Meta Expands AI Fight Against Child Abuse Content Amid Tightened Scrutiny

Meta is expanding its use of artificial intelligence to detect and disrupt child sexual exploitation on its platforms, including advertisements that feature child sexual abuse material (CSAM) or direct users to violating activity outside its services.

The company said the latest measures were introduced after it identified new tactics being used by offenders to evade its systems. These include large language model-based detection of seemingly benign advertisements that covertly "signpost" users towards illegal material, improved systems to understand where an advertisement leads, AI-driven sweeps of ad content, a red-teaming AI agent to probe weaknesses in its defences, and stronger systems to detect offenders who create new accounts after being removed.

READ | Meta Global Team Summoned By Centre Over Algorithmic Bias, Child Sex Content

"Keeping young people safe on our apps is among the most important work we do, and it never stops," Meta said.

The company said offenders are increasingly moving across platforms and changing tactics to evade detection. In one example, Meta said an advertisement can appear harmless on its own while covertly directing users to illegal material hosted off-platform. It said it widened its investigation beyond individual advertisements, disabled accounts linked to the activity, and blocked the external links involved.

In India, Meta said it actioned 5.3 million pieces of child sexual exploitation content on Facebook and Instagram between January and June 2026. More than 98% was detected and addressed proactively, before anyone reported it. Globally, the company actioned 33.2 million pieces during the same period, with more than 97% detected proactively.

READ | Meta Agrees To Report Child Sexual Abuse Cases Amid India Crackdown

Meta said its latest measures include "new Large Language Model (LLM) detection" to identify covert signposting of child sexual exploitation material, alongside technology that examines "not only what" an advertisement shows but "where an ad leads". It has also introduced additional AI sweeps to identify material missed by earlier systems and a red-teaming AI agent that proactively probes its own defences for weaknesses.

The update comes amid heightened scrutiny of Meta in India over child safety and CSAM. The company has held multiple discussions with government officials on content moderation, recommendation systems, human review, and intermediary obligations. Meta has also agreed to report child-safety cases directly to the National Cyber Crime Reporting Portal operated by the Indian Cybercrime Coordination Centre (I4C).

Meta said its existing systems include behavioural signals, image and video matching, and technology to block links to external websites hosting violating material. It said it is also investigating predatory networks at the network level rather than acting only against individual accounts.