Meta introduced new artificial intelligence tools on October 7 to detect ads covertly directing users to child sexual abuse material (CSAM) in India. The company reported it actioned 5.3 million pieces of child sexual exploitation content on Facebook and Instagram in India between January and June 2026, detecting over 98% before user reports, according to medianama.com.
The new AI measures include large language model detection to identify seemingly benign ads that covertly link to illegal content, destination checks to analyze where ads lead beyond their visible content, and AI sweeps to find material missed by earlier systems. Meta also deployed a red-teaming AI agent to simulate adversarial attacks and improve detection capabilities, as detailed in the company’s blog post cited by medianama.com.
This update follows a BBC Eye investigation that exposed Instagram running paid ads promoting CSAM in India. Meta’s actioned content in India accounted for about 16% of the global total of 33.2 million pieces during the same period. The new AI tools aim to close loopholes exploited by ads that appear harmless but direct users to illegal content off-platform, addressing a critical gap in online child protection efforts.
Meta’s enhanced AI ad review system reflects ongoing efforts to combat child exploitation online. The company’s data shows that proactive detection prevented the spread of millions of harmful content pieces in India during the first half of 2026, underscoring the scale of the challenge and the role of AI in content moderation.