MeitY’s new AI Governance guidelines set seven sutras and a six-pillar framework

New Delhi: The Ministry of Electronics and Information Technology has released the India AI Governance Guidelines, a comprehensive framework to enable safe, inclusive and trusted AI adoption. 

The framework includes:

• Seven guiding principles (Sutras) for ethical and responsible AI.
  
  
• Recommendations across six pillars of AI governance.
  
  
• An action plan with short-, medium- and long-term timelines.
  
  
• Practical guidance for industry, developers and regulators on transparent and accountable deployment.
  
  

The document anchors policy on seven sutras: Trust is the Foundation; People First; Fairness and Equity; Innovation over Restraint; Accountability; Understandable by Design; and Safety, Resilience & Sustainability.

The guidelines organise recommendations under six pillars—Infrastructure, Capacity Building, Policy & Regulation, Risk Mitigation, Accountability, and Institutions. They call for better access to data and computing, agile regulation, graded liability, and a whole-of-government approach to oversight.

The Ministry said that India’s approach is application-focused and pro-innovation: empower sectoral regulators rather than regulate the underlying technology, and balance growth with safety. The report added that a separate AI law is not needed at this stage, but consistent enforcement of existing laws is essential.

An institutional architecture is proposed in the report. A small, permanent AI Governance Group (AIGG) will coordinate policy; a Technology & Policy Expert Committee (TPEC) will advise it; and the AI Safety Institute (AISI) will test and evaluate systems, develop standards, and represent India in global safety forums.

The time-bound action plan maps near-, medium- and long-term tasks. In the short term, priorities include setting up governance bodies, drafting India-specific risk frameworks and liability regimes, preparing an AI incidents database and grievance mechanisms, and expanding access to data, compute and safety tools. Medium-term steps include standards for content authentication and fairness, a national incidents database, regulatory sandboxes, and DPI-AI integration. Long-term goals include new laws where needed, global standards engagement, and horizon-scanning for emerging risks.

On risk mitigation, the report seeks an India-specific risk taxonomy and encourages voluntary measures backed by incentives. It also promotes techno-legal approaches—such as privacy-preserving tools and content authentication—to make “compliance-by-design” practical at scale. A DEPA-for-AI-Training concept is outlined to enable consented data use during model training, alongside complementary guardrails like audits and explainability.

The guidelines were launched by Principal Scientific Adviser Prof. Ajay Kumar Sood in the presence of MeitY Secretary S. Krishnan; Additional Secretary, CEO IndiaAI and DG NIC Abhishek Singh; MeitY’s Kavita Bhatia; and Prof. B. Ravindran of IIT Madras.

Prof. Sood said the framework’s defining principle is simple: “Do no harm.” He said the ecosystem will use innovation sandboxes and flexible, adaptive risk mitigation to support growth while protecting people.

S. Krishnan said the approach is human-centric and will rely on existing laws wherever possible. “AI must serve people and improve lives while addressing potential harms,” he added.

Abhishek Singh said the final guidelines follow wide public consultation. “India remains focused on making AI accessible, affordable and inclusive, while building a safe, trustworthy and responsible ecosystem that fuels innovation and the AI economy,” he said.

The guidelines were drafted by a committee chaired by Prof. Balaraman Ravindran (IIT Madras) with members from government, academia and industry, including Abhishek Singh, Debjani Ghosh, Kalika Bali, Rahul Matthan, Amlan Mohanty, Sharad Sharma, Kavita Bhatia, Abhishek Aggarwal and Avinash Agarwal, among others. The document is positioned as a reference for policymakers, researchers and businesses to foster national and international cooperation on responsible AI.

The launch marks a milestone ahead of the India AI Impact Summit 2026, to be held on February 19–20 in New Delhi, which will gather global leaders to discuss AI’s role in driving people, planet and progress.