India Considers Frontier AI Safety Rules as Government Prepares Consultation Paper
The Indian government is preparing to issue a consultation paper on a comprehensive artificial intelligence law in the coming weeks, according to Union Electronics and IT Minister Ashwini Vaishnaw. Officials familiar with internal discussions told NDTV that policymakers are evaluating how to address risks arising from increasingly capable AI systems.
A consultation paper has been announced, but the government has not issued final legislative text or fixed obligations for frontier-model developers. Rather than confining statutory oversight to platform distribution, the upcoming law could go beyond regulating distributed content to examine the systems that generate it.
Proposed Oversight Tiers and Possible Pre-Release Testing
Under the architecture under consideration, regulatory obligations would distinguish between general algorithmic applications and highly capable frontier systems. Establishing separate tiers enables authorities to concentrate supervisory oversight upon premier models while shielding emerging startups from uniform overhead. As reported by Crypto Briefing, developers constructing frontier models could face supervised sandbox requirements and mandatory controlled-environment evaluations to identify hazardous capabilities prior to deployment.
A central technical challenge for policymakers will be establishing objective statutory thresholds to classify frontier systems. Calculating the volume of hardware compute deployed during training provides one quantitative metric, though qualitative functional evaluations remain under consideration. Officials noted that the government could examine whether developers should limit access to particular functions or restrict certain users.
StartupWire reports a possible public consultation process, but an official launch date, portal procedure and comment period have not yet been confirmed.
Controlled-environment evaluations are among the possible approaches described in reporting, not confirmed statutory requirements. This approach would alter India's historical reliance on incremental rules by applying specific pre-deployment standards.
Evolving Policy Approach and Safety Institute Expansion
The proposed consultation could signal a shift in India’s approach to regulating high-capability AI systems, depending on its eventual scope. In November 2025, the government issued the India AI Governance Guidelines, which prioritized an innovation-friendly stance and encouraged voluntary safety practices across the technology sector.
Officials have discussed possible legal obligations for high-capability AI, but specific requirements and any enforcement structure remain undecided pending consultation. According to reporting, international developers including Anthropic have held discussions with the Ministry of Electronics and Information Technology regarding technical safety testing protocols.
Legal Accountability and Enforcement Challenges
Beyond high-capability frontier systems, the consultation paper addresses widespread societal risks including deepfakes, financial fraud and automated impersonation. Administrators emphasized that content disclaimers provide only partial defense, requiring regulators to weigh dissemination circumstances alongside potential real-world injury.
Legal commentators have argued for clearer rules addressing risks from advanced AI systems.