Meta acts on 5.3 million child sexual abuse content items in India in six months
Over 98 per cent of the content on Facebook and Instagram between January and June 2026 was detected before any user complaint, the company said.
Images are not validated/verified
Meta has tightened the safety systems on its platforms to curb online sexual exploitation of children. Between January and June 2026, action was taken in India on 5.3 million (53 lakh) pieces of digital content on Facebook and Instagram linked to child sexual abuse. More than 98 per cent of this content was detected before anyone complained.
Globally, action was taken on 33.2 million (3.32 crore) such pieces of content on Facebook and Instagram in the same six months, and over 97 per cent was caught before a complaint was made. In India the figure was 5.3 million, with the share caught before complaints above 98 per cent.
Two terms recur in the figures. "Action on content" means removing, blocking or otherwise restricting posts, photos, videos or other digital material linked to child sexual abuse. "Detected before complaint" means Meta's automated safety systems identified the material before a user filed a complaint.
Those who exploit children keep adopting new methods to evade investigators, and Meta has added more AI tools to its systems for this reason. The system is being used to catch cases in which seemingly harmless advertisements lead users to illegal content outside the platform. According to Meta, when such advertisements are found, action is taken against the account concerned and the external links in the advertisement are also blocked.
A new detection system that runs on a large language model has been built to catch advertisements indicating child sexual abuse. Along with what an advertisement shows, it checks where the advertisement takes the user. An additional AI check is also running on the basis of information from this process, to catch content missed by the older system. "Red-teaming AI agents" are being used to find gaps in the safety arrangements. Measures to catch repeat violators, who return with new accounts after theirs are removed, have also been strengthened.
Meta relies on several AI and automated tools to find suspicious accounts. These include "behavioural signals", which analyse account metadata and signs related to an account's behaviour. PhotoDNA and photo and video matching technology are also used to find copies of previously identified child sexual abuse content. The company says it has used this technology across its apps since 2011.
Meta also blocks links to external websites that host or produce child sexual abuse content. After a link is blocked, posts, advertisements and comments carrying the same link are also found and removed. Steps have also been taken to stop content with such blocked links from being posted at all on Facebook, Instagram and Threads. Advertisements with such links are rejected at the time of upload.
Meta is working with law enforcement agencies so that offenders in child sexual abuse cases face legal action. The company said in September that it would pass information on child safety cases to the National Cyber Crime Reporting Portal through the Indian Cybercrime Coordination Centre (I4C). The company says this will strengthen the process of getting such information to the agencies concerned.
