Building Scalable AI Through Global South Partnerships
20 Feb 2026 11:00h - 12:00h
Building Scalable AI Through Global South Partnerships
Summary
The session began with Sunil Wadhwani explaining that he and his brother founded the Badwani Institute for Artificial Intelligence in 2018, at a time when AI was still nascent and few resources were directed toward societal challenges [17-20]. After early setbacks they refocused on partnering with government ministries to identify priority problems, leading to a national tuberculosis (TB) initiative prompted by the health ministry’s focus on TB as the leading infectious-disease killer in India [41-44][45-48]. The institute developed a smartphone-based cough-analysis tool that instantly estimates a person’s TB risk, which has become a national standard and is the only such solution worldwide, prompting WHO to call it a potential game-changer [56-62]. They also automated sputum-test analysis across 64 government labs, cutting result turnaround to one day, and created AI models that predict which patients will abandon treatment so that 2,000 caseworkers can target the most at-risk individuals, contributing to a 25 % rise in TB detection last year [63-70][71-73]. In education, they addressed high dropout rates among primary students by deploying an AI-driven reading-proficiency suite that generates personalized exercises; after a successful pilot, the Rajasthan state mandated the tool for three million children [75-84][87-90].
From these experiences Sunil distilled key lessons: scaling must be built with government from day one, solutions need to be designed for large-scale deployment early, and they must integrate with existing digital public infrastructure such as the TB case-management system Nikshay and the school platform Rakshak [92-100][111-124]. He emphasized that tools only succeed when they make frontline health workers’ and teachers’ jobs easier, otherwise adoption stalls [125-127]. Ankur highlighted the gap between innovation and impact and noted the Gates Foundation’s new “Advantage India for AI” pledge to support such collaborations [131-136]. Sunil reported that the institute now serves over 100 million Indians annually, has built more than 25 AI platforms, and is expanding to Africa-launching operations in Rwanda, Ethiopia and Kenya and offering capacity-building for civil servants on AI governance [143-150][155-156].
Lacina Kone added that Africa’s Smart Africa initiative seeks to replicate India’s digital public-infrastructure model, citing India’s Aadhaar and UPI as examples that can inform a continent-wide digital market [198-207][211-218]. Shalini Kapoor described “AI diffusion” as sharing playbooks across borders so that solutions built in one country can be adapted elsewhere, reinforcing the South-South learning loop [179-190][191-194]. S. Krishnan outlined India’s AI Mission model, which provides low-cost compute, sovereign models, and open data pipelines, and pledged to share these resources globally, underscoring the summit’s role in democratizing AI [305-319][320-327]. He also noted the Gates Foundation’s partnership in curating sessions and establishing an international cooperation centre to help other nations implement digital public infrastructure [329-333].
The participants agreed that coordinated government engagement, frugal scaling, and cross-regional collaboration are essential to translate AI innovations into widespread societal benefit for the Global South [92-100][125-127][179-190][305-319].
Keypoints
Major discussion points
– AI-driven health and education solutions in India and the importance of government partnership – Sunil described AI tools for tuberculosis (cough-based detection, automated sputum analysis, adherence prediction) that are now national standards, and an AI-based reading-proficiency suite that has been mandated for millions of children [56-70][80-88]. He emphasized that scaling only succeeded after they began working directly with ministries, planning deployment at scale from day one, and leveraging existing digital public infrastructure such as Nikshay and Rakshak [92-106][111-124]. He also warned that tools must make frontline workers’ lives easier, otherwise adoption stalls [125-127].
– Key lessons learned for impact at scale – The team realized that technical brilliance alone is insufficient; success requires (1) early and humble engagement with senior civil servants, (2) designing for mass deployment (training, logistics, hardware) from the outset, and (3) integrating with government-owned platforms that provide data pipelines and distribution channels [92-106][111-124][125-127].
– South-South collaboration and export of the Indian model – After impacting over 100 million Indians, the institute began fielding requests from Kenya, Rwanda, Ethiopia, Indonesia, Egypt, Mexico, etc., and set up a foundation to support global deployments [139-156][160-168]. Panelists from Smart Africa and Kala highlighted the need for shared playbooks, joint investment, and regulatory harmonisation to replicate successes across the Global South [177-194][198-226][259-267].
– India’s Digital Public Infrastructure (DPI) and AI Mission as a replicable framework – The discussion highlighted Aadhaar, UPI, and sector-specific platforms (Nikshay, Rakshak) as the backbone that made AI scaling possible, and S. Krishnan outlined India’s frugal AI-mission model (low-cost compute, sovereign models, open data) that the UN and other countries can adopt [111-124][289-321][322-327].
– Democratizing AI and the summit’s role in fostering inclusive participation – Both hosts and panelists stressed that the summit aimed to move AI from elite circles to “the rooms” where youth and diverse stakeholders can see and shape the technology, aligning with the “people, planet, progress” mantra and the Gates Foundation’s partnership [128-136][289-298][327-334][357-365].
Overall purpose / goal of the discussion
The conversation was designed to showcase how AI can be democratized and scaled in India to address critical health and education challenges, extract concrete lessons about government-led deployment, and use those insights to catalyze South-South collaboration. By highlighting the role of digital public infrastructure and the Gates Foundation partnership, participants aimed to build a shared playbook that other Global-South nations can adopt, thereby accelerating AI-driven development across the region.
Overall tone and its evolution
– Opening (0:09-14:54) – Optimistic and celebratory, with Sunil proudly recounting breakthrough AI products and their national impact.
– Mid-section (14:55-18:40) – Reflective and instructional; Ankur and Sunil shift to “lessons learned,” acknowledging earlier missteps and emphasizing humility and systematic scaling.
– Panel segment (19:46-32:24) – Collaborative and forward-looking, with participants from Smart Africa, Kala, and the Indian ministry exchanging ideas on pathways, regulatory alignment, and mutual learning.
– Closing (32:25-48:45) – Appreciative and inspirational, highlighting the summit’s success in bringing diverse Global-South voices together and ending on a hopeful, “we can do this together” sentiment.
The tone moves from proud reporting of achievements, through candid self-critique and strategic guidance, to a unifying, hopeful call for collective action across the Global South.
