Meta's AI ad filters miss hundreds of child sexual abuse videos, researchers find
Researchers uncovered more than 250 additional ads with real child sexual abuse material on Meta platforms since August, despite the company's AI detection tools.
A nonprofit study revealed that more than 250 ads with child sexual abuse material have been displayed on Meta's social networks since the start of August, adding to the 53 ads previously reported. Unlike earlier cases, the newer ads featured images of real minors, including a member of a European royal family whose official photograph was transformed into a graphic video. The ads promoted AI "nudification" applications, many operated by Chinese developers, and were shown to tens of thousands of users across the EU, UK, United States, Australia and India.
Meta removed the content after being alerted, but in some instances it took up to a week for the company to act, during which the ads gathered hundreds of impressions. US and state authorities, as well as regulators in the UK, Australia and Michigan, have launched inquiries, and senators have urged Meta to tighten its ad-review processes.
How this was covered
- Right-leaning outlets covered this 2h later
Why it matters
The story shows gaps in Meta's AI safeguards, allowing illegal child abuse content to reach millions of users.
In this story
