Anirudh Suri

Full Transcript

How Local Innovations Could Drive Global AI Breakthroughs, Vilas Dhar, McGovern Foundation

TRANSCRIPT

Anirudh Suri: Hi, and welcome to today’s episode of the AI Futures podcast series for the India AI Summit, where we bring you some of the world’s best thinkers, doers, and leaders from across the sectors working in AI. Today I’m delighted to have with us Vilas Dhar. Vilas is the president of the Patrick McGovern Foundation. He was also a member of the UN High Level Advisory Committee on AI, and, if I may add, he’s a fellow Harvard Kennedy School alum. It’s my pleasure to have you here, Vilas, welcome.

Vilas Dhar: Anirudh, what a delight and a pleasure, thank you so much for having me. I’m excited for our chat.

Anirudh Suri: As we were just discussing, it’s remarkable that we hadn’t met before, even though we’re doing so much work that intersects and aligns. Let me start with a fairly simple question: how has it been being at the summit?

Vilas Dhar: It’s been a really incredible experience, Anirudh, for several reasons. One, as someone who has been working in AI for twenty-five years, I have to admit I’ve never felt so much like a rock star as I have coming to India and being with lakhs of people who are so interested and engaged. I think I’ve taken more selfies in three days than in my entire life. But maybe more seriously, it’s been a really important moment, a moment where we move from a global conversation about AI that for too long has just been scientists, companies, and governments, to one that, as I’ve always advocated for, brings in real people trying to understand how to incorporate this into their lives. And if nothing else, the first global AI summit in the Global South is truly participatory and representative, and I think that’s a fantastic achievement.

Anirudh Suri: I agree. I was talking to someone earlier about this, that finally the rubber has hit the road for a lot of the AI thinkers, scientists, and companies building the new models of our world, for them to actually see the masses of people who’ve attended this summit and understand that these are their end customers, their consumers. I hope that conversation has been two-way, that as much as people have listened to them on panels, they’ve also listened to those end consumers.

A Personal Journey

Anirudh Suri: You’ve had a long personal journey that has brought you here. Let’s spend a few minutes on that. Briefly, walk us through it, how did you end up where you are today?

Vilas Dhar: You’re very kind to ask. My history in this space is almost a collection of both places and times. My parents moved to the United States from India in the late 1970s. I grew up there.

Anirudh Suri: What part of India were they from?

Vilas Dhar: They were from Bangalore, and they moved to a very small town in the United States, in the middle of fields of corn and soybeans, in rural Illinois, about two hours south of Chicago. We were not a rich family, actually quite a poor one, but whenever they could, my parents’ goal was to bring us to India as often as possible. We’d go to a small village just outside Bangalore, a place where telephones were only just arriving, not everyone had one, there was still no running water, power was unreliable. But I saw what happened when a community came together to solve problems by themselves, and it was an incredibly empowering experience.

Then, back in the US, I was very lucky to live close to a national center for supercomputing. When I was eight or nine years old, I was running through Cray supercomputers, I learned to code very early. I saw the incredible promise of what technology, and even what we thought of as AI, could be. That’s defined my whole life: on one side, an optimism about all the things we could build if we did it for the right reasons, and on the other, a bit of impatience and frustration that we know this is possible, and we don’t quite do enough.

Very briefly: I became a computer scientist and a bioengineer, I started my own company, did that for a while, then I became a human rights lawyer, a totally different turn. It shows the two sides of my personality, I wanted to understand systems and power. Since then, my life has been a combination of building technology for good and supporting governments and civil society in learning to use it well. As you mentioned, I went to the Harvard Kennedy School, where I got to spend time in both reflection and active participation.

I was later given an incredible opportunity to lead the first new institution of the twenty-first century that combined philanthropy, the idea that we had capital to invest and give away, with a technically native awareness of what AI would mean for humanity. That’s what we’ve done. I now lead the world’s largest institution primarily focused on how we build AI for good. We’ve donated and invested over half a billion dollars. But maybe more importantly, we’ve helped shape the public narrative about what responsible AI should be: who should participate, who should benefit, and how we build it in a way that makes society better.

The McGovern Foundation’s Work

Anirudh Suri: Give us a couple of examples of projects you’ve funded or promoted through the McGovern Foundation.

