Welcome, welcome, good evening to all our distinguished delegates here for this dinner on DPI and AI. I'm especially happy to welcome Nandan and Rohini Nikkni who will lead this discussion. My name is Arun. I'm a professor at Ashoka University, which is a private liberal arts university in Delhi. I'm also a non-resident senior fellow with RF. My job here today is to really be the moderator of this dialogue, but also to sort of set the scene because I understand from the list that we have been given that many of our distinguished colleagues here come from the respective policy planning divisions of foreign ministries from your respective countries. We of course have significant participation in the room from the private sector, but also from civil society. So there's a good mix, but I'm conscious that this is a crowd that leans more towards the international relations arena as it were, as is expected with a forum like the Raisina Dialogue. So my goal for the next two, three minutes, please allow me to speak for a few minutes to introduce Mr. Nilekani and also Nandan, and also introduce him in a way that hopefully contextualizes some of the discussions that we would like to have with you, and to also perhaps set the scene for some of the things that we would like to hear from you about things that we as India should learn, and maybe some of the developments that you have been following that are happening in India and in Asia, in other economies of the world where digital public infrastructure is also really taking off.
Mr. Nilekani, Nandan is one of the, to most Indians, Nandan is instantly recognized as one of the seven co-founders of Infosys. I don't think even to this crowd Infosys needs any special introduction. It is a consulting and offshore innovation giant that truly changed the way India thought about software and the digital world, and the world thought about India and India in the digital world. So I'm going to divide my introduction and also sort of artificially segregate Nandan's career into three phases. The first involving his co-founding of Infosys and the advancement of Infosys into the giant as it is today was an important geopolitical moment for us. The rise of Infosys roughly coincides with a period of the late '90s and the early 2000s, a time that was otherwise marked by extremely frigid relations between India and the United States. For those who, and I anticipate most of you have memory of this or have heard about it, 1998 was the year that the Clinton administration in the United States imposed sanctions on India following the second nuclear test by India in Pokhran. So it was very much a period characterized by minimal government-to-government engagement and an atmosphere, as I said, of frigidity following the nuclear test. Infosys was among the wave of companies that broke through the bilateral relationship and offered India for the first time, and also the United States, a template that allowed for business to lead political relations and in fact guide the future template for one of the more important bilateral relationships that we have. The importance of Infosys and other companies that were, I don't want to say that Infosys was the only company involved in this endeavor at that time, the role that Infosys and other companies played to troubleshoot what were acutely relevant and time-sensitive concerns. I don't need to introduce or talk about the Y2K crisis to many of you in the room. But the role that Infosys and other companies played to troubleshoot the Y2K crisis in the United States and other jurisdictions of the world essentially broke the ice between India and the United States. And this I think Nandan would concede was not by design but by circumstance. But it's an important circumstance that led Indians at least to understand for the first time that business and societal relations, especially individual person-to-person relations, can truly guide and transform an important and transformative geopolitical relationship.
