As India positions itself as a global AI powerhouse, the government is exploring a new frontier of regulation: direct controls on AI models. This move signals a potential shift in policy, sparking a critical debate in the nation's tech corridors.
What Are 'Model-Level Controls'?
The discussion
around 'model-level controls' represents a significant potential change in how artificial intelligence is governed in India. Instead of regulating only the applications built on top of AI, this approach would involve oversight of the foundational models themselves—the core engines trained on vast datasets that power AI tools. This could mean requiring developers to get government permission or meet specific safety and testing benchmarks before their models are released to the public. The move follows a period of evolving government stances; an advisory in early 2026 suggested prior government permission for releasing 'unreliable' AI models but was quickly revised after industry pushback. Now, the debate has resurfaced as part of a broader conversation about a dedicated AI law.
The Government's Rationale: Safety First
The primary motivation behind considering such controls is safety and accountability. Policymakers are increasingly concerned about the potential harms of advanced AI, including the rapid spread of misinformation through deepfakes, algorithmic bias, and threats to cybersecurity. In early October 2026, IT Minister Ashwini Vaishnaw confirmed that the Ministry of Electronics and IT (MeitY) would release a consultation paper focused on AI regulation. The stated priorities include safety, a 'human-first' approach, and ensuring no one is left behind. This push for stronger guardrails comes as a parliamentary committee is also set to examine 'safe AI' and deepfake regulation, reflecting a growing consensus in government that the risks associated with AI require a more direct 'techno-legal' approach.
The Innovation Dilemma for Startups
While the government focuses on risk, India's vibrant startup ecosystem watches with apprehension. A pre-deployment approval or licensing regime could create significant hurdles for smaller companies and independent developers. The costs and time required for extensive vetting could stifle innovation and hand a major advantage to large corporations with deeper pockets. Many Indian startups are already moving toward smaller, more specialized AI models to manage costs and enhance data privacy under existing laws like the Digital Personal Data Protection (DPDP) Act. Proponents of a lighter touch argue that a heavy-handed approach could choke the very innovation that the government's 'IndiaAI Mission' aims to foster. The core of the debate is finding a balance that manages risk without killing the momentum of a sector where 86% of founders plan to increase their AI budgets.
Industry Seeks Clarity and a Graded Approach
The tech industry's response has been mixed. While there's broad agreement on the need for responsible AI, opinions diverge on the 'how'. Some Big Tech executives are advocating for sector-specific rules rather than a single, overarching law that treats all AI the same. Meanwhile, lawyers and policy experts are calling for a clear legal definition of 'AI intermediaries' to distinguish them from social media platforms, which would clarify liability. The November 2025 India AI Governance Guidelines, which are not legally binding, favor a 'lightweight' approach, emphasizing adaptation of existing laws over new legislation. Many in the industry hope that any new framework will adopt this graded approach, where obligations are proportional to the risk an AI system poses, rather than a one-size-fits-all mandate.
What Happens Next?
The future of AI regulation in India remains unwritten. MeitY is expected to release its consultation paper soon, which will kick off a formal process of gathering feedback from the industry, civil society, and the public. This paper will be a crucial document, offering the first concrete glimpse into the government's thinking on everything from model-level controls to deepfake mitigation. The final outcome will likely be shaped by the ensuing debate, balancing the government's 'safety first' imperative with the industry's call for a pro-innovation environment. For now, developers, investors, and entrepreneurs are in a 'wait-and watch', aware that the forthcoming rules will fundamentally shape India's digital future.
