AI for All: What NITI Aayog's National AI Strategy Actually Said, and Where It Stands Now

NITI Aayog published the National Strategy for Artificial Intelligence in June 2018 under the tagline #AIFORALL. It picked five focus sectors, proposed a two-tier research structure, and argued India should become an AI solutions provider for the developing world. Here is what the document actually contains, what got built, what quietly stalled, and what it means for anyone deciding what AI skills to invest in now.

What #AIFORALL actually is

AI for All is the tagline of the National Strategy for Artificial Intelligence, a discussion paper NITI Aayog released in June 2018. It is not a law, not a budget line, and not a scheme you can apply to. It is a 115-page position document that argued two things: India should use AI to solve problems for its own population rather than chase frontier model supremacy, and whatever India builds should be exportable to other developing economies facing similar constraints.

The phrase itself carries a specific meaning in the paper. Inclusive growth first, technology leadership second. NITI Aayog framed India as the "AI Garage" for the developing world, meaning that a solution which works for a district hospital in Bihar has a decent chance of working in Kenya or Vietnam.

The strategy named five focus sectors. It also named the things standing in the way, which is the more useful half of the document and the half almost nobody quotes.

Focus sector	What NITI Aayog wanted from it
Healthcare	Close the doctor and diagnostics gap in rural and underserved areas through screening, imaging and triage tools
Agriculture	Raise farm productivity and income through yield prediction, pest detection, soil health and price intelligence
Education	Improve access and learning outcomes, cut dropouts, support teachers rather than replace them
Smart cities and infrastructure	Traffic, crowd management, utility efficiency and urban service delivery
Smart mobility and transportation	Autonomous systems, fleet optimisation, predictive maintenance in rail and road

The barriers it named, and why that list aged well

NITI Aayog was unusually blunt about what India did not have. Low intensity of AI research, almost no core research at all. A shortage of people who could actually build, not just use. No clean, accessible, well-governed data at scale. High resource cost and low awareness among the businesses that needed AI most. And an unclear position on privacy, security and ethics.

Read that list again in 2025. Four of the five are still roughly true. Data access improved at the edges through India Stack and sector-specific repositories, and awareness has gone from low to arguably overheated. The talent gap did not close. It changed shape. In 2018 the shortage was people who could train models. Today the shortage is people who can take a model that already exists and wire it into a real process with real constraints and real accountability.

The paper also proposed research infrastructure: COREs (Centres of Research Excellence) for fundamental work and ICTAIs (International Centres for Transformational AI) for applied work, with an apex body to coordinate them. This is the part that moved slowest.

What happened after 2018

The 2018 paper was a discussion paper, and NITI Aayog kept writing. In 2021 it put out two Responsible AI documents, one on principles and one on operationalising them, covering safety, inclusivity, equality, privacy, transparency, accountability and the protection of positive human values. Those are the documents people actually cite when they talk about Indian AI ethics, and they are worth reading alongside the original.

The centre of gravity then shifted. MeitY took the operational lead. The IndiaAI Mission, approved by the Cabinet in March 2024 with an outlay of around Rs 10,372 crore over five years, is the thing that now has money attached. It funds GPU compute, a datasets platform, an application development initiative, foundation model development and a FutureSkills pillar. If the 2018 strategy was the argument, IndiaAI is the attempt at execution.

The ICTAI network never materialised the way the paper described it. The COREs did not get built as a named national system. What did happen is messier and in some ways more real: a genuine startup layer solving narrow Indian problems, Bhashini pushing Indian-language speech and translation, and a handful of serious academic groups funded through other routes. Whether that counts as the strategy working or the strategy being overtaken is a fair argument to have.

Year	Milestone
June 2018	NITI Aayog publishes National Strategy for Artificial Intelligence (#AIFORALL)
2020	RAISE summit; policy conversation shifts towards responsible and social-impact AI
February 2021	NITI Aayog releases Responsible AI principles (Approach Document, Part 1)
August 2021	Part 2 on operationalising responsible AI principles
March 2024	Cabinet approves the IndiaAI Mission under MeitY, with compute, datasets and skilling pillars
2025	Focus moves to Indian-language models, sovereign compute capacity and sector deployment

Who this actually affects, and how

If you are a student or researcher, the practical signal from #AIFORALL and IndiaAI is that India is funding applied AI far more than frontier research. Compute allocation, datasets, deployment in health and agriculture, Indian-language systems. That is where fellowships, grants and jobs cluster. Choosing a thesis on a marginal improvement to a transformer architecture is a harder road here than choosing one on, say, Kannada speech recognition for an ASHA worker app.