The second phase of Nandan's career, I think it does not begin in 2017, but I think 2009 was when he was appointed the chairperson of the Unique Identification Authority of India, UIDAI. UIDAI was the now statutory body that helped institutionalize Aadhaar. Aadhaar was conceived as a biometrics-driven technology to address one of the acute governance problems that India faced for the longest time, which is that of identification of beneficiaries of public benefits and utilities, plugged in most directly with the public distribution system, PDS as we call it in India. Aadhaar transformed the landscape of governance of state-society relations by one, more effectively identifying the deserving beneficiaries of many state schemes, both at the federal level and also at the state level. It's universally adopted in India. And of course, you're more than welcome to ask Nandan about Aadhaar either from a technological perspective or from a policy perspective. But this identification problem that Aadhaar cracked in some respects, which are sort of as I said institutionalized into the UIDAI setup, was also an important moment considering what was happening in international relations at the time. In the middle and the early 2010s, there were a couple of realities that were beginning to be realized by the states. One, states are the primary unit of international relations, but the debates on the Sustainable Development Goals and the Millennium Development Goals reached a crescendo at that point. And the SDGs, the MDGs and the SDGs were nothing but a reflection of the reality that states were accountable to their people but had not done enough to advance their socioeconomic needs. Of course, at this moment to the best of my recollection there's no conversation that's happening between Aadhaar and SDGs to the extent that it is, or the architects of Aadhaar and SDGs, that is happening today. But I just want to place this in context. The second development that was happening around about the same time, particularly in the 2010s, was that states were sovereign until they were not. You had a number of states that were characterized fairly or unfairly as failing or failed. Their primary problems stemmed from their inability for reasons of capacity that had to do with history, legacy of colonialism, entirely justifiable reasons, but ultimately to do with a lack of capacity to address the socioeconomic needs of their citizens. And if you notice, there is of course no simple linear relationship. It used to be the case that the countries that did not have this governance capacity to address the needs of their citizens, those were the countries that also had incursions of sovereignty. Sometimes there were overt incursions of sovereignty or sometimes their sovereignty was taken away by a thousand cuts. And so this development happened at a time when India was step by step modestly bolstering its own sovereign capacity, and it is a big country. So this was a progressive measure whereby we were able to advance our state capacity primarily in the business of identifying and reaching, targeting beneficiaries of state schemes.
The third phase of Nandan's career, if I can sort of frame it in the context of international relations, is the work that he has done along with volunteer groups to set up what is called India Stack. And India Stack is a set of what are called application programming interfaces. APIs are nothing but lines of code that allow software to talk to other software. So you use an API developed by Amazon if your enterprise is running on Amazon Web Service. You use an API if you are an app developer who wants to be hosted on Google Play or iOS store. But Nandan and his team developed India Stack, these series of APIs which include the most famous of them, I'm sure you've heard in your brief time in Delhi, is the Unified Payments Interface. This widely successful and popular payment gateway interface that is used by millions, that has millions of transactions every day. The innovation and the point of difference between how APIs were traditionally seen by big tech, if I can call it that, and how this was being developed in India by Nandan and his team at that time, was whereas APIs were seen as developments leading towards a walled garden, meaning if you were at AWS you would always be at AWS, if you were in Google Play or iOS you would essentially be beholden to a set of terms that were dictated by Apple or Google, APIs in India for the purpose of at least these public utilities, and hence the name digital public infrastructure, was essentially developed to create interoperable playgrounds if you want to call it that. Interoperability is today baked into the Indian digital ecosystem at least in the critical sectors, and the goal of such interoperability is to not keep foreign players out, some of the most successful players in India's critical sectors such as payments have been foreign players, but to ensure everybody has a level playing field. And the idea is to break down in some respects the walled garden that we were seeing at that time pop up at least incrementally in different parts of the world. This is the third and the final phase of what I want to find as I describe it. I'm sure Nandan has many more innings left to play. But I want to cap it off by saying the work increasingly of Nandan and Rohini has turned towards artificial intelligence, and you have already been baptized as it were to many conversations on AI in the one day that you've been here at the Raisina Dialogue. And they have independently and jointly incubated several institutions and nonprofit entities that use AI in sectors that adequately need addressing in India and still have a long way to go, including education, healthcare and so on. And so the final message that I want to say, although I present a view of there were things happening in the international relations scene that India seemed to have come ahead of, I think nothing could be further than the truth. It just so happened that Nandan has been a key figure on many technological developments that have, I think we can say with some humility, placed us ahead of the curve especially on things that are important to you such as sovereignty and interoperability. And the hope is that we can have a fruitful discussion on how some of the DPI and AI-leaning DPI today can help you embellish some of those themes that are important to you from a policy planning or a foreign policy perspective. So with that brief introduction, forgive me if I have spoken too long, I want to hand it over to Nandan to speak for a few minutes after which we will have an open discussion. Thanks.