Speakers
– Ankur Vora – Chief Strategy Officer and President, Africa and India Office, Gates Foundation; expertise in AI for development, global-south partnerships. [S9][S10]
– Sunil Wadhwani – Co-founder & Chairman, Wadhwani Institute for Artificial Intelligence; expertise in AI-driven health, education, and social impact solutions. [transcript]
– S. Krishnan – Secretary, Ministry of Electronics and Information Technology (METI), Government of India; expertise in AI policy, digital public infrastructure, and national AI mission. [S1][S2]
– Shalini Kapoor – Chief Strategist, XSTEP Foundation; expertise in AI strategy, partnership building, and scaling AI solutions in the Global South. [S14]
– Lacina Kone – Director General and CEO, Smart Africa; expertise in continental AI policy, digital market integration, and public-private AI collaboration. [S11][S12][S13]
– Shikoh Gitau – CEO, Kala; expertise in AI implementation for health and education, private-sector AI solutions in Africa. [S6][S8]
Additional speakers:
– Nandan Hillikini – (mentioned as announcing “100 Pathways to 2030”; role not specified in the transcript). [transcript]
The session opened with Ankur Vora asking Sunil Badwani to elaborate on the institute’s work in India, noting that his own speeches had highlighted AI-driven tools for oral-reading fluency and tuberculosis (TB) screening that were “just amazing” and deserved wider attention [1-7].
Sunil Badwani: He recounted how he and his brother Ramesh founded the Badwani Institute for Artificial Intelligence in 2018, when AI was still a niche field and before the advent of ChatGPT [17-20]. While serving on the Carnegie Mellon University board, he observed billions of dollars flowing into AI research [21-23] but recognised a stark gap: the majority of the world’s population lacked access to quality health care and education [24-26]. After an uninspiring early phase in which prototypes failed to scale, the institute refocused on direct collaboration with government ministries, identifying national priorities and redesigning its approach [31-38].
In health, the Ministry of Health identified TB as the country’s top infectious-disease killer, accounting for nearly two million deaths globally and half a million in India [41-44]. The institute mapped the TB cascade and pinpointed three critical bottlenecks: lack of functional X-ray equipment, slow sputum-test turnaround across 64 government labs, and poor adherence to toxic medication regimens [49-55]. Their AI responses were threefold: a smartphone-based cough-analysis app that instantly estimates a person’s TB risk and provides a probability score (now a national standard and the only solution of its kind worldwide, praised by the WHO as a potential game-changer) [56-62]; an AI model that fully automates sputum analysis, reducing result time to one day [63-66]; and predictive algorithms that flag patients likely to abandon treatment, enabling 2 000 caseworkers to focus on the most at-risk individuals and raising TB detection by 25 % in the previous year [67-73].
In education, Sunil described the pervasive dropout problem among primary-school children, especially in grades 1-5, where early ill-literacy triggers a cascade of failure, frustration and early entry into labour [75-84]. The institute built an AI-driven reading-proficiency suite that generates personalised home-reading exercises for each child. After a successful pilot, the Rajasthan state mandated the tool for three million pupils, illustrating how a targeted AI solution can achieve massive scale [85-90].
From these experiences Sunil distilled four key lessons. First, scaling is impossible without early, humble engagement with senior civil servants; the institute must approach ministries with humility and a willingness to learn [95-99]. Second, solutions must be designed for mass deployment from day 1, with explicit plans for training, logistics and hardware distribution [100-106]. Third, integration with existing Digital Public Infrastructure (DPI) is essential – the TB tool plugs into the Nikshay case-management platform and the reading suite uses Rajasthan’s Rakshak system, both of which provide data pipelines and distribution channels [111-124]. Fourth, tools must make frontline workers’ lives easier; otherwise adoption stalls, regardless of technical merit [125-127].
Ankur Vora highlighted the “innovation-to-impact” gap, noting that progress is not a straight line and that the Gates Foundation’s new “Advantage India for AI” pledge aims to fund such collaborations [131-136].
Building on the Indian successes, Sunil announced that the institute now serves over 100 million Indians annually through more than 25 AI platforms [143-150] and is expanding to the Global South. Over the past year, governments from Kenya, Rwanda, Ethiopia, Indonesia, Egypt and Mexico have requested assistance. In response, a dedicated foundation was created, a team was dispatched to Africa, and operations are slated to begin in Rwanda, Ethiopia and Kenya [154-156]. The institute’s long-term ambition is to impact 500 million people by 2040, echoing Prime Minister Modi’s call to “design in India for the world and deliver these solutions to the world” [161-168].
The panel then turned to the broader theme of AI diffusion. Shalini Kapoor framed diffusion as the creation of “rails” that allow AI solutions to travel across borders, much like the digital public infrastructure that enabled India’s rapid adoption of Aadhaar and UPI [179-190]. She argued that shared playbooks would let a solution built in Kenya be reused in India and vice-versa, reducing the need for each country to reinvent the wheel [191-194].
Lacina Kone, representing Smart Africa, expanded on this vision, describing the continent’s ambition to become a single digital market of 1.4 billion people. She noted that Africa can learn from India’s DPI successes-digital ID, UPI and large-scale election management-and that a harmonised regulatory “cloud” is the prerequisite for investment, likening finance to rain that falls only when the clouds are right [198-206][231-236]. The Smart Africa AI Council, launched in 2025, brings together ministers and private-sector leaders across 49 countries to coordinate computing, data, skills, regulation, market and investment themes [217-225].
Shikoh Gitau added that political goodwill is essential for AI to become both a technological and an economic issue. She defined the “collaboration tax” as the effort, resources, and coordination required for cross-border AI projects, and urged governments to absorb this cost [259-273]. She also cited a resonant comment from CV Madoka … CBC, which underscored the need for shared responsibility [259-273].
S. Krishnan, speaking for the Indian Ministry of Electronics and Information Technology, outlined the AI Mission model that underpins the country’s scaling success. The model provides compute at roughly one-third of global prices [305-307], sovereign AI models (the “AI Kosh” model, as we call it, the AI treasury) [327-329], and open-source data pipelines, all subsidised by the state [322-324]. He affirmed that these resources are available for sharing with the UN and other Global-South nations, and that the AI Kosh is a sovereign model built with taxpayer resources [320-327]. Krishnan stressed that democratising AI means opening the “rooms” and “halls” to youth and diverse stakeholders, allowing them to see and shape the technology directly [289-298][327-329].