If you work in policy, government or a PSU, the five focus sectors still shape how proposals get evaluated and funded. Framing a project as an inclusion outcome rather than an efficiency outcome is not spin, it is alignment with the stated national position.

If you run a startup or an MSME, honestly, the strategy document will not change your Monday. What it tells you is directional: the public money is going into compute and datasets, not into subsidising your SaaS bill. The benefit reaches you indirectly and slowly.

And if you are a working professional wondering whether to care at all, here is the blunt version. Policy documents do not create your job security. The gap NITI Aayog identified in 2018, people who can convert AI capability into a working process, is the gap employers are still trying to fill. Reading the strategy tells you where the demand is pointed. Building two or three real things with AI tells an employer you can meet it. If you want that second part structured rather than self-taught, the AI Fellowship runs as an eight-week hands-on cohort built around projects and automation work instead of theory, which is roughly the shape of the skill the strategy keeps asking for.

How to read the document without wasting a weekend

The full PDF is on the NITI Aayog site, free. Most people do not need all of it.

If you are preparing for a civil services or policy exam, read the executive summary, the five focus sectors with one example each, the barriers list, the CORE and ICTAI structure, and the AI Garage framing. That is almost everything that gets asked. Add the 2021 Responsible AI principles and you are covered for the ethics question too.

If you are reading it for work, skip to the sector chapter that matches your field and read the barriers section. Then go straight to the current IndiaAI Mission pillars, because that is where the funding and the tenders actually are in 2025.

One last thing worth saying plainly. The 2018 strategy is a good document that was never given statutory force, and it is more useful as a map of India's intentions than as a description of what exists. Treat it that way and it will serve you well. Treat it as a status report and you will be confused about why the ICTAIs you read about are not on any map.

FAQs

1. Is the National Strategy for Artificial Intelligence legally binding?

No. It is a discussion paper published by NITI Aayog in June 2018, not legislation or a notified policy. It shapes how government bodies frame AI priorities, but it does not create obligations or rights for anyone.

2. What is the difference between #AIFORALL and the IndiaAI Mission?

#AIFORALL is the 2018 NITI Aayog strategy that set the vision and named five focus sectors. The IndiaAI Mission, approved by the Cabinet in March 2024 under MeitY with an outlay of about Rs 10,372 crore, is the funded programme that builds compute, datasets, skilling and model development. One is the argument, the other is the budget.

3. What does NITI Aayog mean by India as the AI Garage for the developing world?

It means building AI solutions under Indian constraints, low cost, patchy connectivity, many languages, scarce specialists, and then exporting those same solutions to other developing countries with similar conditions. The claim is that a product proven in rural India transfers better to Africa or Southeast Asia than one built for Silicon Valley.

4. Which ministry is responsible for AI policy in India now?

The Ministry of Electronics and Information Technology (MeitY) leads operational AI policy and runs the IndiaAI Mission. NITI Aayog continues to contribute on strategy and responsible AI frameworks, but it is a think tank, not an implementing ministry.

5. Are the COREs and ICTAIs proposed by NITI Aayog operational?

Not as the two-tier national network the 2018 paper described. Applied AI research in India has instead grown through university centres, startups and MeitY-funded initiatives, without the apex coordinating body the strategy recommended.

6. What are NITI Aayog's Responsible AI principles?

Released in 2021 across two approach documents, they cover safety and reliability, equality, inclusivity and non-discrimination, privacy and security, transparency, accountability, and protection and reinforcement of positive human values. They are the most commonly cited Indian framework for AI ethics.

Related reading

• National Strategy for Artificial Intelligence (#AIFORALL): What It Says and What It Became
• What Are the Principles of Responsible AI?
• National Strategy for Artificial Intelligence #AIFORALL: What It Says and What Came of It
• Is Data Science Required for Artificial Intelligence? A Straight Answer
• What Is Generative AI? A Plain Explanation of How It Actually Works