Thank you Arun for a great introduction. So I'll briefly speak about what we have done in India on technology and then speak about what we are planning to do on technology. I think several years back there was the idea that how do you give everyone a unique identity. Now you're wondering why would you want to do that? Because in countries where everybody at birth has a birth certificate, that is the unique identity which they can use subsequently. But in a country where in many states more than half the children born did not have birth certificates, they were essentially documentless and stateless and not having that was an impediment for the future. They couldn't travel on a bus or a train or a plane. They couldn't get a job. They couldn't answer to the policeman who stopped them on the road and so on. So that's how the government had the idea that everybody should have a unique ID. And the second reason for that was that India was building a welfare state and as the welfare state started distributing welfare benefits it found there was a lot of inefficiency, fraud and diversion of benefits. So both to tackle the efficiency issue of welfare benefits as well as to give everybody an ID in the new world, they appointed me to run this thing called UIDAI. I was employee number one of this organization and we over the course of about four and a half years we built an ID system which was online, which worked on the cloud essentially, it could be verified biometrically which means you could use a fingerprint or iris or photograph to verify, and all it did was say John is John. It didn't get into what he was entitled to or something, just proved that he was the person he claimed to be. And today in the world of AI this is called as proof of personhood, but we did this 17 years back. So now every Indian has this ID. So there are about 1.4 billion Indians who have a digital ID which works online, they can use it anywhere and it's used about 80 million times a day for verification.
And then on top of this ID we built the capacity to do what's called KYC, know your customer, which is required to open a bank account or to get a mobile connection. And that was used under Prime Minister Modi's leadership to build the world's largest financial inclusion program where close to 700 million people opened new bank accounts using the ID, which in turn laid the foundation for the world's largest direct benefits program where money is digitally transferred in real time to everybody's account. If it's a pension or a benefit or if there's an emergency or there's a climate issue, we have all the infrastructure now, the plumbing to make this very efficient and 100% accurate, goes to the right person. And then when we are thinking about it, we had to figure out how to get people to withdraw money from their bank account. So we built a whole network of a million business correspondents where people can use their ID and get draw money out. Then we said why do people even have to withdraw the money? Why can't they just spend it from the account? So we built the world's largest payment system called UPI, Unified Payments Interface, which is used by 500 million people today to make digital payments. And this system does about 21 billion transactions a month. It's much larger than any global payment system. And it's free. When you pay somebody 10 rupees, they get 10 rupees. So it's a friction-free public utility. And then we also built a system on top of this where people could use the digital footprint of their activity to get access to credit and other things. So what began as a program to essentially advance welfare benefits has now become a program to provide the ladder of aspiration, but not just for one or two people, for 1.4 billion people. So somebody can use the ID. Somebody can get the ID which is very easy to get. With that ID they can open a bank account. They can get a mobile connection. They can start a small store. In that store they can start taking UPI payments all digitally. Then the digital history of the payments, then they can take it to a lender who gives them a loan on the basis of the business they're doing and they can grow their business. And if they want to learn a skill, they can go to a skill place, get a skill, get a skill certificate which they can keep on their phone and show somebody else. So the whole thing is actually a pretty elaborate digital system at population scale, at a low cost, very frugal, and built here and built in a way that its complete sovereignty is right here. So I mean we did start by saying we want to create some sovereign stuff but it ended up becoming a very strategic thing because it provided the sovereign strength in many areas. So this has now become a part of everybody's life and it has contributed we believe to inclusion as well as to the economic growth that you're seeing in India today. So you're getting that growth because you built an inclusive architecture using digital technology that allows everyone to participate in society. So we then said let's take this globally and we worked with other philanthropists, my wife and I, and we have built a whole infrastructure to take this to many countries and our goal is to get to 50 countries in five years. We got to about 39 countries who use some form of the technology approach that we have done here in those countries. So that's another part of that.
