It wasn't about building another feature or this quarter's earnings report. It was about the race to do the right thing.
Chief executives and founders themselves are coming out and, you know, calling out the risks and asking for greater regulation. But at the same time, companies are also pressing ahead with faster and more stronger, you know, models. Do you see a contradiction there?
Responsibility isn't something you bolt on at the end. We have to build it from the start. That's where we think AI can be a tool, a teaching assistant for the teacher and a tutor for the student. We must build AI responsibly and we must think about the impact to ensure that AI happens for and with communities and not to them.
Today we are talking to Lila Ibrahim, who spent eight years as the chief operating officer of Google DeepMind. Earlier this year she took on the role of being the chief AI readiness officer. In 2023, she signed a paper which put risk to humanity from AI alongside risk from pandemics and nuclear wars. Three years on, AI has moved at a very, very rapid pace. I ask her what does she think of the risk to humanity from AI and what is the responsibility that leaders like Google have in this area. We also talk about the initiatives that Google is having in areas like education and agriculture through initiatives like the National Partnership on AI that it signed earlier this year. We also talked to her about what is the responsibility that companies and countries like the US and China have, which control a lot of AI frontier models as well as compute, and the role that countries like India can play when it comes to having an edge in this whole AI race. We also talked to her about her own journey at Google DeepMind and now as the chief AI readiness officer of Google.
Lila, thank you so much for speaking to The Economic Times. Let me begin this conversation by asking you: back in 2023, you wrote, signed a paper which put the risk from AI alongside nuclear wars and pandemics. Three years on, AI has really moved very, very rapidly. How do you feel about that?
Well, you know, part of the reason I signed that statement was, as an engineer, I think about things very practically, and I feel like it was the statement around: should we be treating this with the same care and attention? And the answer is absolutely yes. We have to be responsible stewards of such a powerful technology. AI has the possibility to enable us to find solutions to problems our world is facing. But we can only do that if we're thinking both about the risks and how to mitigate those risks, and the opportunities and investing in those opportunities. And what I really wanted to do was to say we need to be having these conversations out in the open, and we need to not be about any one company or one country. We have to come together to bring diverse perspectives into this conversation. And there is no time like the present. And I'm quite happy to see that around the world we're openly having conversations about what might go wrong and how do we prevent that. We have to build that in from the start, also what could go right. And I think those two in tandem are really critical conversations.
Right. Like you said, it's increasingly we're having those conversations. And if you just rewind the past couple of weeks, then a lot of many companies and researchers, in fact, chief executives and founders themselves are coming out and, you know, calling out the risks and asking for greater regulation. But at the same time, companies are also pressing ahead with faster and more stronger, you know, models. Do you see a contradiction there?
There's two things that we need to be thinking about. One is, as I mentioned, AI has a possibility, and we're already seeing the impact of being able to identify solutions to some of our challenges in the world. That's the race. The race is to get this right. And so we, and at the same time, we have to think about how do we work with governments? How do we work with organizations and with communities to ensure we're having these deliberations? And that's really what I think is, again, so important and part of the reason I shifted out of my chief operating role into this chief AI readiness role, because if we're going to have an impact on AI to unlock these potentials, we need to be working for and with communities. And I've been so excited to see how much is happening here in India and around the world. And I think we need to bring those opportunities and those risks into the conversations as we're developing. You know, responsibility isn't something you bolt on at the end. We have to build it from the start, and we can only do that if we're out talking with communities, as I've been doing this week in India.
Right. Talk to us about your week in India. You've moved from a role as the chief operating officer of Google DeepMind to as the chief AI readiness officer, which is probably the first of its kind designation ever. So talk to us about, you know, what is the charter and what is the idea behind it, and your week in India so far.
Yeah. So first, let me begin maybe by taking you through my journey at DeepMind. I started nearly nine years ago as chief operating officer. I was the first one at DeepMind, very much an AI research lab. So my first focus was really how do we think about both building the organization as well as how do we think about building AI responsibly. It's been an extraordinary journey over this time. You know, everything from AI capturing the imagination of the public worldwide to a Nobel Prize recognition of some of the scientific breakthroughs we've had, and also thinking about how do we ready our own organization for this emerging technology. And last year, Demis Hassabis, the CEO and founder, and I were having conversations about where do we see AI going and what actually is really exciting and where I feel like I can add my perspective into this moment in time in AI's development. And you know, after 35 years in technology, of this through line of always thinking about the technology development and the impact in society, I really wanted to take this opportunity to work more with the communities in which AI is — the communities that are working with AI. How do we get their inputs more into our development process? They're already there, but if we're not — we don't want to widen the gap, we want to bring those voices in now. Now is the time to be making the impact. So I moved into this newly created role, as you said, I think I'm the first ever, which means I get to define what this means. So of course you can't talk about readiness without coming to India. The diversity of the country, both geographically, culturally, is very interesting, and it's a very innovation-forward approach. So I've been talking with, you know, creators. I met with a lot of creators. I've been meeting with students. I've been talking about the impact of AI in agriculture, our work across the different languages. And what I found is there's something really incredible happening here, which is people are thinking about who are the real people with the real problems and how can AI help. So it's not AI as a tool looking for a solution, but rather what are the challenges and opportunities, and then where and how can AI be a part of that solution. I think it's incredible to see the progress here in the country, and also what learnings do we have that we might be able to take externally, and what are the technologies being developed here that we can also take globally.
