India's AI challenge and digital transformation: Full Q&A with Nandan Nilekani

Infosys co-founder Nandan Nilekani in an interaction with Moneycontrol Managing Editor Nalin Mehta at the AIMA event in New Delhi on February 21.

Anchor: And before I request Mr. Nalin Mehta to conduct the session, a few words about him. Mr. Mehta is Managing Editor, Moneycontrol and Non-Resident Senior Fellow, Institute of South Asian Studies, National University of Singapore. He is an award-winning Indian journalist, political scientist and an author who has held senior leadership positions in major media companies and educational institutions, served as an international civil servant with the UN and the Global Fund in Geneva, Switzerland, and has held research positions at universities and institutions in Australia, Singapore, Switzerland and India. It's over to you, sir.

Nalin Mehta: Thanks very much. Nandan, great to be with you here. There are very few people who I think who have profoundly transformed this country and transformed it in so many different ways, as you have. And just before we start, I mean, I don't think you need an introduction for anywhere. And when I say different ways, and this is directly relevant to the conversation, first, as the co-founder of Infosys, so you're one of the progenitors of the Indian IT revolution, one of the big job creators. As the chief of UIDAI, the creator of Aadhaar, pretty much in the underlying plumbing of our society, of our economics, I think some have called you the CTO of India.

You look at UPI, Rs 23 trillion transacted last month, none of that possible without Aadhaar. But most recently, you've also been involved, I think, I was reading your papers recently on the Finternet with the Bank of International Settlements with Augustin, looking at how the vision for global financial systems, your work with the International Energy Agency, which you were involved with on digital energy grid. You've also written books, Rebooting India, Reimagining India, and I think your last book, the subtitle of that is my favorite, which is about remaining calm in a digital world.

So now, coming to AI, there's not too much calmness in the AI world right now. We've just had DeepSeek. We've had the response from OpenAI. You've had responses, political, strategic responses from Washington, from New Delhi, from Beijing. There's a lot of stuff happening in the AI world. So how important, in your view, is AI for India's strategic sovereignty and for autonomy?

Nandan Nilekani: No, I think AI is extremely important. It is strategic, and being a leader in AI is, I think, very important for India. And I think, what I see is also how it can be applied at population scale to solve basic things, how to use AI to make education better so more kids learn, or how to use AI so farmers have access to the right information at the right time in the right language so they can be more productive. So I think, to me, all this AI is as good as its application and its usage. And then I think India will definitely be a leader in applying AI at scale.

Nalin Mehta: So one of the things that the last time when we had spoken on a stage like this was just after the government change in 2014.

Nandan Nilekani: It's 11 years back.

Nalin Mehta: Yeah. And you had, to an audience like this, explained what Aadhaar, the use case for Aadhaar is going to be, India's stack, which created really the foundational things which led to our economy now. I think about $100 billion worth of market value has been generated by companies who have been built using the platform of Aadhaar and so on. So when you look at AI, and we've done the phase one of digital public infrastructure, now that AI is coming in, how do you expand the use cases on scale for our usage as a developing country?

Nandan Nilekani: Well, I think DPI and AI are complementary. DPI has laid the foundation in terms of large systems, a lot of data, all of which are required for the AI. And so AI will be built on top of DPI, and DPI in turn will get turbocharged by AI. So it's a very synergistic relationship.

And I think what DeepSeek has done, I think which is very important, is they have shown that you don't need billions of dollars to create a reasonably good large language model. So I think that's the big breakthrough, and they've also published how they did it. So I think that's very important. They've been very open about it. So I think now the barrier to entry of creating these AI models has significantly come down. And I know that the government of India is very focused on that India AI mission. And I think you will see in the next one year some very significant models coming from India.

Nalin Mehta: So you had said recently that India shouldn't really invest significantly in foundational models or LLMs. Let the big boys in Silicon Valley do that. Let the guys with the chips do that. We should focus on AI for all, for applications on top of that. Have you changed your mind?

