Amitabh Kant on India's AI Moment: Ambition, Accountability & What We Owe the World

To kick off Season 1 of AI-Naama, Wadhwani AI CEO Shekar Sivasubramanian sits down with Amitabh Kant — India's former G20 Sherpa and former CEO of NITI Aayog — for a conversation on what it really takes to build AI that works at scale, for a country as complex as India.

From policy and governance to public infrastructure and the road ahead, this episode sets the tone for the season: real conversations with the people shaping India's AI future.

🎙️ A Wadhwani AI production.

Follow Wadhwani AI:
 Website: wadhwaniai.org
 LinkedIn: 
 Instagram: instagram.com/wadhwaniai

#AINaama #WadhwaniAI #AmitabhKant

🎧 AI-Naama is a podcast by Wadhwani AI, exploring what is shaping our digital present and future.
🌐 Learn more about Wadhwani AI: https://www.wadhwaniai.org/

Transcript

Copy Transcript Copied

SPEAKER_01 0:00

What was physically not possible earlier? It'll be possible to do this using the power of AI. To my mind, AI will not be fit for purpose if it creates inequalities in society.

SPEAKER_00 0:30

Welcome to the inaugural episode of AI Nama, a podcast from Vadwani AI. Here we discuss perspectives of our digital future with a particular focus on AI. We are delighted to invite Mr. Amitab Kant, who needs absolutely no introduction. He was the person who defined the first AI roadmap for India back in 2018. After that, most of us know he was the Sharpa who led the entire expression of India's G20 summit across multiple cities in India. Welcome, Mr. Ramita Abkanth, and we are really honored to have you on our show. So thank you very much for this podcast.

SPEAKER_01 1:16

No, no, I'm truly delighted to be here, and uh such a uh pleasure to see so many young people working on uh use cases for AI in India. And let me uh first compliment the Vadwani AI Center for using AI to transform India because I recall the time when Prime Minister had launched the Vadwani AI Center in Mumbai, and uh since then it has uh done some remarkable work uh to transform learning outcomes, health outcomes, to work on uh tuberculosis in India and uh build up so many use cases for India using AI.

SPEAKER_00 2:00

In no small measure, this is due to the extensive support that we get from the government because um our entire work, as you know, resides on the backbone of the government's vision, which you put together in 2018, and slowly and steadily that's coming to fruition. So it's actually we're able to see uh in the complex society of India, as you know, change has to be perseverant, consistent, steady, and almost completely quiet and systematic. That's what really causes deep change in India. So we're very grateful. I just wanted to hear your perspectives, just in terms of the focal areas in India that you believe both now, I know that you have a deep interest in various areas water, climate, aspirational districts, different things that you work on. But what would you say are some of the three or four areas that uh an organization like ours, while we continue to expand our work, should place renewed focus on?

SPEAKER_01 3:02

Uh so uh, you know, uh the prime minister's vision is that uh India should be a Viksid Bharat, a developed Bharat by 2047. That is when India becomes hundred. Uh, Viksid Bharat means that we should move from a 4 trillion to a 30 plus trillion dollar economy. Your GDP has to grow nine times, your per capita income has to grow eight times, your manufacturing has to grow 16 times. Uh very few countries in the world have been able to grow at this scorching pace of growth. Only Japan, Korea, and China in recent times have been able to do it. All other countries of the world got caught in the middle income trap. My view is that this will not be possible without the use of technology and without using AI to transform India. AI is a general purpose technology which will impact every single sector of the economy, and therefore AI has to get embedded in our every single sector of India. It is like electricity, it is like computerization which had taken place earlier. But more important than that, to my mind, India cannot be a Viksid Bharat without improving the quality of life, the physical quality of life, without improving uh the human development index. And my belief has been that without improving learning outcomes, without improving health outcomes, without improving nutritional standards, uh it'll not be possible for us to do it. Now, learning outcomes, health outcomes, nutrition are traditionally said to be uh slow moving indicators. It takes a lot of time to do transformational work in this. But my view is that the use of AI can make these as fast-moving indicators. What was physically not possible earlier, it'll be possible to do this using the power of AI. For instance, you know, on learning outcomes, it was said that it's not possible to monitor teachers, it's not possible to do uh physical tests on a regular basis, but using the power of AI, we can do this on a uh across multiple languages, we can do this on a weekly basis and put this out, uh, the results out in public domain so that teachers, parents, students all know whether they are going up or coming down in their learning outcomes. And that'll create the challenge for them. And I have always believed that data is critical. I've always believed outcomes are important. Data is the key, creating competition is important, and putting out results in public domain are very important. You know, when I was a collector, I worked very hard on education and health. But I didn't know whether my district had gone up or down because the Department of Statistics data came six years later. Now you get data thanks to AI on a real-time basis. It's possible to do it. And, you know, use the power of smartphones to transfer data. And therefore, I really think that Vadwani AI can do really transformational work, which it is doing in several districts. We need to spread this across many states, use the power of Vadwani AI to make an impact across several states of India.

