Meta said Wednesday that it took action against 33.2 million pieces of child sexual exploitation content on Facebook and Instagram during the first half of 2026. The company says more than 97% of that content was found by its systems before users reported it. In India, Meta acted on 5.3 million pieces in the same period, with more than 98% detected before user reports.

The new AI tools target a tactic the company calls “signposting”: ads that look harmless but direct users to illegal content or other harmful activity outside Meta‘s platforms. Meta has introduced a large language model (LLM) system for spotting these ads, and it now looks at where an ad sends users, not just what the ad contains. It can use that information to block rule-breaking websites or destinations and take action against the accounts behind them. Meta is also running additional AI-driven scans to catch child exploitation content earlier systems may have missed, and says it will keep adding signals as it learns more about how these networks operate.

Other announced measures include a “red-teaming AI agent” that probes Meta‘s own safety systems for weaknesses and an improved process for finding users who return with new accounts after their previous accounts were removed. These changes come as Meta faces continuing pressure, lawsuits, and criticism from lawmakers over child safety. In August, Meta agreed to pay up to $18 billion to settle a child safety lawsuit involving 29 U.S. states.

Meta has also added several child-safety features this year: parental controls for Meta AI, pre-teen accounts on WhatsApp, alerts for parents when children search Instagram for self-harm content, and September WhatsApp controls that let parents limit teenagers’ use of Channels, control who can see status updates, decide who can add their children to groups, and receive notifications about certain group activity.