India moves toward a dedicated AI law

India is preparing a broader, risk-based law to govern artificial intelligence, a shift from its earlier position that a standalone AI statute was not yet necessary. An official told the Economic Times on July 6, 2026, that the proposed law would classify AI systems by risk level, with low-risk tools such as chatbots, productivity software and recommendation engines facing minimal rules, while high-risk applications in banking, finance and healthcare would be held to stricter requirements. The draft is also expected to include emergency powers allowing the government to direct companies to shut down AI systems or disclose technical details during a crisis. No draft bill, consultation paper or timeline has been released yet.

The move follows a narrower but binding step taken earlier this year. In February 2026, the Ministry of Electronics and Information Technology notified amendments to the IT Rules bringing synthetically generated content, including deepfakes, formally under the intermediary due-diligence framework. Platforms must now label AI-generated content, detect and flag synthetic media, and take down unlawful material within three hours of notification, with a two-hour window for the most sensitive cases such as non-consensual deepfakes.

Together, the two developments mark a change in tone from November 2025, when MeitY’s India AI Governance Guidelines had argued that existing laws were sufficient and a dedicated AI statute was not immediately required.

CT Bureau