In closing, Ankur thanked Sunil and highlighted the Gates Foundation’s role in curating the sessions [128-136]. Krishnan reflected that the summit succeeded in moving AI out of elite circles, placing people, planet and progress at the centre, and announced the establishment of an International Cooperation Centre under the National Institute of Smart Governance to support DPI implementation worldwide [329-334]. He concluded by reaffirming India’s commitment to open, secure and sovereign AI infrastructure [340-345].
The rapid-fire segment captured the overall sentiment: panelists expressed optimism that AI’s collaborative spirit-evident in the collective photograph of partners from Italy, Kenya, Anthropic, Google, Carnegie and the Gates Foundation-demonstrates that AI is about partnership, not competition [391-410].
Synthesis of shared pillars
1. Early, humble partnership with government ministries [95-99].
2. Design for mass-scale deployment from day 1 [100-106].
3. Embed solutions in existing DPI (Nikshay, Rakshak, etc.) [111-124].
4. Ensure tools ease frontline workers’ workflows [125-127].
These pillars, reinforced by concrete Indian case studies and the broader policy context of India’s AI Mission and Africa’s Smart Africa initiative, set a clear agenda for expanding AI-driven health and education impact across the Global South [92-106][111-124][179-190][305-311].
The first question around India. One of the things you’ve done and your organization has done is you found ways of taking the power of AI, democratizing it, and making sure it solves problems that we all care about. In my speeches, I’ve talked about you. I’ve talked about oral reading fluency, the tool whereby for less than, and if I’m stealing your thunder, sorry, but he’ll tell you a little bit more. But I’ve been talking about it because it’s just amazing. I’ve been talking about the fact that your TB screening, you can do things that we couldn’t imagine being done before. So can you tell us more about your work in India?
Sure. Hi, everyone. Thank you. Welcome. Thanks for being here. Thank you for having me. I suspect the way I got over here was they needed, Gates Foundation needed someone for this chat. They looked around. They found this guy wandering around with two badges. They figured that means he’s important. Let’s get him. And next thing I’m sitting over here. But thank you so much, Ankur. So, you know, my brother and I launched the Badwani Institute for Artificial Intelligence here in India about eight years ago, 2018. Back then, AI wasn’t a thing. ChatGPT hadn’t come out. But I happened to be serving on the board of trustees of Carnegie Mellon University in the U .S., where I had studied, gotten my master’s.
So and CMU was ranked then as ranked number one in the world for artificial intelligence research and teaching. So being on the board, I could see all the billions of dollars coming into AI from Google and so on and so forth, even in those days. And it always pained me that none of this money was going into AI for society. You know, three, four billion people in the world out of eight billion don’t have access to decent health care, decent education. AI could be transformative. And that’s what we’re talking about today. But at that time, nothing was going on. So I spoke to my brother Ramesh. We decided, let’s launch this Institute for AI in India.
Prime Minister Modi came, inaugurated it, etc. So we hired a really good team of AI machine learning people, spoke to government, identified use cases, started working, and nothing happened. A couple of years went by, we were developing this, what we thought was really neat stuff, but it wasn’t scaling up. So we took a look at the issues, etc. And then we started realizing, look, we’re not approaching it quite right. We’ve got great AI solutions, but there is a lot more, a lot more to actually having impact than just having a nice technical solution. So I’ll, in a couple of minutes, tell you what we’re doing. So the key lessons that we’ve learned. But once we started figuring out, OK, what we were not doing and that we needed to be doing, then things started happening.
So just to give you two or three examples, as Ankur mentioned, we identify our problems, our challenges that we want to focus AI on by working directly with government. We talk to the health ministry about their national priorities for the next three, four, five years. What should we do? We talk to the education ministry and so on. So three years ago, the health ministry told us that tuberculosis is a very high priority for us. It’s the largest infectious disease killer in the world, kills close to two million people a year. Largest infectious disease killer in India kills close to half a million people over here. And for each person that dies, there are 20 others that don’t die, but they live miserable lives and they are infecting lots of other people as they go on.
So the government, the health ministry said, can you help? So we took a look at the whole. cascade of care in tuberculosis? What’s the patient journey like? Where are the three or four or five key pain points? And we identified, okay, diagnosis is number one, because in these economically vulnerable communities where TB happens, you need x -ray machines, you need sputum analysis, and in these communities, you don’t have all this stuff. You don’t have x -ray machines that work and are calibrated and so on. Problem number one. Problem number two, sputum analysis is another way of diagnosing TB, but these samples go to 64 government labs around India where they are ranulized, et cetera, and it takes time for the results to come back to the patient.
And for those that have TB, you’ve lost a lot of valuable time. Third big challenge is there’s a number of patients with TB who are on the medication regimen, but these are very toxic medicines. They really destroy your body while they’re trying to cure you of TB. So a lot of people stop taking these medicines. and they developed drug -resistant TB, which is much worse, 50 % mortality rate, etc. So we started applying AI to each of these issues. On the diagnosis, we’ve come up with a way of detecting tuberculosis from the sound of a cough into a smartphone. It’s instant. It’s quick. We don’t just say yes or no. We give the risk of this person having TB, what’s the probability, etc.
That is now rolling out nationally, and it is becoming the national standard. And by the way, we’re the only country that has this. It doesn’t exist anywhere. World Health Organization has told us this could be a game changer globally. For the sputum analysis, we’ve developed an AI model. So now the sputum analysis in the 64 government labs, totally automated. Results come out within a day, go back to the patient, treatment starts. Perhaps the most challenging thing. These patients who will fall off their medication. We’ve developed AI algorithms that predict well ahead of time which TB patients are likely to fall off the medication. So then the 2 ,000 TB caseworkers in India, which is a very limited number for 4 million TB patients, they can focus on the right people.
