Cert-In Releases AI Security Audit Guidelines with Four Risk Tiers
The Indian Computer Emergency Response Team (Cert-In) has issued AI Security Auditing Guidelines, requiring regular security assessments for AI systems in India. Released on November 4, the 107-page document categorises AI systems into four risk tiers: low, moderate, high, and critical. Compliance is mandatory for all organisations, including foreign companies handling Indian data. Low-risk systems may undergo self-assessments, while moderate, high, and critical systems require independent audits. The guidelines also mandate policies against deepfakes and the maintenance of an 'AI bill of materials'. Critical systems must undergo annual red-team exercises. Cert-In emphasised that compliance standards apply uniformly, with no exemptions for smaller firms or startups.
Published 2 days ago • 05 Nov 2026, 10:54 AM IST