Vilas Dhar: One of my favorite examples is called Climate TRACE. It’s the world’s largest and first integrated platform that collects data on carbon emissions, one of the primary drivers of climate change, and makes it available. I’ll spare you the technical details, but this involved building governance mechanisms for data sharing, technological frameworks, predictive analytics, an entire AI stack around one fundamental question: how do we ensure everybody is equipped to understand their impact on climate change?

It’s now used by thousands of organizations, nonprofits using it for climate advocacy and to model and predict climate effects, private sector firms trying to decarbonize their supply chains who can now ask foundational questions about where their goods come from and how to make net-zero changes at no added cost, fundamentally transforming the carbon profile of their operations, and large investment firms using it to change how they invest trillions of dollars over the next decade. This is AI for good. It’s not about the AI stack itself, it’s about how it’s used to empower better decisions about the future we’re creating.

Innovation vs. Diffusion

Anirudh Suri: I’m glad you shared that example, because one of the key debates going on in India for at least the last two years is this idea of whether we should prioritize AI innovation, especially at the frontier level, or focus on AI adoption and diffusion. At the summit and elsewhere, you’ll generally see India leaning more toward adoption and diffusion, but with some push now on the innovation front through new R&D budget allocations. Where do you stand on this dual approach? It’s not an India-specific question either, it applies to pretty much any country today. How do you think about that divide?

Vilas Dhar: We’ve advised more than eighty governments around the world on this exact question, Anirudh, and here’s where the answer has to land. On one side, I want to push back on any messaging that says India cannot be a primary participant in building frontier AI. Unfortunately, we’ve heard some tech CEOs come here and say almost exactly that. But I know the capacity of this country, I know what’s possible. There may be some barriers and obstacles that slow things down, but at the same time, India could be a frontier model builder. In fact, this week we’ve seen what that looks like with the launch of Sarvam. Sarvam AI, I’m sure your viewers know, is the first kind of open-stack, open-source AI model built here in India that does something really special: at a much smaller size, much lower power consumption, only about a billion parameters, it performs as well as some of the most powerful frontier models built in the West, because constraints lead to innovation.

But let me also talk about the real opportunity in the moment, which is the term “AI diffusion,” which I don’t love, it feels very passive. With targeted and direct investment and policy, like we’re seeing now in India, we can build such powerful parts of the overall AI stack that we’re actually talking about interdependence rather than overdependence. That is, India creating the public goods around application layers, around new and more power-efficient mechanisms of deploying AI at scale, ones that actually allow for participatory mechanisms, so that rather than being consumers of a global AI stack, India becomes a participant in building a significant part of it. I think this is an incredible opportunity, and I’ve met organizations and firms across the country this week that are really investing in this. From an international perspective, we’re ready for it, and cheering for India in this work. At the UN High Level Advisory Body, one conversation we always had, years ago, before we were really in the thick of it, was how do we ensure every country participates at the level of competency where they uniquely provide a voice and a wisdom? That’s what we’re seeing come to fruition here in India.

Finding a Niche

Anirudh Suri: I love this idea of interdependence, every country finding its own niche in the global AI ecosystem. But it’s easier said than done, because whether it’s countries in Asia, Africa, or really any country other than the US or China, I mentioned this in my own book too, we have these two tech superpowers competing on every layer of the tech stack, AI, bio, cyber, you name it, and they have the resources to do it. But every other country has to find its niche, for lack of resources, which can of course drive innovation. So, as you’ve seen countries around the world, including India, going through this intellectual churn at the summit, what about the rest of the world? What countries have you seen where this exercise of finding a niche, of creating interdependence, is working well?

Vilas Dhar: What a great, thoughtful question. Let me give you a macro answer and then a micro one.

