If you've been following our coverage of India's AI regulation shift, here's the next piece of the puzzle. The India AI framework everyone has been waiting for is finally taking shape — and the headline is that MeitY is building a technical secretariat, not a regulator. Reports this weekend say the forthcoming governance framework will establish this body inside the Ministry of Electronics and IT, and its design tells you a lot about how India plans to walk the line between safety and speed.
01.So what is India actually building?
Instead of creating a new AI regulator with the power to fine companies or block products — the route Europe took with its AI Act — India is expected to house a technical secretariat within MeitY. Think of it as a coordination engine rather than a policing unit. Its job will be to set guidelines across areas like data provenance, environmental metrics, and safety and security, and to make sure various ministries aren't pulling in opposite directions when AI touches their turf.
And crucially, it won't carry regulatory enforcement powers. That's not an accident or an unfinished draft — it's the whole design philosophy.
02.What the secretariat will actually do
Three responsibilities stand out from the emerging details. First, guidelines on data provenance and safety — rules of the road for how AI systems handle data, where it came from, and how safely it's used. Second, environmental metrics, which is India quietly acknowledging that the data centre boom behind AI has a power and water bill attached to it. Third — and this is the one I find most interesting — an AI-incident database.
Imagine a national record where AI failures, misuse cases and safety incidents get logged the way cyber incidents get reported to CERT-In. That database, done well, could become the single most useful public resource for anyone building AI in India.
03.Why India is not going the full regulator route
MeitY Secretary S. Krishnan spelled out the thinking at Semicon 2.0 this week. "Existential threat that AI poses is something that people recognise," he said — but he was equally clear that India cannot afford to sleep on the technology's benefits in manufacturing, healthcare and governance. "We need to reap those benefits."
He also confirmed that the government's position has genuinely evolved. A while back, ministers felt it was "too soon" to regulate AI. Now, the ministry is seriously examining how to regulate its spread, and the minister himself has said the time has come to look at AI regulations seriously. India is also participating in international coordination efforts like Pax Silica, so the domestic framework is being built with one eye on global alignment.
04.The thing nobody's talking about: timing
Here's my honest take. The interesting part isn't the structure — it's the calendar. This comes within days of India scrapping its requirement for approval of AI model launches, a move we covered earlier, with key governance bodies expected to be in place around December. Read together, the message to builders is consistent: build first, follow guidelines, and don't expect a licence-raj for AI. If you're running a startup in Bengaluru or a student team in a tier-2 college, the compliance burden being sketched out is light by global standards.
Of course, a body without teeth is only as good as its guidelines. A lot depends on whether the incident database and safety guidelines get real technical depth — and whether ministries actually coordinate instead of building their own fiefdoms.
05.How this compares with the rest of the world
The contrast is sharp. The EU went heavy with a comprehensive AI Act and steep penalties. The US is stuck between an industry slowdown pact, antitrust lawsuits and an administration that calls safety fears a hoax. India is choosing the middle path — institutional scaffolding without hard enforcement. We walked through the broader regulation signals from MeitY earlier this week, and this secretariat plan fits that direction of travel.
For a country with India's ambitions — a ₹1 lakh crore RDI scheme, AI skilling for a crore people, sovereign AI infrastructure — a light-touch framework is the choice that keeps all of those programs pointing the same way.
Key Takeaways
- India's AI framework will create a technical secretariat inside MeitY, not an independent regulator
- Its remit: guidelines on data provenance, environmental metrics and safety, plus cross-ministry coordination
- An AI-incident database is planned — potentially the most useful piece for builders
- No enforcement powers, at least for now; key bodies are expected around December
- The approach balances recognised existential risk with India's push to capture AI productivity gains
Frequently Asked Questions
What is India's AI technical secretariat in MeitY?
It is a coordination body expected to be set up inside the Ministry of Electronics and IT as part of India's upcoming AI governance framework. It will coordinate across ministries, frame guidelines on data provenance, safety and security, and environmental metrics, and host an AI incident database.
Will India's AI secretariat have enforcement powers?
No. The secretariat is deliberately designed without regulatory enforcement powers — unlike Europe's AI Act regulator model. It follows India's decision to drop its earlier plan to mandate approvals for launching AI models, with key bodies expected to be announced around December.
How is India's approach different from the EU AI Act?
Instead of creating a new regulator that can fine companies or block products, India is choosing a light-touch coordination engine housed within MeitY that sets guidelines and ensures ministries are aligned, rather than policing AI products.
Related Reads
- India's AI trust gap: the SAS-IDC report findings — the data behind why a secretariat and incident database became necessary
What This Means for You
If you're a founder or developer in India, this is broadly good news — the government is building infrastructure for AI governance, not gates. But "no regulator" doesn't mean "no rules": data provenance and safety guidelines will eventually shape procurement, government contracts and enterprise deals, so building clean data practices now is free advantage later. And if you're just getting started with AI agents and computer use, the practical skills — prompting, cost control, safety basics — matter more than policy headlines. Our GPT-6 Astra — Zero to Hero Guide covers exactly that ground, including a full safety chapter on running agents responsibly, so the day guidelines do arrive, you're already compliant rather than catching up.
By the way, if terms like agents and computer use still feel fuzzy, our GPT-6 Astra guide for beginners explains them all without the jargon.