Report says Meta ran ads showing AI child abuse imagery on its apps

In short: A new report says Meta allowed hundreds of paid ads on Facebook and Instagram that showed child sexual abuse imagery, including AI-made videos using real children’s photos.

What happened

The Tech Transparency Project (TTP) said it found 332 ads this year that contained child sexual abuse material on Meta’s platforms, Facebook and Instagram. Many of the ads promoted so-called “nudify” apps, which use AI to digitally alter photos to make it look like someone is nude (like a fake costume put on a photo, but used in a harmful way).

TTP said some ads used photos of real children taken from the internet, including a press photo of a young member of a European royal family and images copied from an Instagram account of a 14-year-old influencer. TTP said these images were turned into explicit videos. The report also said the ads often targeted preteen girls.

Meta removed the ads flagged by TTP and Wired, but TTP said some removals took days. In one case, TTP reported an ad the day it started running, but it stayed up long enough to go from 4 views to 330 views. TTP also said Meta sometimes removed one ad but left up an identical copy.

Why it matters

These ads are not just offensive, they can be illegal and they can harm real children whose everyday photos get turned into sexual content without consent. The report also raises questions about Meta’s ad review system, since ads are supposed to be checked before they are shown to people.

Source: Arstechnica