India is signalling a major shift in its approach to artificial intelligence, moving from a light-touch stance to potentially requiring developers of the most powerful AI systems to prove they are safe before release.
What Is Being Proposed?
The Indian government is actively
considering a new regulatory path for artificial intelligence that would create specific obligations for developers of 'frontier' AI models. This refers to the most powerful and capable systems, like those developed by OpenAI, Google, and Anthropic. According to officials, the plan involves asking these developers to demonstrate the safety of their models before they are released to the public. This marks a significant change from India's previous 'pro-innovation' stance, which relied more on voluntary guidelines and existing laws. The proposal is part of a forthcoming consultation paper from the Ministry of Electronics and Information Technology (MeitY), which will gather feedback from industry and civil society to shape a new AI law.
Why the Shift in Stance?
The move reflects a growing global consensus that the risks associated with advanced AI have reached a critical point. Union IT Minister Ashwini Vaishnaw stated that the time has come for new regulations, citing concerns about cybersecurity, deepfakes, and other potential societal harms. This shift was reportedly accelerated by recent incidents where autonomous AI agents from international labs breached Indian organisations without being detected by the targets themselves. Furthermore, India's national cybersecurity agency, CERT-In, issued a high-severity warning earlier in 2026 about the risks of frontier models, which can be used to automate cyberattacks and create sophisticated phishing campaigns. The government aims to create a 'techno-legal' framework that balances safety with its ambition to be a leading AI development hub.
What Would 'Safety Demonstrations' Involve?
While the details are still being discussed, the proposed safety measures would be multi-layered. They could include mandatory pre-release evaluations to assess a model's capabilities and potential for harm. Developers might be required to conduct extensive testing in controlled environments, or 'sandboxes', to identify risky behaviours before a public rollout. The framework would likely be risk-based, meaning the level of scrutiny would depend on the power of the AI system in question. This would allow lower-risk applications to continue developing with less friction while focusing regulatory attention on the models with the greatest potential to cause widespread harm. The government is also considering what defines a frontier model, looking at factors like the computing power used for training as well as capability-based assessments.
The Global Context and India's Role
India's potential new rules do not exist in a vacuum. They align with a broader international trend toward AI regulation, seen in the European Union's AI Act and various initiatives in the United States. However, officials have been clear that India wants to forge its own path, avoiding what some see as the innovationstifling approach of Europe while still taking risks seriously. Officials from the IndiaAI Mission have stressed that India, as a leader of the Global South, must be part of the global conversation setting safety standards, rather than having rules dictated by the US and China. This move is also supported by the establishment of the IndiaAI Safety Institute, which is intended to develop the capacity to test and audit these complex models domestically.
The Innovation vs. Regulation Dilemma
The central challenge for policymakers will be striking the right balance between public safety and economic ambition. A framework that is too strict could increase costs and slow down innovation, particularly for homegrown startups hoping to compete on the global stage. The government has indicated that the industry will be expected to shoulder the primary responsibility for ensuring safety. This suggests a model of co regulation, where companies are legally required to manage risks, with the government providing oversight. By making its intentions clear through the upcoming consultation paper, the government is giving the burgeoning Indian AI industry a chance to help shape the rules that will govern its future, ensuring that the guardrails put in place are both effective and practical.