SPEAKER_00 6:55

Thank you. We really actually you caught one of the more nuanced aspects of AI, which people tend not to discuss, uh, which is earlier when you had to evaluate anything, you would use statistical quantity control and do sampling. So if you want to find out how good something is happening, you would look at 3% and 5% because that is all the physical capacity allowed. Today you can do 100% sampling. AI and digital transformation allows you not to bother anymore with statistics and go straight to what is happening in 100% of the cases. Yeah. So when we do, for instance, a vulnerability mapping, Pan India in TB, we map the entire country. We don't map small parts. So you are absolutely correct in terms of the perspective of trying to integrate and use data. The other part, which is equally fascinating, is contextual data. Often we try and segment our thinking process into saying this is health, this is education. But from a human perspective, it is the confluence of health and human and economics and agriculture that forms the basic foundational livelihood of people. So, from that perspective, what you're saying is absolutely on top. It's multi-dimensional. It is indeed multi-dimensional. Another different perspective, I know that this is a very big area of discussion in today's world, which is just the context of energy consumption for creating some of these massive models and the benefits thereof or the problems thereof in society. In other words, you've got a small set of companies that spend an enormous amount of energy, they consume a lot. From our perspective, we think about it differently. If we think about it saying that, look, try to focus on contextualizing the solution rather than building these big, big models. At some point, you know, that should taper off, and you really need to focus on impact, which is really contextual. To your point, multilingual, multicultural, multi-data points, India-based solutions seem to be important. What's your perspective on that?

SPEAKER_01 8:59

So, you know, right now, if you look at uh uh what is happening in the world of AI, uh 90% of the new models are coming from the United States. About 5% are coming from uh China and 3% are coming from Europe. So uh innovative models are coming from the United States, but look at the extent of uh energy being used. Uh for instance, just to give you an example, OpenAI's deal with NVIDIA is uh for uh close to about 11 gigawatt of power, which is like I mean, uh uh all the top uh Blackwell uh models of uh uh uh NVIDIA being used. And uh uh you know the amount of computing power being used because uh the first mover advantage battle is being fought out uh amongst several companies that I'm gonna be a leader uh and therefore I'm gonna use a massive amount of computing power uh between four or five companies, massive amount of computing power because I'm gonna take the lead, I'm gonna go then go and do an IPO, I'm gonna become. But you know, if you look at the history of technology in the world, uh first movers are never the winners. I mean, you can see this with uh internet, Alta Vista started it, it went bust. The battle was won by Google. If you uh Clive Sinclair did the first computers, he went bust. Uh, battle was won by Apple and by Microsoft through Windows and Steve Jobs. So first movers never win the battle. So it's not about computing power. Computing power, today the amount of energy being used is more than the energy being consumed by Japan. It's more than the energy being used by the entire city of New York. So it's become an energy guzzler. What has to be done is to have sophisticated, more sophisticated software with limited amount of energy. Correct. And limited amount of energy, and AI has to converge itself with renewable energy and with nuclear energy. It's important that we use renewable energy to fuel more and more of AI, and data centers must be driven by renewable energy so that it becomes compatible with climate. That's number one. Number two, AI must find solutions for many of these challenges of climate today. Correct. In terms of uh better cooling, better uh distribution of energy, uh, how do we do more sustainable development? AI itself will be able to find many solutions for this. And therefore, I'm a long-term believer that in due course, uh, this technology, the march of this technology is inevitable. It's for the betterment of human beings, but the human beings must ensure that this technology is transformational for the benefit of society uh in terms of uh transforming uh many of the social indicators, uh, which is critical, especially in a country like India. We are a large country of 1.5 billion. Uh the battle being fought is for supremacy of companies. The battle being fought in the AI world is in raising resources, the battle is being fought is for large language models, but the applications of those large language models are really critical. How do we use AI to make ourselves a better society is really the key. And therefore, uh the work being done by Vajwani AI is pathbreaking. And to my mind, uh it will really demonstrate many of the use cases for the global south and for the emerging markets of tomorrow.

