The company said the new measures are designed to identify ads that either feature child sexual abuse material, known as CSAM, or covertly direct users to violating activity outside its platforms.
The move comes as predators increasingly use advertising to evade detection, posting seemingly benign promotions that link to external websites hosting illegal content.
"Signposting" — the tactic Meta is hunting
Meta calls the tactic "signposting": the ad itself may contain nothing illegal, but it acts as a gateway, using innocuous-looking imagery or captions to funnel users toward abuse material hosted off-platform.
To counter it, the company introduced a large-language-model-based detection system trained to recognize when ad content strongly suggests a violating destination.
The company also says it has improved its ability to assess where an advertisement leads, rather than examining only what the ad shows. That shift lets Meta block violating destinations and take action against the accounts behind them.
After identifying offending networks, Meta says it widens investigations beyond the reported ads, mapping related activity across its platforms, disabling accounts, and blocking off-platform links.
Sweeping, red-teaming, and banned-user detection
Alongside the signposting system, Meta has added AI-driven sweeps to surface exploitation content that earlier systems may have missed.
It is also using a red-teaming AI agent that proactively probes its own defenses for weaknesses, and has strengthened its ability to detect banned users who create new accounts to slip back onto the platforms.
"Keeping young people safe on our apps is among the most important work we do, and it never stops," Meta said in its announcement.
The scale behind the announcement
The announcement came with new enforcement figures. Between January and June 2026, Meta said it actioned 5.3 million pieces of child sexual exploitation content on Facebook and Instagram in India, with more than 98 percent identified and addressed proactively before anyone reported it.
Globally, it actioned 33.2 million pieces of such content in the same period, with more than 97 percent detected before user reports.
The update lands amid heightened scrutiny of Meta in India over child safety, where the company has held 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 run by the Indian Cybercrime Coordination Centre, known as I4C.
By Hamza Faisal