Nandan Nilekani: No, I think it's really at what price point. Would I spend a billion dollars building a large language model? No. But if I can deliver a large language model in 50 million, sure. Why not? So it's also about the fact that the technology is moving so quickly that it's dropping in price. It's becoming more efficient. So there is a time at which, and frankly, LLMs will be commodities. It's not really, you can already see that. Every day there's a new announcement of somebody or the other. So it's going to be a race of all the people. And I think now we can actually build models.

But we should not take our eye off the ball, which is how is this useful to society? What is it going to do that makes a difference? So I think if we can do an LLM for 50 million, I would recommend that we do it. But then we should also look at how to use it.

Nalin Mehta: So with DeepSeek, and you talked about this, it's been talked about like a Sputnik moment, right? When the Russians did that, it led to America taking the moon shot. And it had huge profound impact, not just in space, but Silicon Valley and all kinds of aspects of the economy. Now, at the moment, China seems to be investing significantly in core tech, from various things, from AI to robotics, other things. Can India really afford not to go big time into this?

Nandan Nilekani: No, no, I think we should be. All I'm saying is that we should do it at the right time, at the right cost point. And I think now the building of core models is now at a quality and cost point, which anybody can do it. $50 million, so many companies can spend that. So I think now is the time to do that. But again, the challenge with AI is, how do you make it work? How do you make it work at population scale? How do you make sure that it's very cheap to operate? Because they say inference and reasoning and all kinds of things, but they're very expensive from a compute perspective. And we can't afford those costs.

So if we want a farmer in a village to speak to his phone, in Hindi or Bhojpuri or Maithili or whatever, and get a very reasoned answer on what he needs to do so that he can improve his productivity, and I want to do that at a few rupee per transaction, that's a whole different ballgame. And that's where we have to go.

Nalin Mehta: So one of the things that DPI did was that it completely transformed our rural economy in multiple ways. For example, with direct benefit transfers, what it did to the political economy, to politics. But it also created financial inclusion. It led to 600 million bank accounts. Which other specific use cases you have in mind, let's say for agriculture, education, and so on, where once you add AI to it, for example, with PM Kisan, when you add AI to it, how will that change?

Nandan Nilekani: It's already using AI, by the way. PM Kisan already uses AI for questions. I think there are many uses of it, but I'll take three, which I think have the highest multiplier effect for India. First is language. We are a country with 22 official languages, and a few hundred languages and dialects. And that, in some sense, when we talk to each other, we have to speak in English or in Hindi, because that's pretty much the common language that we all have.

But if I'm a farmer sitting in Odisha, and I speak only Oriya, or I'm in Bengal, I speak only Bengali, how do I get access to knowledge? How do I get access to the internet? And I think we can use AI and all the new things like LLMs to make it possible to have voice communication in a language of your choice with the computer. So that dramatically expands access. You don't have to be literate. You don't have to read. You can just talk. And, you know, it can be so subtle that, you know, in India, every 50 kilometers, you know, the dialect version change. You can handle all that.  So the very fact that we can build AI at scale for language, and remove language as an impediment for access itself is huge. So a billion people can now communicate with a computer.

And one of the things that I've supported for the last six years is something called AI for Bharat at IIT Madras, which is a philanthropic thing. And they have built the largest database of Indian language in the world. And that is now being used by everyone else to train their models. So that's like a core, it's like a DPI. It's a core infrastructure of open data on Indian language anybody can use. So language is one clear, huge, massive benefit.

The second is in education. We know that our children are still not fully learning basic literacy and numeracy. Of course, the recent ASER report, you know, after 20 years, is starting to show some improvement in reading and numeracy and all that. But nowhere in the scale we want for our future. And we think that AI can play a huge role here. Because with AI, you can do very micro-diagnostic of where exactly the child is not, like an ECG of their learning, and say, okay, this child doesn't know this vowel or this syntax or this whatever. And we have built, one of my NGOs, X-Step has built that, and it's being rolled out in three states.

