Meta deploys AI to catch ads leading to child abuse material

meta introduced new ai tools to detect ads and accounts directing users to illegal child abuse material on facebook and instagram. in the first half of 2026, meta acted on 33.2 million pieces of child sexual exploitation content, with 97% detected by its systems before user reports. the company uses a new large language model to detect 'signposting' ads that may look normal but direct users to illegal content. meta also employs ai-driven scans to find child exploitation content and a 'red-teaming ai agent' to test its safety measures. these efforts aim to prevent bad actors from using ads to send users to illegal content outside meta's platforms.