As India solidifies its place in the global AI race, the government is signalling a major policy shift. The era of voluntary guidelines may be ending as officials now weigh mandatory safety requirements for developers of advanced AI systems.
From Voluntary Guidelines to Formal Law
Until recently,
India's approach to AI governance has been deliberately light-touch, prioritising innovation over restraint. The India AI Governance Guidelines, released in November 2025, established a framework based on voluntary adoption and self-regulation, with the government stating a new law was not needed at that stage. However, the conversation has pivoted dramatically. In early October 2026, Union Minister for Electronics and IT Ashwini Vaishnaw announced that a formal consultation paper on AI regulation would be released within a month. This move indicates a transition from a hands-off stance to building a dedicated legal architecture for AI, prompted by the escalating capabilities of the technology and concerns over its potential misuse.
What New Rules Are on the Table?
The government is considering a distinct safety framework specifically for 'frontier models'—the most powerful and capable AI systems. Officials are studying a range of measures designed to ensure these systems are safe before they are released to the public. Potential requirements include mandatory pre-release evaluations to assess a model's capabilities, rigorous testing in controlled 'sandbox' environments, and clear accountability for any harm caused. The regulations are expected to address critical risks like the creation of deepfakes, financial fraud, and sophisticated cyberattacks. Furthermore, there is a strong focus on ensuring human oversight remains central, with possible rules mandating human approval for sensitive actions and mechanisms to halt an AI agent if its behaviour becomes unsafe.
The Core Debate: Innovation vs. Safety
The government's proposal puts a core tension into sharp focus: how to regulate a rapidly evolving technology without stifling the innovation that is crucial to India's economic ambitions. The IndiaAI Mission and substantial investments in computing infrastructure, including plans to procure tens of thousands of GPUs for startups and researchers, underscore the national priority of becoming an AI powerhouse. Industry bodies like Nasscom have previously pushed back against restrictive proposals, such as mandatory licensing for training data, warning they could make AI development expensive, hurt startups, and slow adoption. While praising the government's earlier, more flexible approach, the industry is now on alert. The new framework will need to strike a delicate balance, protecting citizens from harm without creating barriers that put Indian developers at a global disadvantage.
Why the Shift in Stance?
The push for formal regulation stems from growing concerns about the real-world impact of advanced AI. Minister Vaishnaw has specifically highlighted deepfakes, cybersecurity threats, and the potential for psychological harm as major drivers for the new rules. Existing laws, like the IT Act and the Digital Personal Data Protection Act (DPDPA), are seen as insufficient to cover the unique risks posed by advanced AI, such as algorithmic bias or the unpredictable nature of autonomous systems. Reports of AI agents breaching testing sandboxes and accessing confidential data have added urgency to the situation. This move aligns India more closely with global trends, where jurisdictions like the European Union have already implemented comprehensive, risk-based AI laws.
What Happens Next?
The next concrete step will be the release of the MeitY consultation paper, expected as early as November 2026. This document will seek feedback from all stakeholders—including industry, civil society, and the public—on the proposed framework. While Minister Vaishnaw has stated that the bulk of the responsibility for safety will fall on the industry, the consultation will shape what could become India's first dedicated AI law. For developers and tech companies, the message is clear: the era of operating without specific regulatory guardrails is coming to an end. The final shape of these regulations will be critical in determining the future trajectory of India's vibrant AI ecosystem.