This is impacting now tens of millions of people. Just in the last year, the rate of TB detection, thanks to our cough against TB, has gone up by 25%. You may think that’s bad news, you know, higher numbers, but now we can treat these patients. We can get them on the right, you know, clinical care protocols. That’s one example. Education. Throughout the global south, there is a very high dropout rate of young children from schools, very high, in grades 1 through 5. Problem in India, problem everywhere. We got a call from a very large state government in India that said, we’ve got this issue, can you help? We sent a team in. We’ve got a call from a very large state government in India that said, we’ve got this issue, can you help?
and we had to analyze what’s causing this high dropout rate. We learned that the single biggest reason for this high dropout rate is an inability of these very young children, 7, 8, 9 years old, to be able to read If you can’t read, it affects how you do in every subject, right? Science, history, geography, you struggle, you start failing, you get frustrated, and these are, again, poor communities. Your parents say, forget school, what’s the point? Come work in the field or work in the kitchen, and that affects the rest of their lives. We’ve come up with an AI -based suite of tools that… …and for the child that goes to the teacher, and for the child, we come up with personalized exercises…
stories that they can read at home, but which help them to get better at their specific area of weakness. Each child is different. We were in pilot with the state. They were so impressed, they made it mandatory for all 3 million school kids in that state, in that age group. State of Rajasthan saw it recently. Here, in that age group. So that’s the kind of scale one can get. What’s the difference between what we were not doing in our first 2 or 3 years versus what we’re doing now? What we learned is, number one, the only way to scale is government. You have to work with government from day one. Working with government isn’t easy, right? It’s easy to say it’s challenging, it can be frustrating at times.
but you have to understand how to navigate it. How do you work with senior civil servants? You know, start with an, you know, approach them with humility, not like you have the answers. You’re trying to understand the problem. You want to work with them. Secondly, think scale from day one. You can’t develop an AI solution and then say, oh, I want to use it on one million people. There are issues you have to think in right in the beginning as to how will the scale out? How will large scale training happen in the field? How will frontline health workers or, you know, teachers or other governments? How will they use this? So thinking that way is very upfront.
And in fact, with government, what we do now is once we identify a problem, even before we work on the technical solution, we plan that deployment to scale as to what will happen. And we make government accountable for a lot of it. Just as we’re accountable for the technical side. The other really key learning has been. And that government, and this relates, Ankur, to what you were just saying, has developed a lot of digital public infrastructure. Aadhaar is like the great example that we’re all aware of, right? UPI, United Payment Interface, incredible example of that. So there are lots of things in health care, in education and agriculture where the government has developed this digital public infrastructure.
And it’s critical. We didn’t know this. This was probably our key finding. It’s critical to find a government platform that you can integrate into. So the examples I’ve given of TB, government has a wonderful platform called Nikshay. It’s like a case management system for tuberculosis patients. We’ve integrated everything. We developed algorithms into that platform. The education, this early childhood reading proficiency, each state has a platform. Rajasthan, as an example, has a platform called Rakshak. Rakshak for 70 ,000 schools, 400 ,000 teachers, 8 million students. we plugged our algorithms into that platform. So if these platforms hadn’t been there, we’d be struggling to scale any of this up. The final learning that we’ve had, and maybe this is the most important, is all these technical solutions are great at a macro level to bring down TB, to improve reading proficiency, etc.
But at the end of the day, if the person using this tool, the frontline health worker, the teacher, if it doesn’t make life easier for them, in addition to improving education for the child or healthcare for the patient, it won’t happen. You can push all you want from the top that, oh, you must use this, but there’s got to be pull. They’ve got to want to use it, and that happens only when you make life easier for them.
Thank you very much, Sunil. We’ll do the next question a little bit quickly. But I do want to just… acknowledge a few things, call out a few things. One is this journey between innovation and impact. I love what the learnings you talked about, because we keep on sometimes focusing on the innovation part, and we think that the road from innovation to impact is a straight road, and it’s not. It’s possible, it’s probable, but it’s not guaranteed, and we need to work hard at it. And so your learnings get to it. It’s also one of the reasons why we love our seven -year partnership, and hopefully it’ll be much more as we think about, as some of you know, the Gates Foundation yesterday announced a new initiative, a new pledge around AI for AI, which is Advantage India for AI.
And the idea is to make investments in India for the global south, and we’re looking forward to partnering with you. So, Sunil, one of the places where we do partner, and we’re talking about things, in fact, earlier today, we were talking about work in Ethiopia and Rwanda and Kenya. Can you talk a little bit more about… how you think about your work in the context of South -South partnership and how do you take the learnings you have from one place to the other place?
Sure. So when we got started, our goal was only India, right? My brother and I, we are from India, our hearts are still here. So we weren’t thinking about any place else. But what’s happened is as our AI solutions have been scaling up in India quite dramatically, and we are today impacting over 100 million people a year, we’ve developed over 25 AI platforms in partnership with government. We’ve started over the last year getting a lot of inquiries from governments around the world in the global South saying what you’re doing in India, we need in Kenya or in Rwanda or in Indonesia or Egypt or Mexico. By the way, in India, we don’t just develop these solutions. We also do a lot of capacity building, meaning training of senior civil servants on how you can use AI.
What it’s good for, not good for, etc. We help ministries develop data governance standards, use case frameworks and so on. Then we do the actual solutions development. That’s the biggest chunk of what we do. But then we have a big deployment team. We have close to 100 people making sure that these things get deployed. So we were thinking only India. But over the last year, we started getting all these inquiries. And we finally said, look, we set up this foundation to have impact globally. So now we’ve we’ve have we sent a team out to Africa to meet with several countries. We are starting operations this month in Rwanda, Ethiopia and Kenya. And I’m glad to see a colleague over here from Smart Africa.
We will be partnering in this work. We’re very excited about that. And then beyond that, we expect to be going to a number of other places today. As I said, we’re impacting maybe 100 million people in India. Our goal is by the year 2040. To be impacting 500 million people. We are very excited about our partnership with you at the Gates Foundation. So the Prime Minister Modi, if you heard him yesterday in his speech, he gave a brilliant speech. As part of that, he said he said for the last several years, I’ve been saying make in India for the world. He said, no, I want to add to that in this age of AI design in India for the world, develop in India for the world and then deliver these solutions to the world.
And that is what we’re trying to do. That’s the evolution now thinking. And again, we’re excited, very excited about the partnership with you.
Thank you, Sunil. Are we I think we have a change of plans. Thank you so much. And Sunil, if you could please stay on stage and if I could invite our panel up. Shalini Kapoor, Chief Strategist from the XSTEP Foundation. Lacina Kone, Director General and CEO of Smart Africa. And Shikoh Gitau, CEO of Kala. Thank you.
yeah good good to start okay thank you so much thank you anchor. Thanks for your time uh and here we are and thanks uh Sunil for you know spending some more time with us, uh thanks Shikoh we have been meeting bumping into each other and thanks thanks Lacina thanks for being here So we’ll get into some of the discussions on the South -South collaboration that is spurring innovation and that is diffusing AI into all the sectors. AI diffusion is about the routes and the rails which need to be laid in AI the similar way digital rails were laid in the DPI time. And now they can be shared. They are playbooks. They can be shared.