But I think now the challenge obviously is about AI. Now clearly there are a lot of issues on AI but we think that the real value of AI for a country like India and I think for many countries is not the technology but how useful is it for human beings, otherwise what's the use of all this stuff. So our focus here is on AI diffusion. In other words, how do we put this to work for people, because one of the reasons why Indians are willing to accept technology, they have seen it doing good for them and we want to continue in that strain. And we think AI there is a race to the top and a race to the bottom. The race to the bottom is the promise of huge job loss, the fact that you have all this fake stuff, fake news, fake this, fake that, the addictive nature of AI, the fact that you're really seeing all the harms of AI. And the race to the top is all the promises being made. AI will cure cancer, AI will solve world hunger and all kinds of things. But in the reality in the real world today, the race to the bottom is much faster than the race to the top. And therefore there's likely to be as big a backlash on AI as anything else because in some sense you can argue that the whole backlash against globalization is because middle classes in the west did not see any value from globalization. So they protested through the ballot box. Similarly it'll happen with white collar. If everybody's going to lose their job they're jolly well going to ask the politicians to put constraints on AI. So we have to make sure AI works for people and in India that's our plan. Our plan is to identify major areas, agriculture, education, healthcare etc. and actually apply AI at scale using the same principles we applied earlier in digital public infrastructure now to AI and we're also learning how to do it quickly. For example, we did an entire thing on agriculture support for farmers in the state of Maharashtra which took us 9 months to build and then we took that same approach to Ethiopia where we were able to build it in 3 months and then we built it recently for Amul which is India's largest milk cooperative, 2.6 million farmers, 22 million cows, 1400 vets. We had to solve the problem of how do we get the latest veterinary knowledge to the farmers and we built that in 3 weeks. What I'm saying is that we think now we have the methodology and an approach to take AI to work positively for people and we think that's going to be the long-term future of AI is how is it going to be used. So in broadly therefore I'll end here by saying that what has happened in India is the use of technology at the scale of an entire population to create tangible material positive benefits for people and that has led to both social cohesion, inclusion as well as economic growth and we'll continue that with the AI journey. So I'll stop here and take questions.
While you chew on that and also your food, let me start then with a question because I think I mentioned to you that many of our colleagues here today are from the policy planning division of their respective foreign ministries and one of the concerns that came up in one of the early sessions today is, you know, we can creatively develop uses for artificial intelligence but we're still living in a world where there are these two powers who have a great deal of control over supply chains and so how is India thinking about deploying these use cases while navigating some of the challenges that invariably associate with, you've heard what's happening with Anthropic and OpenAI and also at the infrastructure layer.
I think I personally believe, and you may think I'm a contrarian, but AI technology will be a commodity. That what you're seeing now, I mean there's no winner take all in AI. And you know when you have a race in technology because you believe that if you're first there then you'll be able to catch the whole market, and you can say that has been true in earlier eras, only two people who make mobile phone operating systems, the only three people who provide cloud services in the west and so on, but AI will essentially get commoditized because of two, three reasons. One is the massive capital being thrown at it both in America and in China and therefore you'll have a large number of very large companies that build technology. In the US you have five companies that build frontier models, maybe a sixth, and in China you have at least five building frontier models. And also the knowledge of building this is disseminating around the world very fast. So if you have enough data and enough chips you can build these AI models. So I believe that that's not where the action is. The action is really about usage, what do you do with this stuff. Now the moment you commoditize the lower levels of AI then you're less dependent on anyone because you can always choose to replace one supplier with another supplier or one country with another country and people will not be able to control the access to this. So I think and nobody can own the whole stack because it's such an interconnected world that there'll always be different parts of the stack which, so even if somebody shuts off a company then somebody else will do it. So I personally feel focus should be make this stuff useful to people and the rest will take care of itself.
Perfect. And this is also an issue that I'm happy that he brought it up, that is of concern to a lot of our colleagues who are concerned about sovereignty. And the basic argument that you see not just from Nandan but some of his colleagues that I've had a chance to interact with, it's fine to say that you invest in lower levels of the AI stack and perhaps you should, there is an argument for that if it is possible, but as long as you build use cases it is possible for you to one, strategically orient your economy towards those use cases, develop, advance, even the use cases change, and when there is no vendor lock-in, there is no vendor lock-in because you are ultimately driven by the clarity of purpose in regard to your policy planning. Of course there may be market factors and also geopolitical factors that sometimes create lock-in scenarios. But I think we're at a point of time and I'm specifically looking at my European colleagues where we are acknowledging a bit more the challenges of being locked into certain technologies.
