The Indian government is considering a new regulatory framework for frontier AI models that would require developers to pass safety tests before public release. This proposal marks a shift from the previous voluntary approach established in the November 2025 India AI Governance Guidelines toward formal legislation.
The Ministry of Electronics and Information Technology is preparing a consultation paper, expected to be released in the coming weeks, to gather feedback from industry and civil society. Reports suggest the document could arrive by November 2026, with the minister’s announcement occurring between October 23 and 28, 2026.
The proposed measures may involve mandatory pre-release evaluations and sandbox testing to mitigate risks such as cyberattacks and synthetic content generation. The IndiaAI Safety Institute, which is already operational, may see its oversight role expanded as part of these regulatory developments.
Major AI developer Anthropic has reportedly engaged in discussions with the Ministry of Electronics and Information Technology regarding model access and safety testing. The company has advocated for the implementation of mandatory transparency rules for AI developers.
