Meta has pledged to tighten its defences against child sexual abuse material, deepfakes and other AI-generated harmful content, as the company wrapped up a second round of talks with the Centre on content moderation and platform accountability, people familiar with the discussions told NDTV Profit. The discussions, involving senior Meta executives, Union IT Minister Ashwini Vaishnaw and officials from the Ministry of Electronics and Information Technology, centred on the company's content moderation practices and measures to strengthen user safety. Meta assured the government it would enhance due diligence to detect and remove CSAM, deepfakes and misleading AI-generated content, and outlined improvements to moderation systems addressing risks from generative AI. The government maintained that compliance would be judged by outcomes rather than assurances, with proposed measures closely monitored. A key point of discussion remained the Centre's insistence on a three-hour deadline for taking down specified categories of harmful content after it is flagged or identified, applying to CSAM, deepfakes, misleading AI-generated material and other objectionable content under India's evolving regulatory framework. The engagement is expected to continue, with technical consultations likely to examine how recommendation algorithms, automated moderation systems and AI-driven detection tools operate in practice.