removed the ads and said they violated its rules. Meta also said its systems review ads mainly using automated tools, and that many of the ads had minimal reach, with one ad shown to at least 2,563 accounts in Europe based on available data.

Why it matters

These were paid ads, not regular user posts. That means someone paid to place them, and Meta’s ad review system approved them. The report raises questions about how well major platforms can stop illegal and abusive AI-made content from being promoted and distributed through advertising.

Source: Wired