India considers safety framework for frontier AI models

The Indian government is considering a new regulatory framework for artificial intelligence that would impose stricter requirements on high-capability frontier models. According to reports, the proposed approach may involve a two-track system that distinguishes between standard AI applications and powerful frontier models capable of significant harm, such as enabling cyberattacks or generating realistic synthetic content.

Proposed measures under examination include pre-release evaluations to assess model capabilities, controlled-environment testing to identify risks in lab settings, and sandbox requirements where developers trial models under supervision before public deployment.

Union Electronics and IT Minister Ashwini Vaishnaw has announced that a consultation paper regarding the forthcoming AI law will be released in the coming weeks. This document is intended to gather feedback from industry stakeholders, civil society, and the public to shape the final legislation.