Right. Let me ask you, since you spoke about, you know, India being very diverse, of course, culturally as well as as far as languages are concerned, India is also a very difficult country, you know, when it comes to some of these solutions because there are just so many languages, dialects and things like that. How do we sort of create a model? How do we get a model to work, say maybe in an area like agriculture, which works for people, for farmers at the grassroots level and not just great in a lab?
Yeah, it's a great question and I think about this a lot. And actually this is one of the areas that really excites me because I actually have some farmland as well. I grew up in a farm community in Indiana in the United States. And so one of the things that we did is it's a very common approach to what we take with all of our technologies is we have to understand what the challenges are. So we've been talking with farmers and crop breeders like once every two weeks to try to have the conversations of what are the challenges that they have and where might AI be useful and where won't it be useful. And one of the — we've identified a few things. One is our program called Agrifund, which is actually working to map out the different farmlands that can provide guidance on device to farmers that have to deal with when's the right time to plant a crop, how to deal with disease, how to think about unlocking economic opportunities. But again, that's kind of hand-in-hand. And one of the key underlying factors of that is language accessibility. Given again the diversity of the community here across India. So we have a research team based in India at Google DeepMind that's been working at different dialects and different languages across the region. And so we've been able to take those learnings of bringing new languages into the applications such that those learnings can now be applied to other countries and other languages and dialects around the world.
In addition to the work with farmers around Agrifund, which has now reached 600,000 farmers in India and has been exported to 11 other countries based on the technology developed here, we're also very excited to announce a partnership with ICARISAT, which is headquartered in Hyderabad, and we're working on ensuring that some of the orphaned crops such as sorghum, which feeds jowar — is that the correct name? — which, you know, helps one of the millets. Yeah. One of the millets. The ultimate of millets, right? Providing food for 500 million people, also could be used for biofuels or to help feed livestock. Completely under-researched crop. And especially the ones that are indigenous here in India and in Africa and the Global South. So we're working with ICARISAT to do a gene mapping that will be open source, much like the language work we're doing, so that the innovation can happen on top of that, where we can start thinking about how do we make disease-resistant, drought-resistant crops that can survive the global warming and the climate issues that people here and the farmers here are dealing with. I think it's a great example of working with the community for the community and thinking about how we also might be able to take those learnings and that technology elsewhere to help advance so that AI can continue to create solutions for real people.
Right. So I think that in some ways continuing the Google playbook of innovating in India and taking and replicating solutions worldwide.
It's a great blueprint. It's a great blueprint. And I think the collaboration of the local community has been really fantastic for that opportunity.
Right. Could you talk about some of the solutions where you feel that, you know, some of the research or some of the development that is happening in India to solve some of these very local grassroots solutions could be taken outside, you know, outside of India and to other countries where they could, you know, solve, say, population-scale problems?
Yeah. The Agrifund work with the farmers that has reached 600,000, and we've now taken that to nearly a dozen countries across Asia. The work with languages of being able to take adding a new language, which used to take months, now down to weeks and then like days. So I think that's quite significant of taking that globally. We were just meeting with Atal Labs as well and thinking about the tinkering labs, and there's so much innovative work of how do you keep the teacher at the center. I saw a demo this morning. I'm going to visit a school. I think there's a really interesting opportunity there of how to take this tinkering model to inspire students to innovate and be part of creating solutions and entrepreneurial work to think about what that might look like at a broader scale beyond India. There's no shortage of opportunity. I mean, the innovation here is so inspiring. I was fascinated when I was talking to the students the other day. I was meeting with some of the university students in three different states within India, and they all had different perspectives on how AI was supporting their studies, how it was supporting their lives and their livelihoods. And I think this is really what's so exciting about what I'm seeing here and the people I'm talking with, especially the AI natives. They really feel for the first time they have a technology that can be personalized to them, that helps them bring their curiosity and their creativity and the reality into what they're trying to build for their life.