So imagine that scale, if you can improve the literacy and numeracy of children. That's huge. Society will benefit hugely with more literate and numerate people, right? And the third thing is, we think, in agriculture, so we're creating this thing called Open Agri Network, which brings all agri knowledge at your fingertips, with reasoning and inference capability, and in a network way, so it can come from different locations. And the government of India has a project called Vistar, which, based on the same infrastructure, many of the states now want to do it. So I think if we can have every Indian communicate with each other effortlessly through language, if every child can learn better with AI, and if a farmer can improve his earnings, good enough.

Nalin Mehta: I think you've pointed out recently that by the end of this decade, more Indians will be using AI than anybody else in the world. Now, and I think one of the arguments you've made is that it's not a solution issue, it's more about distribution. How do we distribute the solutions to a much larger scale? So how does private sector and government work on that?

Nandan Nilekani: No, I think the early population scale applications will probably come from government. I mean, PM Kisan already uses AI.

Nalin Mehta: Some 3 lakh farmers were talking using AI when the data was...

Nandan Nilekani: And you can see the spike. PM Kisan is a cash transfer, right? So whenever there's a cash transfer, there's a spike in usage, because the farmer wants to know what's happening to his money, and he's conversing with the AI. So it's already right there, live. So population scale applications will come from government. And what that does is it ensures that we know how to do this at scale cheaply, because that is a big thing about our DPI also.

The fact is, today you talked about UPI. UPI does 17 billion transactions a month, 400 million active users, 50 million merchants, and you can make a one-rupee payment. People are paying one rupee, and there's no transaction cost. So the ability to build at that scale cheaply is something we have perfected. And we'll apply that... India will apply that same thinking to AI. So I think the government will be actually the pioneer in applying this to different cases.

And then the private sector will do it for that. For example, if you look at UPI again, now NPC has implemented an Indian language voice command for a payment. So I can speak to the phone and say, transfer money to this person. So all that will start coming.

Nalin Mehta: So we are going to take questions for Nandan from the audience. So I think there are chits on your table, so please do send your questions if you have them, and once they're with me here, we will put them before Nandan to answer. Let me... On the question of government, America has started this Target program, right? In terms of investments in AI, we have our AI mission, we have a national minerals mission, which is now being started up. And after DeepSeek came out, further investments have been announced by the government of India on GPUs and various aspects of AI, both on the hardware side and other side. What will it take to create an Indian equivalent of Stargate? I mean, to really scale up from the government side.

Nandan Nilekani: Indian equivalent of...

Nalin Mehta: Stargate in the sense is...

Nandan Nilekani: Oh, Stargate. Stargate is not a government thing. It's three private guys. It's three companies, so I won't call it a government thing. It was announced in the White House, but that doesn't make it a government thing. So I think... I don't think we need to go at that level, because, again, DeepSeek has shown that you can do this much more frugally. But I think what they're doing, you know, what the government is doing is the right thing. They have the AI mission, they've got competitive bids on GPUs, so they're doing a price discovery of GPU compute, which is coming out very attractive, so that's reducing the cost of using this. They have invited people to bid on their LLM strategy. They're going to support them. Many people will bid. So I think it's all the right... And then they're building all the applications, PM Kisan, you know, the Vistar and all that. So I think... And, of course, they support AI for Bharat. So AI for Bharat gets funding from me and also from the government of India.

Nalin Mehta: So in terms of... Let me ask you about the two recent things you've been involved in on applications of AI. When you write about... When you co-authored this paper on the Finternet, what DPI did was it sort of put... It created the building blocks, which people could then build... It created a digital highway where people could kind of drive on that highway. So when you're talking about Finternet, one of the arguments I think you make is that the traditional financial world is kind of behind what's happening at the moment. So how do you use aspects of DPI or aspects of technology, tokenization and so on to bring traditional transactions like stocks and other bank deposits into this? How do you use AI for that? If you can explain that argument.

Nandan Nilekani: Yeah, so the idea of Finternet is essentially bringing in the ability to tokenize all kinds of assets. So assets can be bonds, deposits, art, horses, real estate, whatever. And what tokenization does actually goes back into the future. In the sense, if I give you a 100 rupee note, that is a token. Because that note carries everything about that. All the properties of that note are in that note. So when I give it to you, you know it's 100 rupees, you know it's issued by the government. Everything is all packaged there. Or if I give you a file with all my real estate papers, that's a token. So with that file you can go for it.

