Meta has acknowledged limitations in its current automated enforcement systems, saying they "often rely on straightforward patterns" and "aren't as good at capturing nuance," following an NDTV investigation that exposed sexually suggestive videos involving minors, AI-generated exploitative content, coded hashtags and captions, and networks directing users to Telegram channels. The tech giant disclosed it is developing a new generation of AI models to strengthen enforcement against child sexual exploitation on Facebook and Instagram. Meta said its current systems largely rely on probability scores and established patterns, while the new models analyze the full context of a post, including intent, tone, syntax and language variations. The company said it is already testing the upgraded models, finding they deliver "more accurate and comprehensive enforcement than our current systems, reducing mistakes." The disclosures come amid mounting scrutiny in India, where Meta has been summoned by the Ministry of Electronics and Information Technology over child sexual abuse material and the temporary removal of a Facebook video posted by Prime Minister Narendra Modi. Meta's chief global affairs officer, Joel Kaplan, apologised over the temporary restriction on Wednesday.