SPEAKER_00 13:16

Actually, you again bring about a very important point. The way I conceptualize it is you've got tech and innovation, lab-based work, and you've got market, which is noisy, uncertain, ambiguous. And really, it's the middle player who needs to have the conviction, the commitment to bring large-scale social change responsibly. The real celebration of AI can only come when the respect is accorded to the human being using it. So this is a complex thing, and it goes to the heart of what you just said. Stop focusing on this end of the spectrum, start focusing on the other end of the spectrum. And that, of course, requires a certain in India a certain concept of Seva Bhav, abgo bed ke karna hai, lugong ki balay keli karna hai. So that perspective, how do you see and how do you see it playing out in the future?

SPEAKER_01 14:03

So it's a very powerful force. AI will be a very, very powerful force for making a better society of tomorrow. Uh, it will be a transformational force. Uh we are talking at a point of time when the world is going to see the biggest rise in productivity. The world has never gonna see a more transformative force, and we are fortunate to be living at this point of time when the world will be transformed. Uh, we are seeing conflicts in one part of the world, we are seeing breakdown in supply chain on the other part of the world.

SPEAKER_00 14:41

Absolutely.

SPEAKER_01 14:42

Uh but between conflicts, between inflation, between breakdown of supply chain, we are also seeing the rise of the biggest force which will ensure that the world will become a better world of tomorrow because of the rise of productivity. And therefore, how do we enable this use of this productivity for making a better society is critical, particularly in the context of India. And I'm saying this because India accounts for almost close to 20% of the global data. The key to AI is data. India accounts for 20% of the data. India accounts for 20%, more than 25% of the new AI projects on the GitHub. Uh, and India, uh, to my mind, today contributes more than 33% of the data for uh models like uh Chat GPT, 33% more than the United States of America. So Indian data is fueling much of the innovation across the world. The second thing which is very important is to realize that, you know, I mean, if you look at what does AI require, it requires computing power. Now, computing power is being driven by Nvidia. Uh it requires chips, which is being made in Tewan by TSMC. Correct. It requires lithography machines, which come from the Netherlands. Correct. It requires helium gas, which comes from Qatar. So AI was not monopolized by one country, it requires uh you know several companies to countries to work together. And in all this, data is critical, which India controls. India has a major player. Yes. India also is a major player as far as talent is concerned. Uh, if you look at the latest Stanford AI index, India contributes about 2.8 times more talent than the United States or the world global average. United States is 2.2, India does 2.8 more than the global average. And for the first time, the Stanford AI index says that actually global talent is coming back to India. Correct. It's not that India is exporting talent, but global AI talent is coming back to India and is coming back to India for startups, it is coming back to India for uh global capability centers, 2750 odd global capability centers in India. So you have data, you have talent, and the third critical aspect is that India has built up very strong digital public infrastructure.

SPEAKER_00 17:32

Yes.

