Meta rolls out AI ad-review tools to curb child sexual exploitation content
META•Meta introduced AI-driven ad-review measures targeting ads that direct users off-platform. From January through June 2026, it actioned 33.2 million pieces of child sexual exploitation content on Facebook and Instagram, with more than 97% detected proactively.
1. New review measures
Meta’s controls include large language model-based detection, expanded analysis of external destinations, stronger detection of repeat offenders and red-teaming to probe defenses. The company plans to report child safety cases directly to India’s National Cyber Crime Reporting Portal, alongside workflows linked to NCMEC.
2. Content actioned
From January through June 2026, Meta actioned 33.2 million pieces of child sexual exploitation content on Facebook and Instagram, with more than 97% detected proactively. In India, it actioned 5.3 million pieces, with more than 98% detected proactively.
