Meta is using new artificial intelligence tools to detect child sexual exploitation content more effectively on Facebook and Instagram, including cases where offenders use seemingly harmless advertisements to direct people to illegal content outside its platforms.
The company said it acted on 5.3 million pieces of child sexual exploitation content on Facebook and Instagram in India between January and June 2026. More than 98% of the content was detected and addressed before it was reported by users.
Globally, Meta took action on 33.2 million pieces of such content during the same period, with more than 97% detected proactively.
How are offenders using ads?
Meta said it has found cases where advertisements appear harmless on their own but are designed to direct users to illegal material hosted on external websites. Such advertisements can be difficult for automated systems to identify because the ad itself may not contain clearly violating material.
To address this, Meta has introduced new AI tools that look beyond what an advertisement shows and analyse where it directs users.
One of the new systems uses large language models to detect what Meta calls “signposting” — seemingly ordinary ad content that may be used to covertly direct people towards child sexual exploitation material or other harmful activity.
Meta is also conducting additional AI-based sweeps of advertisements to identify content that earlier systems may have missed. It has deployed an AI “red-teaming” agent that tests its own safety systems to identify weaknesses and discover new tactics used by offenders.
The company has also strengthened systems to detect repeat offenders who try to create new accounts after their earlier accounts have been removed.
What happens when Meta finds such content?
When Meta detects or is alerted to potentially exploitative content, it investigates the material. If it violates its policies, the content is removed and accounts linked to malicious activity can be disabled.
Meta also uses technologies such as PhotoDNA and other image and video matching tools to identify known child sexual abuse material. The company said these systems have been used across its platforms since 2011.
Meta also blocks links to external websites hosting violating material. Once a link is blocked, the company searches for and removes other content containing the same link, including posts, comments and advertisements.
The company said it reports identified child sexual abuse material to law enforcement with the assistance of the National Center for Missing & Exploited Children (NCMEC), in line with applicable law.
In India, Meta said it will also report child safety cases directly to the National Cyber Crime Reporting Portal operated by the Indian Cybercrime Coordination Centre (I4C).
The company said the fight against child exploitation requires continuous changes to its systems as offenders change their tactics and move across platforms.