So digital tokenization is taking an asset and wrapping it around with all its attributes so that it's freely tradable. And in some sense, that is coming out of the crypto world. So the crypto world took things like public blockchains, the concept of immutability, which means that you cannot change an object. So that means you can then, if I give it to you, you can't change it. But they applied it to the cryptocurrency world, which then went into another spin of, you know, who are they answerable to and all that. So that's different. So we're taking the crypto capability but applying it to the regular world.

Nalin Mehta: So central banks will have to come in, in some form of digital currency.

Nandan Nilekani: In fact, the paper I wrote is co-authored with the chairman…

Nalin Mehta: CEO of the Bank of International Settlements.

Nandan Nilekani: Which is a sort of group of all central banks. And I launched it in Basel also last few months back. So the idea is that, then I'll give you an example. For example, banks can be easier to lend to real estate or land, because land can be tokenized. It's easier to lend to art, so art can be tokenized. It's easier to lend to this. That's one example. Second is, it allows you fractionalization. I heard an interesting proposal yesterday from somebody saying, there's this stadium and we want to tokenize it and fractionalize it so all the fans can buy a piece of it.

Now, think about it. How do you do this, right? So you can take an asset, which is a stadium, some football stadium, and make it into a tokenized asset and fractionalize it into, you know, 100 euro or whatever it is. And then you can buy and sell those fractions. Suddenly you have made the asset liquid. So we have all kinds, or you can use it for 24 by 7 trading. If you look at what's happening in the stock markets, you know, it's open only, say, 9.30 to 3.30 or whatever. You can use this to create 24. So there are many, many uses of this.

Nalin Mehta: But then it will also require different kinds of regulation around digital currency?

Nandan Nilekani: Oh, yeah. See, all these... No, you don't need digital currency. You can still make the payment using the regular currency. But the asset is, it's traded.

Nalin Mehta: In terms of your work with the International Energy Agency, we wrote the foreword for this vision paper last week on what you're calling the digital energy grid. You wrote the foreword for that. How does AI change fundamentally the distribution of energy in the way that you've outlined?

Nandan Nilekani: I think AI does not change, but AI can be... Because the grid is changing with AI or without AI. That's nothing to do with it. But if you do the digital energy grid the way we have designed it, and it's there in the paper, you can apply AI much easier to it. So what is this? I think you should ask the question, what is this? Why do you need all this?

Nalin Mehta: Exactly.

Nandan Nilekani: So I think the way to think about it is that this is the age of electricity. You know? All over the world, electricity is a growth market because consumption is going up. AI data centers need electricity. Electric cars need electricity. Tomorrow, instead of boilers, you will have heat pumps and so on. So the whole demand for electricity is going to go up. So people are calling the next 20 years as the age of electricity because electricity demand is going to go up some six times in the next few decades.

But the big change is in the old model of electricity, you had a single producer, some thermal power plant or whatever, and you sent it over the wire and you paid your monthly bill. In the new model, every consumer will be also a producer. So in your house, if you have a rooftop solar, you have a two-wheeler or four-wheeler EV battery, you can produce electricity, you can buy it from the grid, you can sell it to the grid, you know, the whole net metering and all that. You can store it in the battery, you can charge your battery when the sun is shining, you can sell to the grid when it's at night, when the demand is higher. So every person in the world who has all this at home can become a producer, consumer, buyer, seller and store of electricity. So there'll be millions.

If you look at the Prime Minister's program on rooftop solar, that's one crore, they're talking about 10 million rooftops. Every one of those persons can be in the business of buying and selling electricity. It's a whole new world. So how do you coordinate this massive network of players requires a very different kind of thinking. So that is what the DEG is about. On top of that, you obviously apply AI to every decision of that. So AI is a consequence. I mean, AI can be done without this, but this will turbocharge your ability to apply AI to manage the grid of the future.