And AI diffusion actually concept came from it started with generation. Jeffrey Hinton talking about it. He’s a professor in Georgetown in D .C. When he talked about that, how actually electricity was created in. Germany, but it was diffused across India, across the USA, where USA made so much of strides into it. So like electricity, AI is a general purpose technology. It’s a GPT, which is there. And between invention and impact, there’s a big layer of adoption and diffusion, which needs to be there so that AI gets diffused into society. And when it gets diffused into society, something which can be built in Kenya can come to India, something which is built in India can go to Kenya, because there are playbooks which could be leveraged.
Not everybody needs to build everything. Not everybody needs to build the entire stack. How do we learn from each other? And how does that South -South collaboration happen? That’s what is the focus. So I’ll start with the pathways. What are the pathways to scale? and I’ll start with Lesina, that you are leading Smart Africa and you help coordinate across a lot of nations, different stages of AI, somebody in pilot, somebody in production, somebody has solved data, somebody has solved language, somebody has solved voice AI. What do you think are the opportunities for the South -South collaboration in building these pathways together?
Yeah, thank you very much for inviting me. In fact, that collaboration, first of all, before we even talk about South -South, the collaboration is a sense of the creation of a Smart Africa. Because if you look at Africa through just on the Kenya, which is 50 million, Ghana, which is 30 million, or Nigeria, which is 240 million, you’re missing the point. But if you look at Africa as a 1 .4 billion people, but be able to leverage that 1 .4, you need a collaboration. and the scale. So coming from that, when you look at the continent of Africa, which is technically speaking the global south, I don’t want to get there, in the global south, and the south -south collaboration is very important because we do not need to reinvent the wheel.
India has shown the world what the DPI actually means for 1 .4 billion people. It’s working digital ID, a country which is able to organize an election for 850 million people to vote. You have to actually kudos. So we don’t need to reinvent the wheel. Africa can learn a lot from India, even with the use cases. But why use cases particularly? Because there’s a lot of similarity between India and Africa in terms of culture, in terms of value, that we all know we are into the AI, we are into digital transformation. It’s not just… You create a… But the luxury of the luxury of the society is to be able to have a population, inclusion of the population, ethical of the technology, exactly.
So coming from that, that’s one of the reasons to reboost, actually, Smart Africa Initiative. The creations of the Africa AI Council came into play where last April, on April 4th, 2025, 49 countries came together to actually sign the declaration. And subsequent to that, the AI Council came to life on November 12th, 2025 in Guinea after our board of directors, which are represented by the head of the state, announced that, accepted that we could. We had a first meeting already. Then the council consists of 15 members. is not only driven by public, but it has seven ministers coming from seven different countries and eight private sector members. Why? Because in our constitution of Smart Africa, private sector first. We do believe that the government should be creating a conducive environment for the private sector to excel.
It cannot be dominated by public sector. And underneath of the council, we have, of course, six thematic groups, mainly computing, so we can look at the collaboration of South Asian computing power. We can look at it in the data set. We can look at the skills. We can look at the regulation, which is the governance. We can look into the market and we can look at investment. And when it comes in terms of investment, something we need to know. The investments in the prior technologies, the investment cycle is too slow for the AI. Just look back 12 months ago. Where were we? And where we are today? So this is something we need to look at carefully.
We are looking at the three aspects. One, the government needs to be creating a conducive environment for private sector to chip in. The private sector needs to be executing, but they can only execute, as I said, everyone’s cry because they said finance is the issue. I always said finance is not the issue. Finance is the last thing you should think about. You know why? Because I said financing is like the rain. For rain to fall, you need certain condition of the cloud. Those clouds are the regulatory environment, the conducive environment for business private sector because private sector does not like unpredictability. So the third thing, the philanthropics. The reason why I want to speak about it, the philanthropics need to serve as a 2D risk area because these are some of the things government is the last thing to invest in a technology because they want to make sure it’s going to work.
But if you don’t throw business, it’s time for them to accelerate, to use that as a de -risking while the private sector can chip in and so on and so forth. Thank you.
Yeah, thank you so much. And you talked about DPI, you talked about the private sector, public coming together. It’s the entire ecosystem. And on 18th, actually, Mr. Nandan Hillikini actually announced 100 Pathways to 2030, which is a clarion call. If you ask me, it’s a clarion call for people to join to create pathways simply because, you know, if you climb Mount Everest, suppose Edmund Hillary has climbed Mount Everest. And do you think he’ll come back and he’ll say, I’m not going to tell you how I climbed? What is the route I took? Where did I go? Where did I go? Where did I not go? What did I see? He will talk about, right? He will talk about them so that it is easy for other travelers to come in.
So pathway is like that. that if someone has done the AI pathway, others should learn from it, benefit from it. So, Shiko, you were with us on the stage when you joined and you said that, you know, from the Global South, 100 Pathways to Scale, you would like to join us. Please tell us, how can that diffusion help? How can this, you know, we can work, collaborate to, you know, to get the AI use cases from pilot to production?
I can finish? Okay. It’s that collaboration. As I was saying, this idea of how do we bring this multiplicity of thinking together, given that we have the same exact challenges. We have challenges around, we have multiple languages. We have culture and diversity. We have things that we need to be able to work together. How do we collaborate together? And for us, the biggest takeaway is how do we make AI not just a technology, but a political and economical issue? Yeah? That was the biggest one, because the people are there. The builders were there. The researchers were there. The policy makers were there. But we need that political goodwill to be able to make this work together.
And something that CV Madoka from I’ve forgotten the organization. CBC said that struck really a chord to everybody including myself is we need to start having a conversation on what is called the collaboration tax and it’s something DG when we were coming in we were talking about I’ll define collaboration tax as this effort resources and things that you need to put together to be able to collaborate with each other. It’s what the government should be doing it’s what that political part of AI should be doing to bring the collaboration together and how do we make these people come together without the effort of I mean not the effort, the pain of collaboration and that’s what we need to be talking about because the resources, the people are there people are willing to collaborate and work together we saw it this week and as the minister said I said while you’re chasing the Guinness World Record of having the most number of people also chase for the diversity that these countries are seeing thank you so much for bringing Africa and the world to India Yeah.
Thank you, Shikoh. Thank you so much. We’ll take a small break in the panel discussion and we’ll have Mr. Krishnan come here and talk about how scale and collaboration can help in the South -South and what is transferable from India. I know he’s like busy and across all. So over to you, Mr. Krishnan.
