India Scraps AI Model Launch Approval Requirement

If you're an Indian founder losing sleep over how to get your AI model past a government gatekeeper before launch, you can relax a bit. India has scrapped its proposal to mandate approval for launching AI models — a decision that removes one of the biggest compliance headaches the industry was bracing for. It's a clear signal that India wants to be the country where AI gets built, not the one where AI waits in a queue.

01.What was scrapped, exactly?

The proposal on the table would have required AI developers to seek approval before deploying their models in India. That idea is now dead. Reports this week confirm the approval requirement has been abandoned as part of a deliberate regulatory shift — the thinking being that pre-launch gates slow down innovation without necessarily making anyone safer. The move aims to foster innovation and streamline how AI actually gets deployed, while still expecting compliance with existing ethical and legal standards.

Honestly, this is the right call for a country at India's stage of the AI journey. A licensing regime would have hit small teams the hardest — a Bengaluru garage startup can't afford a compliance department the way a multinational can.

02.A secretariat, not a regulator

So what's coming instead? Reports indicate India's forthcoming AI governance framework will establish a technical secretariat within MeitY — the Ministry of Electronics and Information Technology — rather than a full-fledged AI regulator with teeth. The secretariat's job description is practical rather than punitive: setting guidelines for data provenance, environmental metrics, and safety and security; coordinating across ministries; and hosting an AI-incident database.

Notice what's missing: enforcement powers. The secretariat won't be able to fine anyone or block a launch. It's closer to a technical coordination body than a watchdog. That's a very different philosophy from, say, the EU's heavier-handed approach — and a deliberate one.

03.December deadline for governance bodies

Alongside the secretariat, the government is moving to stand up two inter-agency bodies by December: the Artificial Intelligence Governance Group (AIGG) and a Technology and Policy Expert Committee (TPEC). The AIGG will focus on coordination and building an India-specific risk assessment system, while the TPEC brings in technical and policy expertise. The timing isn't random — the goal is to have policy consensus in place ahead of the India AI Summit. We covered MeitY's earlier signals that AI rules were coming here, and this is the shape those rules are finally taking: coordination first, enforcement later, maybe never.

04.What MeitY's top official is saying

For a sense of the government's mindset, look at what MeitY Secretary S. Krishnan said at Semicon 2.0 this week. He acknowledged that "the existential threat AI poses is something people recognise" — that's what frontier labs are working on — but he was blunt that India cannot afford to overlook the technology's potential across manufacturing, healthcare and governance. "We need to reap those benefits," he said.

Krishnan also confirmed that the government's position is evolving: the Minister has gone on record saying maybe the time has come to look at AI regulations seriously, and the Ministry is examining how to regulate the spread of AI — with coordination at both national and international levels. Translation: serious about safety, allergic to slowing the ecosystem down.

05.Why India is choosing the light touch

Think about the context. Indian AI startups are seeing a surge in VC investment. The government is pouring money into research through the RDI scheme — the ₹1 lakh crore push we wrote about here — and building sovereign AI capacity. In that environment, a pre-launch approval regime would have been shooting your own runners in the legs before the race. The lighter framework keeps the door open for rapid experimentation while quietly building the plumbing — incident databases, provenance guidelines — for accountability later.

Related reads

• India's new AI technical secretariat in MeitY — the institution now steering this policy.
• What IT Minister Vaishnaw said about AI regulation — the thinking behind the light touch.

Key Takeaways

• India has scrapped the proposal to mandate approval before launching AI models
• A technical secretariat inside MeitY will replace the idea of a full AI regulator — with no enforcement powers
• It will set guidelines on data provenance, environmental metrics, and safety, and maintain an AI-incident database
• The AI Governance Group (AIGG) and Technology and Policy Expert Committee (TPEC) are expected by December
• MeitY Secretary S. Krishnan: existential risk is real, but India must reap AI's productivity benefits

Frequently Asked Questions

What This Means for You

If you're a student with a model idea, a startup founder, or a developer in India, the message from Delhi is refreshingly simple: build first. There's no permission slip to wait for, no queue to stand in. That's rare — most of the AI world right now is debating slowdowns, lawsuits and licensing regimes while India is quietly rolling out the red carpet for builders. What that freedom doesn't change is the value of knowing what you're doing — a model deployed badly can still burn user trust, money and reputations. The GPT-6 Astra — Zero to Hero Guide is the practical next step for exactly this environment: prompting, computer use, AI agents, cost optimization, business ideas and safety, taught from zero.

One more thing — if you are still getting up to speed on GPT-6 Astra itself, our complete GPT-6 Astra walkthrough explains the model, the pricing and the privacy basics in one sitting.