Nalin Mehta: So in terms of your work, what do you spend most of your time on? You said you do a day a week for Infosys?

Nandan Nilekani: Yeah, so I'm the non-executive chairman of Infosys. So I spend 20% of my time on that. And I spend another 30, 40% of my time on different types of DPI, AI, all this stuff which I just talked about. The rest of the time, I don't do anything.

Nalin Mehta: So when you're not doing anything, how do you reflect on, what are your thoughts on intellectually? See, a lot of people are talking about AI as a huge disruptor on the scale of, say, what the steam engine was, or what electricity was, and that's changing society as we know it. In terms of jobs, will that affect everybody? Do you see this as a huge problem or as an opportunity?

Nandan Nilekani: No, I see it as an opportunity, but then I tend to be optimistic about everything, so that doesn't mean much. But I think that, yes, obviously it'll have an impact on some jobs, and some jobs which can, but many jobs will not, there'll be only a few jobs that'll be totally taken out. There'll be many jobs where a lot of the tasks will be taken out, but not all the tasks, so they still have work to do. So it'll help human beings to become much more productive and effective in their jobs. So I think it's a net positive.

And also, it's going to create new jobs that we didn't anticipate earlier. Now, if you look at India, you don't have enough teachers, you don't have enough doctors, you don't have enough nurses, you don't have enough of every skilled resource. So if you can use AI and amplify the ability for people so that more people can learn, more people can get healthcare, more farmers can get better information, then it's a very net positive.

So it all depends on how you architect it. You can either do it in an extractive way, where some guy has all the data and then he makes money off you, or you can do it in an inclusive way where everybody benefits. So even DPI at the heart of it, if you think of it from an ideological point of view, is about creating inclusion at scale. So AI also should do inclusion at scale.

Nalin Mehta: So let me ask you this question, that when you came to government, as the head of Aadhaar, you set up the UIDAI, it was a great example of a technologist coming in the private sector, coming in and transforming. I think you had one piece of paper saying, create a unique ID system, but nobody really knew what that meant. A unique identity system. One paragraph, right. We have another technologist of a different kind, of a different makeup altogether in the White House right now, changing US government. So there are people saying India should also have a transformation, a tech transformation of a different kind. So how do you look at what's happening there, and what kind of intervention do we need now at our end?

Nandan Nilekani: Well, first of all, I think, what we did over the last 15 years, 1.3 billion people with Aadhaar, 80 million times a day, 700 million people, new bank accounts, mobile network, Jio used that for acquisition of 100 billion customers, direct cash transfer running into a few hundred billion dollars, UPI. So a lot of stuff has happened in a sense, but it's not happened in a confrontational way, let's put it that way. So I think the difference is probably that.

Nalin Mehta: That's a good answer actually. So I was wondering if you'll sidestep that. So okay, so let's take some questions. There's a question from the audience. Considering the change that is happening due to AI, how can the educational institutions design the curricula in order to engage better with students and with relevant skills?

Nandan Nilekani: Well, one is of course, we have to contemporize the education in universities and that's about classroom. But also I think the bigger question we have to ask ourselves is, what do people do in a world of AI? In a sense, if AI is going to do a lot of things which human beings did, then what do people do, right? That's a core question. And I think there are two or three things that are very clear. One is the human skills of empathy, motivating, leadership, collaboration, those will become even more valuable, right? So I think our institutions have to build those capabilities in our people. It's the human skills. Human skills are not going to go away. They'll become more valuable because you can have all the AI in the world, but if you can't get five people to work together and you can't collaborate, then you can't get anywhere. So I think building those human skills is going to be required.

Second is first principle thinking. Being able to step back and go to first principles and analyze something. That again, AI can't do that because AI is more mechanistic in its approach. And then of course, creativity. How do you get people to be creative? Because that again, it's not clear that AI can do that much. So I think we have to look for what is it that we have to do in a world which is full of AI and still be useful, relevant, satisfied with life and all that. I would rather go there than say learn skill A, B, or C because skill A, B, or C may not be relevant five years from now.

