Meta unveiled a new AI‑driven CSAM detection framework on June 30, 2026 after noting that offenders were embedding covert links in seemingly innocuous ads. The update includes large‑language‑model algorithms that flag ads that merely look normal but funnel users toward illegal material hosted off‑platform.
Meta activated additional AI sweeps that scan all ad content for hidden links, and deployed a red‑team AI agent that actively probes Meta’s own defenses for blind spots. The company also strengthened account‑level monitoring to flag users who create fresh accounts after removal.
India’s data shows Meta actioned 5.3 million CSAM items on Facebook and Instagram from January to June 2026, with 98 % caught before any user report. Meta has agreed to forward every CSAM case to the National Cyber Crime Reporting Portal run by the Indian Cybercrime Coordination Centre.
Meta’s content team has held several consultations with Indian officials on moderation, recommendation algorithms, and intermediary duties, and the company plans to hold quarterly reviews of its detection efficacy. The next step is a scheduled audit by the Indian government in September 2026 to assess Meta’s compliance.