I was just going to do one more thing, which is thank you, Shalini, and thank you to the panel for allowing us this small break. For those of you who don’t know, Secretary Krishnan from the Ministry of Métis is over here. He’s had probably one of the most amazing, successful weeks this week. So please join me and give him a big round of applause for. Secretary Krishnan, thank you so much for being here. As a proud Indian, I’m quite excited about the fact what happened this week. I’m also, as somebody who cares about the agenda of this global South. everything that happened in India this week put the Global South agenda right front and center.
So thank you for doing that. There were so many announcements made this week about how we’re going to make progress in the months and years to come together. Would love to welcome you to give a little bit more context of the announcements that were made and the achievements that were achieved this week. Thank you. Welcome.
Let me first apologize to the panel for having sort of stepped in abruptly, but between juggling many things going on across the summit, I think this is a very important session as far as I was concerned, because if this particular summit was about one thing, it was about the Global South. The fact that… India representing the Global South… could actually dare to host this event and also dare to host it on this scale. The one thing we were very clear about is summits thus far have basically been about country leaders. It’s been about CEOs and it’s been about some experts getting together in closed rooms and not really having the opportunity to do what or to actually showcase to people as to what the possibility of the technology is.
And this particular event gave us that opportunity. I think we were very clear that what we wanted to do was to let people into the rooms. We wanted to make sure that people, especially youth, had this opportunity to come and listen to the best minds, there are on artificial intelligence as technology. and to every possible perspective on how this technology can work for everybody. And the second thing, of course, as you’re well aware, we said people, planet, progress. And two aspects of it were very important. One was, or three, if I can. One is democratizing access to AI, all the AI infrastructure and resources. That was one key aspect. The second key aspect was including those who are not ordinarily given access to this, those who are excluded.
And the third key aspect is putting humans at the center of this process to make sure that this is a technology that works for people. And I think the prime minister was very clear and emphatic in his address yesterday where he put people, or manna, right at the heart of AI. So to enable this to happen, of course, multiple things have to happen. We have to find frugal ways to innovate. In order to make these resources available, we have to make sure that we have the resources. We believe that our own AI mission model, the India AI mission model, is one of those frugal ways in which both the compute infrastructure and the model infrastructure and the data set infrastructure can be created for each country because some of this needs to be on a specific basis for regions and for countries.
In India, we are a subcontinental scale. There are 22 official languages and many other languages which need to be taught or which need to be understood. And we understand this cultural and linguistic diversity better than any other region in the world. And we can, at a continental scale, we can contribute in that effort. That is one key element. The second key element, of course, is to create compute in a way so that it doesn’t, I mean, people are not enabled. I mean, people are not able to build moats. around it. That, you know, the implication that you need the kind of resources to do this, that nobody else can do it and only we can do it is not an approach we wanted to take.
So we created this model where the private sector is encouraged to invest. We created this model where access to it is something which the government subsidizes. In the process today, AI compute in India is available at a third of the price that it is available in the rest of the world. I think that has been the significant achievement. The United Nations asked us saying, would it be possible to, once you build it out on scale, would it be possible to share with the rest of the world? This is something that we have committed to them saying, as in when the size is adequate so that we can meet other requirements, we will be happy to share them.
We are happy to share the model even now. And the AI Kosh model, as we call it, the AI treasury is something that we are happy to share even now. we are happy to share the fact that the models that we have supported in India and which have been built out as sovereign models in India that again is technology that we are happy to share with the global south it’s something we can enable some of it is something we’ve built with our own resources so it is in a sense completely sovereign unlike in many other places it’s something that the government has paid for from taxpayer resources and we can use and the third element of course is the data sets and how they are shared now that framework is again something we can certainly share the most important thing and I think that’s what is also showcased so eloquently in the expo is the range of applications which have been created out of this and there are close to 900 startups across all those halls who have done a variety of things even in the main hall we with the Gates Foundation we have set up a lot of applications and we have set up a lot of applications and we have set up the African village which is such a showcase even to the leaders fundamentally about applications which can work there and fundamentally for people to see applications which have worked in different parts of the world which can be taken elsewhere.
So all of those are available. Those are resources which we want to share. These are resources we want to actually give. And most importantly, as I said, I think if there is one thing that this summit reflects, for the first time, we’ve actually democratized AI. We’ve shown you what democratic AI looks like when people are let into the rooms, when people are let into the halls and they can see for themselves as to how this would work. So it gives me immense pleasure that, you know, and a very, very key partner for us in all of this has been the Gates Foundation right from the very outset, right from the planning stage of how we wanted to do this.
This particular part of the set of sessions on the Global South is something, that we work closely with, curated carefully. We have put together sessions which will be relevant to this group, and we have always made sure that in addition to this, of course, on every occasion, whether it is in the space of DPI or whether it is in the space of any of the other applications, we are in a position to support it. Under one of our organizations, the National Institute of Smart Governance, we have now put in a center which is fundamentally focused on international cooperation so that they can actually provide support to other countries where DPIs are to be actually implemented, how to ground them.
And we believe that probably the most effective way of dealing with this is to actually be able to cooperate amongst ourselves. So. So that we are able to take it out. We are able to learn from each other. We are able to contribute to each other. And that is something we are now really ready to do. India knows what it is to be deprived of or denied technology. India knows what it is to actually try and work your way past it. We have managed to do that. We have managed to democratize it. We have managed to make it available to people at scale. We have tried to keep it open source. We have tried to protect it in a number of ways from cyber attacks in each of those areas.
So in this entire technology stack, there is experience. There is the way that we leapfrog different stages. So I think if we work together in the AI space, likewise, there is so much that can be accomplished. And we undertake as a nation, I think I can say with responsibility that we will have devices and we will have structures through which we can sort of deepen this cooperation. We can deepen the support. We can enable this in a number of ways and continue to stay engaged through the Gates Foundation, through the other institutions to actually make this happen. So thank you very much. Thank you to Gates Foundation for curating this particular event. And thanks to all of you for participating in this.
I mean, it’s one thing to arrange it, one thing to organize it. But another thing for all of you to actually come up here, put up with some of the inconvenience which would have been caused. India is not a very convenient country at best of times. but India is a country with spirit and India is a country which fixes things
Okay, we have some time with us. How much? Five minutes. And we need to have a question to Sunil. He has been working at, you know, so many interesting, I mean, I listened to the stories of Wadwaning AI. and they inspire you thoroughly. So Sunil, I’ll give it to you. I want people to hear your message that how can these work that has been done in India, how can it help Global South?