Nalin Mehta: So there's another question. What steps are being taken on safeguarding information, particularly data protection? Because that's a huge concern around privacy and also sovereignty.

Nandan Nilekani: In the context of?

Nalin Mehta: In the context of all kinds of AI applications which people are using. So what steps should government take for data protection and so on?

Nandan Nilekani: I think that's a very important question because AI to be effective has to be trained on a lot of data. And therefore, people will be rightfully concerned if their data is being used and so on. So I think that, and fortunately in India now, we have a very good DPDP bill, data protection bill, and I think that will also govern the AI usage. But ultimately, it's privacy is about personal identifiable information. As long as you have anonymized it so that you can't trace it back to a person, then you should be able to process large amounts of data without affecting any human being.

Nalin Mehta: Well, you know more about that debate than most people because you spent a lot of time fighting that debate.

Nandan Nilekani: Second longest case in the Supreme Court.

Nalin Mehta: That's right. There is a question, and it's an interesting one, that you've spoken about use cases of AI in fintech and agriculture, but what about e-commerce? What about ONDC in particular, which is something that you've been engaged with? What do you see the prospects of that going forward, and any use cases of AI in that?

Nandan Nilekani: No, no, I think you can have specific use cases. I mean, for example, on the buyer's side, AI can help the buyer come quicker to the right purchasing decision. On the logistics side, AI can be used to figure out what's the maximal route to deliver something. So there are going to be a lot of, but they're all point uses of AI to solve something in that system. I mean, today, what we don't realize is that India has been using AI for a long time. We are one of the bigger users. If you look at Aadhaar authentication, you can authenticate with a biometric, fingerprint, iris, or face, and one of the concerns is what if somebody fakes it?

And the algorithm for authentication in Aadhaar checks for liveness, that it's a real human being and not a photograph, and that’s all using AI.

Nalin Mehta: There is another question around misuse, which is about, there's a big concern regarding AI being misused by different actors, especially in terms of deepfakes. So how does one go around protecting it? What kind of safeguards government should build?

Nandan Nilekani: No, I don't know, I mean, I don't know. That's a genuine issue of deepfakes, and we had fakes, we have deepfakes, we have scams. India's the only country which has a scam called digital arrest. I don't think anybody else has digital arrest as a scam category. So we have all kinds of things happening here.

I think, at least in some sense, Aadhaar actually solves that issue of, if you want to show that somebody's a live person, then you can actually do an Aadhaar authentication. So before AI was created, we already had the framework for that. So I think AI will have a lot of it, deepfakes, misinformation, hallucination, there are many issues.

Nalin Mehta: There's another question, what about biases in AI? This is something that you've talked about, especially with languages and Indian language models. So Saket has a question that, how does India address the issue of biases and monitor ethical and so forth?

Nandan Nilekani: I think that's actually one of the best reasons why you should build your own LLMs. Because if you just bring outside LLM, then you bring its biases with it. So if you ask me, what's a very good argument for building India's own LLMs, it's because we can train it on Indian data, Indian context, and that will actually help.

Nalin Mehta: So on that question, in the last couple of months, several people in the tech world have come and said that they disagree with your previous view on LLMs. How do you, do you want to respond to that?

Nandan Nilekani: I was saying-

Nalin Mehta: Or that's too simplistic.

Nandan Nilekani: See, if you look at the big, big four guys, between them, they're spending $300 billion on building up their compute and AI and buying chips from NVIDIA and all that. All I said was, we don't need to get into that thing. That right now, we use AI to quickly make it work. Because, see, making it work is tough. And when the models become commodities, then you can always build a model, which is exactly what is happening. So, all I'm saying is there's a time and space for it.

You don't start by spending billions of dollars building a model without a use case in mind. You use existing technology, make it work, learn how to use it. And then as the technology prices collapse, you build it yourself.

Nalin Mehta: There's another question regarding health, which is something we haven't talked about too much. It's regarding Ayushman Bharat and the digital mission, that hospitals, according to Anmol, who's asking this question, are still reluctant to integrate AI because of various reasons. So how should we curb this reluctance to engage with AI because it might under-report profits, according to this question?

