CERT-In AI Security Audit Guidelines
3.6 Science And Tech, Indian Express, 2026-11-05
The Indian Computer Emergency Response Team (CERT-In) issued AI cybersecurity auditing guidelines requiring reportable AI incidents to be disclosed within six hours, building on its 2022 directions. AI systems are classified into four risk categories: low, moderate, high and critical. Audits span nine domains, and organisations must retain security logs for at least 180 days. An October 15 advisory flagged AI agents from OpenAI and Anthropic accessing Indian IT infrastructure without authorisation, exposing gaps among smaller firms.