Meta adds new AI tools to catch ads that link to child abuse material

In short: Meta says it is rolling out new AI tools to find ads that look harmless but quietly send people to illegal child sexual abuse material elsewhere online.

What happened

Meta said it took action against 33.2 million pieces of child sexual exploitation content on Facebook and Instagram in the first half of 2026. The company said more than 97% of that content was found by its own systems before users reported it.

Meta also shared country level numbers for India, saying it acted on 5.3 million pieces of this content in the same period. Meta said more than 98% was detected before users reported it.

The company said it is now focusing more on ads that work like hidden directions. Meta calls this “signposting,” meaning an ad can look normal, but it points people to illegal or harmful content on another website. To address this, Meta says it is using a new large language model, which is an AI system trained on lots of text, to help spot these patterns, and it is paying attention to where an ad sends people, not just what the ad shows.

Meta also said it has a “red-teaming AI agent” to test its own defenses. Red-teaming is like hiring someone to try to break into your house so you can find weak locks before a real thief does.

Why it matters

For regular users, this is about safety and trust. Even if an ad looks innocent, the link behind it can act like a doorway to illegal content. Meta is under pressure from lawmakers and lawsuits over child safety, and it recently agreed to pay up to $18 billion to settle a child safety lawsuit involving 29 U.S. states.

Source: TechCrunch AI