Nandan Nilekani: No, I don't know what is this use case.

Nalin Mehta: I think he's talking about private hospitals.

Nandan Nilekani: Under-reporting, private. I don't understand. But I think, you know, when you do large-scale transformation, whether it's Ayushman Bharat or whatever, you have to find out ways so all actors have a stake in the success.

Nalin Mehta: So, see, coming back to the larger rubric of this conversation was around DPI to AI. One of the things that happened with DPI was that you could create different building blocks of it. You could build Aadhaar. Then the financial inclusion happened. On top of that, NPCI came in. The UPI thing happened, which allowed DBTs and all of that. But it took Nandan Nilekani to build the initial part of it. It took Dilip Asbe to work on NPCI. It was a private, multiple people. I think it's a question that the director of IIM Bangalore, who's here, he'd, I think, asked this, wrote this thing a couple of days back. Who are the next Nandan Nilekani and who are the next Dilip Asbe?

Nandan Nilekani: I can't solve that. But, see, I think we were fortunate that we had a bunch of people who have been working together for 15 years and have a belief that DPI, and now DPI plus AI is the way to go for a large country like India. So we had the original team at UIDAI, and by the way, NPCI started at the same time as UIDAI. That's one of the great coincidences of India. 2009, UIDAI was set up. Around that time, NPCI was set up. And we worked together and built the KYC and the Aadhaar-enabled payment system and the DBT. All that was built in that time.

Nalin Mehta: And the banks were involved, the RBI was involved in that.

Nandan Nilekani: Yeah, yeah, no, RBI was a great support. RBI has been a fantastic support for this. And after I finished, I became an advisor to NPCI. So I worked with Dilip and Pramod and all that, and that's how the UPI thing happened. So it's a bunch of guys who've been working together who have a common sort of goal.

Nalin Mehta: Do we need that kind of movement to be recreated with the similar people or different kinds of people, but to sort of kickstart things in the way that DPI got kickstarted?

Nandan Nilekani: In India?

Nalin Mehta: With AI?

Nandan Nilekani: Of course, of course. I know, it's already happening. There are many people working on AI quite selflessly today. I mean, if you look at Professor Mitesh Khapra at IIT Madras, he's building this AI for Bharat, which is part of the Bhashini program. It's amazing what he's done.

Nalin Mehta: There is a question about, I'm sorry, I'm trying to understand the handwriting, on AI applications in safely and preempting fixed asset failures. I'm not quite sure what that means. Sorry? On safety.

Nandan Nilekani: I'm sure you can do it. I'm not an expert either on the railways or on safety, but I'm sure you can apply AI there.

Nalin Mehta: So let me ask, and we'll have to conclude. So last question on this, you wrote a piece two or four days back or I think last week, talking about the IT Act, talking about how we've had a first phase of reforms with DPI and other things, and this could be now getting the second phase. How do you see the IT Act and how do you see the use cases of AI in particular now on the larger financial underpinning of the economy?

Nandan Nilekani: Again, AI is already in some ways used in the financial system, both in income tax and in GST. AI is used to catch fraud and so on. So that's already happening. I think the Income Tax Act is a big step forward because it's basically a lot of simplification.

Nalin Mehta: From 5 lakh word to become 2 lakh word.

Nandan Nilekani: And also they've reduced the sections. So because we have, if we really want to have ease of doing business and all that, we have to make all these laws very, very simple. And I think that's a big step in that direction.

Nalin Mehta: So final word, Nandan. We have a lot of concern about the economy at the moment. We've just had economic numbers out. There is a lot of uncertainty globally, strategically and so on. How positive or how gung-ho are you about India's economic prospects going forward?

Nandan Nilekani: Well, I don't know about the short term thing, but I'm very confident that if India does everything right, it can get to six to 8% growth compounded.

Nalin Mehta: Thank you. On that note, I think ladies and gentlemen, please give Nandan Nilekani a big hand.