We have one question here, Matt, please just introduce yourself.
Yes of course. Hi my name is Anna Tonko. I come from New York and I'm not on the policy planning side. I'm on the private sector side. I run my own company called Sustainable Impact. I'm an entrepreneur and increasingly my work is focusing more and more in the space of accelerating consequential innovations and especially in the deep and the clean tech space. And so some of the broader themes that I'm seeing, Mr. Nilekani, I think we're in the cusp of mass adoption for fusion energy. There will be breakthroughs within the next five years or so in fusion energy production, in converting this to commercial space. On the other side we're looking at BCI, the brain-computer interface technologies that will allow on the consumer level people to control essentially computers with their thoughts. And so as I reflect on an excellent introduction that I think was done on your illustrious career, I can't help but wonder when you looked at mass adoption, so you were in charge of creating trust in Aadhaar and in many systems that required complex explanations for people who probably would not have had these before. How do you ultimately create on a societal level trust in new systems and create inherent buy-in and understanding and things that are not familiar that have as you've said many positive but also negative drawbacks and how do you do this across different cultures? So it's maybe a bit of a more on a philosophical front but I'm curious if you also have any practical reflections from your experience.
You know, I think adoption at scale was obviously one of our big, big challenges. But we discovered, and again this was because the factor conditions in India, it may not be true in a more developed economy, that people were hungry for an ID. People needed it because it was a bottleneck for them. It was an impediment. It was coming in the way of their lives. So they didn't have to be convinced about the need for this. So I think similarly, we have tried to use technology as an empowering and liberating way of giving people access to things they didn't have before and once they get that we didn't have any problems in selling the idea or getting mass adoption. Right? Just to give you a sense of the scale of what we did, at peak enrollment in Aadhaar was one and a half million people a day. So we had 35,000 enrollment stations around the country each enrolling 50 people and they were all full, you know. So that's the scale at which we had to do it. So if your purpose and intent is right, if you can explain clearly the value proposition, if you can show the benefits, and you keep it simple and hide the complexity behind, we think that you can drive mass adoption.
Hi, I'm Charlotte from Northstar. I live in Australia. I'm also private sector and actually my question was very similar to yours. So great kicking us off. Now Nandan, I'm really curious to hear from you in addition to this question around change adoption which we see to be a challenge in almost every industry with human behavior shift. I'm curious to hear what do you observe in other leaders having sort of been capable to horizon scan and see problems and challenges that aren't yet here to be able to envision and create and innovate solutions that aren't yet available. What do you see other leaders could be learning or really need to start stepping into in order for us to really face the challenges that are coming down the pipeline next?
Well, I think you know in India, we're very fortunate because of the alignment of multiple factors. You know, we had a political leadership that understood technology and believed in technology. We had technologists because India was, as he explained, so much of the world's development happened here. We had the talent to actually make all this happen. And I think by and large people had nothing. So this actually they saw something as positive. So I think in other countries they may not necessarily be at the same stage of where it is. But I'm convinced that if you have the right intent, if you design these things in a way that it's inclusive, that's empowering, that's liberating, and if you're able to consistently give this message, you can really get people to adopt and accept.
