AI news · Meta · ads safety
Meta uses AI to catch ads that point to child abuse material
New language-model checks look at where an ad leads, after Meta found harmless-looking ads steering users off-platform.
Meta said on Wednesday 7 October 2026 that it has added new AI tools to its ad review to catch ads that quietly steer people to child sexual abuse material elsewhere online. The company’s post describes a tactic it calls signposting: ad content that looks harmless but points to illegal material off Meta’s platforms.
The measures include a large language model trained to spot signposting, checks on where an ad leads as well as what it shows, extra AI sweeps for material earlier systems missed, and a red-teaming AI agent that probes Meta’s own defences. Meta is also tightening detection of banned users who return with new accounts. Meta says ads that carry a blocked link are rejected at upload.
Meta acted on 33.2 million pieces of child sexual exploitation content on Facebook and Instagram in the first half of 2026, Meta says more than 97% was found before anyone reported it, TechCrunch reports. TechCrunch notes Meta agreed in August to pay up to $18 billion to settle a child safety suit brought by 29 US states.
For advertisers, the practical change is that Meta says its review now looks at where an ad leads as well as what it shows, and acts against the accounts behind violating ads.
Sources
- Meta: Measures We’ve Put in Place to Fight Child Exploitation. Checked
- TechCrunch: Meta rolls out new AI tools to detect ads that secretly lead to child sexual abuse material. Checked
Alexandre S.
Alexandre S. is a creative ad strategist and copywriter with ten years' experience in digital marketing across multiple channels.