So as Mr. Kone of Smart Africa said 15 minutes ago, the challenges that we have in the Global South generally, really Africa, India, other countries are similar. The values we have, more importantly, are very similar. The strengths that we bring to the table in terms of our talent, our youth, et cetera, are similar. So I think it’s mutual learning. It’s not one way. It’s not that we’ve developed great stuff in India that can just be, you know, taken over. It’s mutual learning on both sides. It’s a mutual sharing of ideas. There are lots of very good things happening in Africa and Asia. There are lots of things happening in Africa and Asia that we can learn from over here.
On the technology side, as you were saying, Shalini, we’ve been fortunate. We’ve had a government that is very pro -technology. There’s a tremendous range of digital public infrastructure that we can access in India that enables, that provides data pipelines. It provides digital distribution systems without which none of this AI can scale. There’s been a very clear regulatory framework for AI that’s been developed in India that really helps. And most important, there’s an openness in government. And I think it’s driven by the prime minister’s vision and belief that technology and AI can truly transform societal development. So those, to me, are the big things, more than individual AI solutions that really make a difference. And I see that happening in Africa in many countries.
Sure. Thank you so much. I think we are literally… We are literally at the eve of the summit getting over. It’s been a fantastic week. meeting the best of the people, listening to the best of the sessions, navigating the traffic, yes. But like Secretary Krishnan said that, you know, we fix everything. So I just want to have one last question to each of the panelists. What’s the best thing you liked of the summit? And to you, Sunil, first. I mean, one moment, one feeling that you will carry forward.
I will give you a counterintuitive answer. AI is making the world move faster and faster and faster. And all the traffic challenges we’ve had over here are teaching us patience. You will get there. Things will happen. Life will go on.
Thank you so much. Shikol, what’s the one feeling you’ll travel back with, back to Africa?
I think my best moment is, and I’m going to pick and be selfish, in the moment in Oberoi when I stood and saw this diverse sea of faces, I think about 300 people, and we’re all celebrating like, we can do this as the Global South. This is happening in our TAF, and this belief that the Global South has something to offer into this AI conversation. So it’s no longer a two -horse race, it’s a multiple -horse race. Thank you.
For me, it’s, you know, our vision is to transform Africa into a single digital market. Coming here this time, it shows me already our future. Having 1 .4 billion people on the one regulation. How does it feel like? So you know what I’m talking about. That’s one of our obstacles, says regulatory harmonization in India, you do not have that. multicultural, you have a multicultural multilingual, you have multicultural what does it feel like, including the traffic in the morning as well of course thank you
And my best moment for the summit was that back in Oberoi on 18th evening, several partners across Italy, Kenya Anthropic, Google Carnegie ORF, Gates Aikstep stood next to Nandini, Kenya and we all came together for 100 pathways till 2030 and we all were together and we were not doing non -collaboration pictures which is going on in Insta, we were all together so it was AI is about collaboration not competition, that’s the theme thank you thank you really enjoyed thank you thank you thank you you are the best man you are the best it’s a pleasure more photos more photos Thank you. Thank you.
And we make government accountable for a lot of it. Just as we’re accountable for the technical side. The other really key learning has been. And that government, and this relates, Ankur, to what you …
EventAnd we make government accountable for a lot of it. Just as we’re accountable for the technical side. The other really key learning has been. And that government, and this relates, Ankur, to what you …
Event_reportingSunil Wadhwani shared concrete examples from Wadhwani AI’s work, including AI systems that diagnose tuberculosis from cough sounds and assess children’s reading abilities in 20 seconds, both enabled b…
EventThe University of Oxfordhas warnedthat AI in healthcare, primarily through chatbots, should not operate without human oversight. Researchers found that relying solely onAIfor medical self-assessment c…
UpdatesCooperation, sharing of technology, and learning are important for effective implementation at scale. Changing mindsets during implementation is crucial for success. In summary, DPI enables citizens,…
EventThis challenges traditional hierarchies of expertise in technology development and advocates for genuine inclusion of diverse perspectives, not just token representation. It suggests that technical ex…
Event“We will move from pilots to platforms, from fragmented data to interoperable systems, from experimentation to execution, from intention to investment.”<a href=”https://dig.watch/event/india-ai-impact…
Event– **Importance of Partnerships and Open Collaboration**: All speakers stressed that no single organization can address these complex challenges alone. They highlighted the need for multi-stakeholder a…
EventOn a positive note, the potential for South-South cooperation and shared learning experiences in the field of DPI was highlighted. Success stories from India and Africa were cited as examples of the f…
EventJohn OMO: Thank you very much, Mohamed. I really appreciate being here. Context. Africa has SMEs contributing conservatively between 50 to 60% of our GDP. Out of that, just about 20% of the SMEs are u…
EventAbhishek Agarwal: Yeah, like what we need to do, like a lot has already been spoken, and I would say that if I have to list what we need to do individually as countries. In India, of course, our focus…
EventAnd accessibility has to be also broadened in terms of multi -modality and also, where necessary, include a human in the loop in the service delivery cycle as well. So some of the inclusion dimensions…
Event-India’s Global DPI Model and Knowledge Transfer: The discussion highlighted India’s role in sharing its DPI framework globally, particularly with African nations and the Global South, emphasizing the…
EventAbhishek Singh, Under-Secretary from the Indian Ministry of Electronics and Information Technology, emphasised that operational implementation requires enhanced regulatory capacity for testing AI solu…
EventThe speaker mentions India’s national ID system (Aadhaar) and payment system (UPI) as examples of DPI enhancing sovereignty.