Thank you. My name is Tin Har. I'm based in Paris. But I'm going to speak under my previous capacity and I was the managing director of the Better Than Cash Alliance. India really has become the leader of the world in terms of not only technology but inclusion. So really looking at technology for inclusion and really put India on the map as the leader of the world and you personally have also led this. So it's not only India but you personally and both of you as you know your personality by taking time to inspire other countries and having worked with 30 countries in the global south many of them said well if India can do it I can do it. And so that inspired because you know it allowed countries to project themselves with a development model that was not a
A couple of things. One is that when we did this, this was unknown, right? We're just doing something that had not been done before. So when you do it, then it becomes known. That itself reduces the risk. But equally, we have spent a lot of effort in packaging what has been done in a way that other countries can adopt it very fast with less friction. For example, the ID system we talked about, Aadhaar, an open-source version of that was built at Bangalore at IIIT-B with funding from a philanthropist called MOSIP. And now MOSIP is getting rolled out in 28 countries, but they didn't have to build it. It's already built and it's available as open-source software. And then we set up a consulting group to help people implement in these countries. So we've done a lot of work in building the nuts and bolts of dispersion of technology. And so I think there's a lot of work done to reduce the friction. We understand that not everyone has our advantages, but if we can make it simpler to use, automate it, simplify it, then we'll increase the adoption.
Thank you. I think I might be the first questioner from the public sector. I'm Sandani, Director General for Policy Planning at the Korean Foreign Ministry. I was wondering what your perspective might be on the ongoing debate within Washington about the wisdom of selling H200 chips, Nvidia chips, or advanced chips to China. The debate taking place obviously between whether or not the country should seek to make China continue to be dependent on US-based chips or whether it would be wiser to curtail altogether and curb the exports of advanced chips to China. Where do you sit on that argument?
I'm a humble plumber and you're asking me to answer geopolitical questions, but I think these are perhaps efforts to restrict the pace of AI diffusion and to retain technological superiority or technological control. But I don't see it's not going to work. AI is too big and too distributive to be controlled by any country or any company. And one is of course because there are so many of them. You know, if there are two people it's a duopoly, there are 10 people it's competition. So I'm very happy when there are 10 people because they're all competing. But also because if you look at what one of the arguments being made is that the Chinese companies are doing distillation, which means they're using existing frontier models from the US and extracting in a way by asking questions and using that to train their models. You can't stop that. So the nature of this technology is that it cannot be controlled. It is going. So I think if they are trying to prevent Nvidia chips, it's going to delay the inevitable. And today if you look at, like in this week, DeepSeek is going to launch version 4 which they say they are going to launch it on the Huawei chip. It's not even on Nvidia. So if the Chinese are able to make the full stack, DeepSeek models are running on Huawei chips, they have broken loose. So I think this is not something that you can control. I think it's just going to spread and the action will move to how this stuff is used. I mean, do you ask who's your electricity provider? You don't know who it is. Some guy in Korea is providing electricity, right? Same thing would happen with AI.
Thank you very much. Excuse me. Ruben's pleasure, head of the policy planning unit also at NATO headquarters. Thank you very much for the invitation and it's been really very enlightening to listen to you, sir. I had over the years heard about the India Stack and what you've been building here and it's really truly a privilege to hear it from you. In this sense, being of course from the public sector, from a public policy planning variety, I am struck by the sort of larger governance nature of this issue. And when I look at what China for instance is doing on smart cities and exporting its technologies through the means of technology but also exporting a governance model, I sort of see there's elements. This is not just technology, you're clearly also helping to export a way of dealing with between the state and the citizen, between the way that the social contract and how it's embedded into the technology. And I'm wondering, you mentioned a number of countries that have been now picking up on your work and taking it up. Is that geopolitically, is the ideological dimension a valuable point for them that they see what they like in the Indian model and not say the smart cities that we see in China?
Yeah. So you know, right from the beginning we have not tried to create something which is a lock-in for countries and that's not our ambition. Our ambition is not to make them dependent on anybody else. So we are essentially distributing this technology as open-source, no strings attached. You can take it, use it, do what you want with it. So I think once people see that our philosophy is not control but our philosophy is to genuinely share the value of what we are doing, I think that is helping. You could argue it's part of India's soft power and you could argue that India is getting a good name because of this. So that's certainly true but there's no financial benefit nor is there a control benefit because we really think that if the world's countries are going to implement technology at speed and scale, you can't do it by just going to a bunch of vendors. It's not going to happen. You need this kind of ability to build your own capability. So that's the goal.
