Meta Launches New AI Tools Against Child Abuse Content
In the first half of 2026, Meta intervened in millions of harmful contents and deployed artificial intelligence systems to detect hidden redirections.
Meta announced that it took action against millions of pieces of content containing child abuse on Facebook and Instagram in the first half of 2026. The company stated that it has launched new artificial intelligence tools aimed at detecting hidden ads and accounts that direct users to illegal materials.
Millions of Contents Intervened
Meta announced that it took action against 33.2 million pieces of child abuse content on the Facebook and Instagram platforms in the first half of 2026. It was stated that the vast majority of the detected contents were found by the company's systems before user complaints.
Specifically for India, it was stated that 5.3 million contents were intervened in the same period, and a high proportion of these were also caught automatically before user reports.
Targeting Hidden Redirections with Artificial Intelligence
The new artificial intelligence tools developed by the company target individuals who redirect users to illegal materials outside the platform through seemingly harmless advertisements. Measures are being increased against the tactics that malicious actors constantly change to evade detection.
Large Language Model and New Scans
Meta has deployed a new large language model system to identify ads that look normal but contain suspicious redirections. By focusing not only on the ad content but also on the targeted destinations, it aims to block rule-violating websites.
Security Tests and Child Safety Steps
The company is using a red-team artificial intelligence agent to find vulnerabilities in security measures while improving its systems to prevent owners of banned accounts from returning with new ones. These steps are carried out alongside increasing pressures regarding child safety and other auditing features being implemented.