The macro answer: first, let’s acknowledge the world is reshaping so dynamically that I fully expect India to operate as a great power within the next decade. I’m not sure the answer will be that India has to give up on competing across the whole stack, in fact I think the exact opposite is true, it may just take a little time to get there. In the meantime, what we’re seeing is a regional diversification of lived experience. For example, there’s work happening in Senegal that looks at restructuring local agricultural practices, significant investment in how they buy seeds, how they help farmers understand water cycles, how they harvest well. What they’re doing isn’t happening in isolation, they’re drawing on lessons from around the world, from work in Trinidad around sugarcane farming, where there’s a center of excellence for AI in agriculture, from places here in India, and from agricultural frameworks in the US. When they crack the problem, not just the technological solution but the full behavioral transformation, how do they package it up and export it in a way that teaches the rest of us what they’ve learned? This is the work that needs to happen, and one intervention we’ve supported is a global network of exchange centers at the United Nations level, centers of excellence that can exchange both practice and code, and provide both technological and behavioral frameworks.

At the micro level, let me give a bit of a different answer, and this brings us back to why we’re here at the summit. We can talk about the topics or what international competition looks like, but we also have to come back to where this is actually being originated. Let me give you a vignette. Last year, at the Paris AI Summit, I had the chance to launch a new 2.5 billion dollar initiative on building AI capacity in the Global South. I’ve been in India many times in the intervening year, and I’ve found my own source of pride in these young students in universities and colleges across the country. I’ve visited places like MICA in Ahmedabad, PSG in Coimbatore, Amrita University in Kerala. At these places, of course there’s a lecture and a discussion, but then I get to spend time with students who are figuring out what role they want to play in shaping our society, enabled by AI. The ideas I get are often not the ones I hear elsewhere, like “here’s a marketing solution using AI for advertising” or “here’s an AI solution that will make me a hundred million dollars.” It’s often, “my family faces this problem, and I’ve been thinking about how to build a device, an algorithm, or an application that helps solve it.” That, to me, is a source of hope for how we solve this question of what “competitiveness” should look like. If we can build a world where people know they have the agency and power to build AI solutions to solve their family’s problems, that gives me more hope and confidence about an AI-enabled world than anything I’ve seen come out of Silicon Valley.

Setting the Agenda

Anirudh Suri: I usually ask people who work across multiple countries in AI this question: are you worried that too much of our agenda for what AI should be used for is set in Silicon Valley, and then, let’s say, China? And if it is, how do you fix that? But in a way, what you’re saying is, don’t worry so much about that, let Silicon Valley and its equivalents elsewhere keep doing what they’re doing, but let other parts of the world take the lead in setting the agenda closer to the problems they’re actually facing.

Vilas Dhar: I wouldn’t go so far as to say don’t worry about it, I do worry about it, I feel it, it’s often my 4 a.m. worry, that we’ve over-depended on a very tight concentration of AI agenda setting. But let me say it a bit differently. It’s not just the agenda, it’s also the framework of decision-making and governance. So many frameworks have been created about how we govern AI, some from tech companies, some from governments. But governance is muscle memory, we have to actually learn the practices of governance, and that comes through collaboration on data sharing, algorithmic design, and other things we’ll do together, we have to do it together.

Anirudh Suri: This idea that today countries are competing, and companies are competing too, from your sector, civil society, as a voice, is your sense that people, voices like yours, can create a different nature for this competition? Because right now it seems like a competition much like, and I try to draw parallels to history to provoke a bit, a kind of neo-colonialism with AI: capture as many markets, capture as many use cases as you can, because if you don’t, a rival will. How do you see this nature of competition possibly changing? It doesn’t seem very healthy for the world necessarily, healthy competition is good, but maybe there need to be guardrails so it’s healthy competition for the good of the world, and not just competition for its own sake.

Vilas Dhar: You’re absolutely right that this competition dynamic is taking us away from what matters. When people talk about the “AI race,” look, I still have a childlike curiosity about the world. I was with my dear friend Shekhar Kapur earlier this week, and he said, “why do we ever grow up? Children ask all the right questions.” So I still ask the right questions: when you talk about a race, can anybody tell me what we’re racing toward? We don’t really know, right? It’s almost this idea of provoking an emotional response, saying we have to go faster and faster, but toward what, we don’t know.