Hi, I'm Kevin Baron. In my previous life, I was a Pentagon reporter covering the military for 15 years. And the biggest story in AI right now in the United States is Anthropic and the war with the Pentagon. And I wonder if what you're saying or how what you're saying applies to the world of national security and exquisite technologies where there are only a few vendors who can provide the type of AI that the Pentagon or the CIA may need. The whole argument was Anthropic was the best and that's why they were the one and now they're not. In a world where we have major political figures battling over a few companies who are the only ones able to provide what they provide.
No, I think Anthropic is very good. I know Dario pretty well and I respect them, but they will find others. If the American government doesn't, if Anthropic doesn't play ball, they'll find somebody who plays ball and who's not going to stand on saying you can't use it for this and you can't use it for that. So I think nobody has the power to control this because there are many people operating. Look at the US, you have Anthropic with Claude, you have OpenAI, you have Meta, you have Elon Musk with xAI, whatever. I mean, you have Google with Gemini. That's five right there. And there are at least three other guys who claim they're going to build frontier models. Go to China, there's Qwen from Alibaba, there's DeepSeek, there's MiniMax. So this technology is just going to diffuse. In India, last week we had a company called Sarvam which launched an Indian large language model with 100 billion parameters. Now a year back, 100 billion parameters was cutting edge. Today a trillion parameters is cutting edge. But what they built here is what somebody else built last year. So this stuff is just going to keep, you know, you can't bottle this stuff and say only I have it and you only buy it from me. It's not going to happen.
I'm Rajiv Mantri. I work in the private sector in India. So my question was more on where does DPI go from here in the AI era. So you explained the different layers that enabled first the identity layer, then the payments, then the lending, and then few other things which have now come on top of even that lending. So as we see more AI diffusion, what types of DPI could emerge? I guess we're looking at fifth, sixth, seventh derivative now in terms of if you look at the first foundational derivative as Aadhaar.
No, I think basically AI in DPI or whatever is how do you accelerate rapid adoption? How do you accelerate rapid diffusion? And therefore it's how do you create the rails by which you can quickly replicate and implement something? I'll give an example of a DPI kind of thing. One of India's challenges is language. We have 22 official languages and a few hundred languages and dialects and a lot of small tribal languages which very few people speak. Now computers couldn't deal with that. But now we have worked and created an open-source dataset of Indian languages which has actually collected samples of voice in 22 Indian languages which is a public good. Anybody can use it. So you can be Meta, you can be Sarvam, they all use it. So by creating that public good, you're accelerating the diffusion of use of Indian language in AI, which means that you're improving inclusion. So today there, you know, there are a Bhil tribe in Nandurbar in Maharashtra, also they are in Madhya Pradesh. In three weeks they have built the AI capability in Bhili which is their language, unimaginable, which means that now a Bhil farmer will get data or information on the chatbot in their language. We couldn't have done this before. So these are all using AI to massively increase inclusion. So I think we'll have those kind of ideas all of which will help in propagating AI at speed and scale, learning from our experience earlier.
Can I ask a related question? What would, if a country wanted to embark on this DPI journey and maybe they'll be, I don't know for a lack of a better term, AI native, what should they do differently from India that learns from the drawbacks that we face?
No, I think you know, what is the problem you're trying to solve? I mean, in India we had to solve the problem of identity. So we had to start there. In the West you don't need to solve that problem. But I think there is a huge need to use technology to create interoperable systems which reduce friction for people. People to get a job, people to get access to education, people to get more learning for a farmer, whatever it is. And that's how you will create a positive environment for AI because AI can do huge gains and it's going to do huge harms and if you don't watch it, it's going to do harms and then you're going to have a big backlash. So I think governments and policy makers have to look at how do we use AI to create positive benefits for the citizens. The starting point for different countries will vary but you know, if you sit down for some time we'll figure out how we can apply it to a different country's needs.
Hi, my name is Alexandra Sidel from the Australian Foreign Ministry. I have a question about scale because you're talking about millions and millions and millions of people. The countries in our region are often micro states and I'm interested in whether what you propose that works for billions of people can work in countries that are micro in nature.
