Meta Rolls Out New AI Tools to Combat Child Sexual Abuse Material

Meta has launched new AI tools to detect ads and accounts that secretly lead users to child sexual abuse material (CSAM) and other illegal content online. This initiative follows the discovery of ads on its platforms that, despite appearing normal, were designed to direct users to harmful content elsewhere, as reported by TechCrunch. The company confirmed on Wednesday that it acted against 33.2 million pieces of child sexual exploitation content on Facebook and Instagram during the first half of 2026, with over 97% detected by its systems before user reports.

In India specifically, Meta addressed 5.3 million pieces of child sexual exploitation content within the same period, with over 98% identified proactively by its systems. The latest AI tools are particularly aimed at identifying "signposting," a tactic where bad actors use seemingly harmless ads to funnel users towards illegal content or other harmful activities off Meta's platforms. This new approach means Meta is now scrutinizing where an ad sends users, rather than solely focusing on the ad's content itself, to block offending websites and take action against associated accounts.

To enhance detection, Meta has introduced a new large language model (LLM) system designed to uncover these sophisticated methods used by perpetrators. The company is also implementing additional AI-driven scans to catch CSAM that older systems might have missed, continuously adding new signals as it learns about evolving network operations. Furthermore, a "red-teaming AI agent" has been deployed to test Meta's internal safety measures, proactively identifying weaknesses that bad actors could exploit.

Meta is also improving its systems to track and remove individuals who attempt to return to its platforms with new accounts after previous ones have been disabled for violations. These new measures are introduced amid ongoing pressure on Meta regarding child safety, including lawsuits and criticism from lawmakers. In August, Meta agreed to pay up to $18 billion to settle a child safety lawsuit involving 29 U.S. states.

Earlier this year, Meta introduced several child-safety features, such as parental controls for Meta AI and pre-teen accounts on WhatsApp. In September, WhatsApp further enhanced parental controls, allowing parents to limit how teenagers use Channels, manage status update visibility, control group additions, and receive notifications about specific group activity.

What to watch: The effectiveness of Meta's new AI tools in reducing the spread of child sexual abuse material.

Editor's note: The draft is comprehensive, well-structured, and incorporates all statistical and contextual data from the provided full text.