Meta purges 5.3 million child abuse files in India with new platform safety tools

Meta's reported proactive detection rate, exceeding 98% in India and 97% globally for the first half of 2026, suggests a reliance on automated systems to identify violative content. This indicates that the company's approach is primarily driven by its internal AI tools rather than depending on user reports.

The new safeguards address increasingly sophisticated tactics by online predators, who reportedly use advertisements that appear harmless but direct users to illegal content outside Meta's platforms. In response, the company has upgraded its ad review systems and deployed additional AI tools, including large language model (LLM)-based detection systems, to identify such 'signposting' content.

These LLM-based systems are designed to detect material that, while not overtly illegal, is suspected of directing users to child sexual exploitation content or related harmful activities. Meta also states it has improved its ability to assess advertisement destinations, allowing it to block links to violative content and take action against the accounts responsible, preventing further dissemination of such links across its platforms.