So instead of talking about competition as the primary dynamic, why don’t we talk about leadership? Instead of competing in an AI race, why don’t we ask, what is our vision of what this world should look like, and who elaborates that vision? Because the great challenge we have right now is that we have a small set of visions, ones we’ve heard even at this summit, from the titans of tech companies, that are often about either how to make the most profitable outcome, or scaring you a little with existential fears, and we heard both of those this week. My voice, the voice of civil society, and the people we represent and share the stage with, very rarely starts with either of those things. It doesn’t start with profit or fear, it doesn’t start with hype or horror, it starts with meaningful and pragmatic hope. What does it look like to build a society with greater economic participation, educational access, that actually builds AI tools that let people live better lives? These aren’t tropes, they’re not passing slogans, they have to be turned into practical matters.

Anirudh Suri: Absolutely.

Vilas Dhar: And that narrative exists, absolutely, because I see it with hundreds of thousands of people around the world who are actively building it. Maybe what we should do is not ask “where is that conversation,” but say, “it exists, how do we bring it into the limelight?” And how do we properly bound the amount of power and resonance we give to tech CEOs to shape what billions of people around the world think about the story?

Anirudh Suri: I think too much of that narrative gets shaped, as you rightly say, too narrowly, and the other voices get drowned out. I think it’s very important that the kind of work you’re doing through your foundation, and individually, gets highlighted into the mainstream more. In a conversation I was having earlier with Jeetu, one of the things we ended on, amongst all this hype, let me share a couple of anecdotes from the last few days. So much hype around tech and AI at an AI summit, but my favorite moment of the last four days was last night at the Swiss Embassy, where the Swiss president hosted a dinner with classical music, an Indian and a European performer. The evening was magical, and I told the musicians afterward that nothing in the last four days had given me as much joy and peace as that half hour with them. I mentioned I went to a Quaker college in Philadelphia, where we started and ended every meeting with a minute of silence, to ground you, to center you, and I told them their music had done that for me that night.

Vilas Dhar: Beautiful.

Anirudh Suri: I mention that because I think tech, Silicon Valley, forgets that there’s a whole world of wisdom out there, from music to the arts to the common person solving a specific problem for their family, as you were saying. If that doesn’t get incorporated, we’re living in our own real ivory tower bubbles. I’m glad you pointed that out. The other thing I was discussing with Jeetu was how important values are in all of this as tech shapes our world. It’s very important that our values remain clear, and that we use that to shape the kind of “tech for good,” or “AI for good,” you were talking about. This has been great, Vilas, any last thoughts?

Vilas Dhar: Maybe just a quick closing comment. I think, as an accomplished podcaster, maybe you shouldn’t hold too tightly to the minute-of-silence lesson, I’m not sure your viewers would love it. But think about what kind of conversations we could have if we started from values and principles first, not just from what tech will do. Let me say one last thing, and I’ve so appreciated this time together. I appreciate the warm welcome in India, but I leave Delhi more hopeful than I arrived, and I think that’s a really powerful testament to what this meeting has been. I’m not so much hopeful about what frontier models will do, I’m not a techno-optimist who thinks they’ll fix everything, but I’m hopeful because I’ve seen, across this week, people stepping forward to build solutions, government ministers willing to build policy, technology companies here in India willing to build products that make the market better and more robust, and people like you who want to bring it to the world and have these deep conversations. Thank you so much.

Closing Recommendations

Anirudh Suri: Thank you so much, Vilas. We end every podcast with two quick questions: books you’d recommend our viewers and listeners read, and a couple of podcast guest recommendations.

Vilas Dhar: On the book side, I’ll give you the answer I always give: a book called Siddhartha, written in the mid-nineteenth century by a German philosopher, about our Buddha. It’s a beautiful book about a personal journey of realization, longing, and agency, and I can’t think of anything more powerful, as we think about technology, than starting within ourselves.

On podcasts, I hear there are some I should be listening to, that you host, that are pretty amazing, everybody’s told me about them. I’d also point to people staying at the forefront of things, like Ezra Klein in the United States, but I’m also hopeful about a growing international set of podcasts from places like Kazakhstan, Senegal, Chad, Saudi Arabia, and here in India, giving us global context. I’m always happy to share sources on LinkedIn as we go.

Anirudh Suri: Perfect, thanks so much for those recommendations. I’m going to pick up Siddhartha, and I know Ezra’s book, Abundance, is on my list to read soon too. So thank you for those recommendations, and lovely having you.

Vilas Dhar: What a delight, thank you.