SPEAKER_01 17:33

So you have a massive amount of data being generated. GST is fully digital. Uh uh you know, uh in India transport movements are uh fast tags are all digital, everything is digital in India. So India is technologically leapfrogged. No other country has been able to do what India has done through the digital public infrastructure. We built the rails, we built the digital infrastructure layer on top of which we've allowed the private sector to innovate and compete. And therefore, my view is that with the digital public infrastructure, with the talent, with the uh data, it's very important uh that a country like India builds up applications. Uh it builds up applications for Indian users. And it's very important that we are able to uh use it for transformational purposes, not merely for India but for global south. I'm saying this because to my mind, AI will not be fit for purpose if it creates inequalities in society. 100%. AI will not be fit for purpose if it is not multilingual in a country like India. India's uh regulatory uh rules and norms are very progressive. India probably has the best regulation on data, it's very progressive, it's uh it's pro-innovation, it uh pushes the boundaries on innovation. Uh unlike uh the regulatory regime in say EU, India's AI regulations and data regulations are all very, very pro-innovation based. So the regulatory regime, the data, the talent, all this and the digital public infrastructure which India has built up uh will enable India to leapfrog in the years to come as far as use cases are concerned.

SPEAKER_00 19:35

Thank you again. Thank you very much for the perspectives that you bring. I totally agree. It's almost as though the diversity and the richness of the country has now been given a tool to express itself in the world. And I think all the factors, like you said, the data, the talent, the regulatory framework, and the availability of a set of use cases offers ourselves an extraordinary opportunity in history, like you correctly said. And I do have to compliment you as well as the government. See, in 2018, the world was a different place. And to have even conceptualized that something like this can start making a change, it does take time. So it's eight years ago, but we were able to establish a lot of foundational elements. Coincidentally, we were also formed around the same time. So again, I would like to express our thanks to you because of people like you who have the ability to think a little more about the future and what the future offers. And then organizations like us, which are very well um thankfully um aligned and understand what it means to bring about large-scale social change. We are working closely with you. So any concluded remarks from your end? We would be delighted to hear it.

SPEAKER_01 20:46

No, first of all, I would like to compliment the Vadwani brothers. You know, I've Ramesh and Sunil, I met them when uh the Vadwani AI Center was launched. Uh they had the vision at that point. Eight years back, they had the vision to use the power of AI to transform India. I don't think any Indian across the world ever thought of that. No. So they had the vision, they had the commitment, they had the passion, they had the energy, and they had the resources to put into India. No one else has done what they've done. So I really want to compliment them for this vision. Uh, secondly, I want to compliment you and your team, which had the ability to really work closely with government. And working with the government is not easy. It's very complex, it's very difficult. And uh your ability to work across government departments and your ability to work with state governments to find solutions to many of the pressing challenges of India is really remarkable. And when we find solutions to the challenges of India, you're finding a solution not for the 1.5 billion people of India, but you are finding a solution for the next 5 billion people of the world who will move from poverty to middle class in the next two decades. And therefore, what Wadwani AI is doing is to find a solution for the next 5 billion people of the world. You are moving slow-moving indicators to fast moving indicators, and you are transforming lives of citizens of the world.

SPEAKER_00 22:26

Thank you. Again, thank you very much for your kind words, and thank you very much for taking time to spend with us. We indeed will be guided by some of the principles that you've mentioned. And a meeting like this uh validates what we do. It gives us an energy, a renewed energy for the work that we do, and we truly look forward to a lasting relationship in terms of direction and guidance in terms of the work that we do. Again, thank you very much, Mr.

SPEAKER_01 22:54

No. I'm always available to you for any any time you want. Thank you. Thank you. Thank you.

SPEAKER_00 22:58

And that's a wrap of our very first special episode of AI Nama by Vadwani AI. Thank you for watching, and we're really grateful to Mr. Amitab Khan for the wonderful conversation that we had on AI. Keep watching. Until next time, take care. Bye.