EventThe AI Impact Summit held in New Delhi brought together ministers and senior officials from multiple countries for discussions on artificial intelligence governance under India’s leadership. The summi…
EventBernard Maissen: Yes, thank you. Hello, everybody, dear panelists. Nina, thank you for giving me the floor. In the global technological and economic competition, countries have a strategic interest in…
EventThere is an acknowledgement of the need for more alignment and coordination in the field of AI regulation. Efforts are being made to bring stakeholders together and promote coordination. For instance,…
EventWe are committed to work together on this through knowledge sharing, co -operation and collaboration. creation and capacity building so that AI solutions are locally relevant, inclusive and accessible…
EventFinancial mechanisms | Artificial intelligence | Capacity development Role of Philanthropy and Public‑Private Partnerships
EventThe tone was consistently celebratory, inspirational, and optimistic throughout the discussion. Speakers expressed pride in young innovators’ achievements, excitement about India’s AI future, and grat…
EventThe tone throughout the discussion was consistently formal, optimistic, and collaborative. It maintained a ceremonial quality appropriate for a launch event, with speakers expressing gratitude, shared…
EventThe tone is consistently optimistic, collaborative, and forward-looking throughout the discussion. Speakers emphasize “limitless potential,” mutual benefits, and shared democratic values. The atmosphe…
EventThe discussion maintained a consistently optimistic and collaborative tone throughout. Speakers expressed enthusiasm and pride in their achievements while emphasizing partnership and mutual support. T…
EventThe discussion maintained a consistently positive, encouraging, and professional tone throughout. It began with excitement and anticipation, continued with focused technical questioning during present…
EventI give this example because I’m fairly confident that when you look it up and when you try it yourself it will work. And I know it will work, by the way. That is, it will fail rather. On the current v…
Event_reportingThere is a mentor talking to them. So we thought we’ll improve the student report first, but that didn’t work out so well. So we, our mentor actually helped us a lot, Akilesh from Tech4Dev. And we ite…
Event_reportingcomplex ones are referred appropriately and millions of lives are saved. AI will not just speed up innovation. It can help bring that innovation to community clinics, to health workers like my parents…
Event_reportingAnd so we thought the biggest use case, the biggest investment would be on making the speakers better for Ray -Ban Meta version 1. And we did that, and the music and audio quality was much better. But…
Event_reportingThe overall tone was optimistic and forward-looking. Panelists were enthusiastic about the potential of smart factory technologies, while also acknowledging challenges. The tone became more practical …
EventThe discussion maintained a balanced, thoughtful tone throughout, combining cautious optimism with realistic concern. Panelists demonstrated technical expertise while acknowledging significant unknown…
EventThe discussion maintained an optimistic and ambitious tone throughout, with speakers expressing confidence in India’s ability to compete globally in AI development. The tone was collaborative and solu…
EventThere has been successful collaboration with stakeholder in Africa and other sub-regional activities
EventThe tone was largely informative and collaborative, with panelists sharing research findings, policy perspectives, and recommendations. There was a sense of urgency around the need to take concrete ac…
EventThese key comments transformed what could have been a standard ceremonial closing into a meaningful reflection on the philosophy and purpose of global dialogue. Brende’s reframing of disagreement as v…
EventThe tone is consistently celebratory, optimistic, and forward-looking throughout the discussion. It maintains an enthusiastic and grateful atmosphere, with speakers expressing appreciation for partici…
EventDistinguished colleagues, as we come to the close of the summit, I want to particularly say how grateful I am for the opportunity to gather again and spend time over the last few days with such a comm…
EventAs the host country representative, Aukrust highlighted Norway’s commitment to digital inclusion through “investing in digital public goods” to narrow divides between wealthy and poorer countries. He …
Event“While serving on the Carnegie Mellon University board, he observed billions of dollars flowing into AI research.”
The knowledge base notes that, as a board member, he could see billions of dollars pouring into AI from companies like Google, confirming his observation [S15].
“The Ministry of Health identified TB as the country’s top infectious‑disease killer, accounting for nearly two million deaths globally and half a million in India.”
WHO data cited in the knowledge base confirms TB as the leading infectious-disease killer worldwide, though the exact death figures differ from the claim; the source provides the broader context of TB’s impact [S116].
“The institute mapped the TB cascade and pinpointed three critical bottlenecks: lack of functional X‑ray equipment, slow sputum‑test turnaround across 64 government labs, and poor adherence to toxic medication regimens.”
The knowledge base lists the same three challenges-limited X-ray machines, costly and time-consuming sputum analysis, and toxic drug regimens leading to poor adherence-supporting the institute’s identified bottlenecks [S115].
“The TB AI tool plugs into the Nikshay case‑management platform, a Digital Public Infrastructure.”
A related source describes a government DPI called Nixia (likely referring to Nikshay) that serves as a patient-management system for TB data, providing context that such integration with a DPI exists [S5].
The discussion shows strong convergence on four pillars: (1) government partnership and integration with public digital platforms; (2) South‑South mutual learning and shared playbooks; (3) democratization and inclusive participation; (4) reliance on robust digital public infrastructure. These shared positions span AI, ICT‑for‑development, capacity building, and enabling environments, indicating a cohesive vision for scaling AI impact across the Global South.
High consensus – most speakers reinforce each other’s points, creating a unified narrative that AI impact will be driven by public‑private collaboration, open infrastructure, and cross‑regional learning. This alignment suggests that future initiatives are likely to be coordinated, leveraging government platforms, shared resources, and South‑South partnerships to accelerate AI‑enabled development.
The discussion shows modest disagreement centered on the preferred scaling model (government versus private sector) and the perceived primary constraints (finance versus regulatory environment). Participants share a common goal of expanding AI impact in the Global South, but differ on the mechanisms and priorities to achieve it. Unexpected divergences arise around the role of financing and the concept of a collaboration tax.
Low to moderate – disagreements are largely about emphasis and strategy rather than fundamental opposition, suggesting that consensus can be built through coordinated policy design that balances government leadership, private‑sector dynamism, regulatory readiness, and adequate financing.
The discussion pivoted around a few core insights: the necessity of government partnership and scale‑by‑design, the power of concrete, nationally‑backed AI deployments, and the need for systematic diffusion pathways supported by regulatory and political frameworks. Sunil’s early remarks about moving beyond pure technology set the tone, which was expanded by Shalini’s diffusion metaphor, Lacina’s regulatory focus, and Shikoh’s ‘collaboration tax’ concept. Krishnan’s framing of democratic AI tied these strands together, positioning the summit itself as a model of inclusive, frugal AI. Collectively, these comments shifted the conversation from showcasing isolated innovations to constructing a replicable, cross‑regional ecosystem for AI impact in the Global South.
Disclaimer: This is not an official session record. DiploAI generates these resources from audiovisual recordings, and they are presented as-is, including potential errors. Due to logistical challenges, such as discrepancies in audio/video or transcripts, names may be misspelled. We strive for accuracy to the best of our ability.
Related event

