Great music. Hello everyone. I have the best job in the world because I'm joined by three wonderful leaders who are trying to understand what they believe will change, what still holds us back, and who will benefit from this next era of discovery. It brings together three perspectives: Noubar on building new companies and scientific platforms, Lila on developing AI capabilities for science, and Kiran on biotech drug development and patient innovation. Thank you all very much for joining us. Noubar, let me start with you. What do you think, in your wildest dreams, AI will help us discover that we still can't imagine?
We can't imagine it, but you can. It's a difficult question because I think the answer is everything. In other words, I think the way we'll discover everything in the future will be fundamentally different and won't be limited by human imagination. In nature, nothing is discovered through human imagination. It simply exists. And now we're going to be able to do it at multiples of what humans have been able to do, and it's an exciting but also terrifying perspective.
Lila, do you think AI is accelerating science, or is it actually just radically changing it?
I think it's helping us understand the world in a way we never imagined, as Noubar said, very much like a microscope that can zoom in and go very deep, or a telescope that can zoom out. And I think perhaps the limitation is our ability to ask creative questions, to be bold enough and have the courage to ask.
What do you think the next AI breakthrough will be?
Well, it's definitely going to be very different from what we're doing today because right now biology is a very empirical science, and I think AI will allow us to do it very differently. It will allow us to collect and connect very disparate dots. That's why I'm thinking about developing and innovating new drugs that don't focus on single targets. You know, it's not about drugs aimed at a single target, but I truly believe that by combining the understanding of biology and the leveraging of AI, you'll be able to holistically analyze the networks, the signaling channels, and the neural circuits. And I believe that's how new drugs will be developed—multi-specific and multifaceted drugs. But Noubar, is there something that remains intrinsically human in scientific discovery?
Well, maybe I'll rephrase it this way. I know we're in an era where we're trying to restore life and existence to the human being because it's being challenged by this other form of life we've created—at least the thinking form. I think the way of being human is going to change in the scientific fields in the following way. I think today we're obsessed with artificial intelligence. We can call it superintelligence, AI, whatever, but I call it artificial intelligence because it's a form of artificial intelligence, just as it is. Then we have human intelligence, but I think what we haven't realized is that there's another form of intelligence around us, which are the forms of intelligence of nature. Every living cell, every cell that processes information, is also a form of intelligence. I think once we realize that, we'll realize that what has been biology—humans trying to understand the intelligence of nature—is now being complemented by humans trying to understand the intelligence of machines, and we're not going to be good at either one, in my opinion. And we're going to have to discover how to use machine intelligence to better understand natural intelligence, whether it's the immune system or viruses. I mean, people often argue with me about how I can say a virus is intelligent. Well, a virus that could wipe out most of the planet's immune system through a pandemic—if that's not intelligent, then our bodies must be pretty weak. The fact is, it's much more intelligent than the viruses in the computers that have us so impressed today. So I think humans are going to have to make room for understanding, but also to leverage machine intelligence to better understand natural intelligence, and why not potentially the other way around?
But does that mean there could be a discovery that we can't explain at some point?
There are already tons of discoveries we can't explain. It's called the modern pharmaceutical industry. We probably don't really understand how more than half of the drugs work. We can give a version of how they might work, but we actually don't understand the mechanism of most drugs. But through rigorous testing, through controlled experiments, we can demonstrate that they're more effective than harmful, and that they're regulated and approved. So I don't think humans understand. By the way, I don't understand how my iPhone does most of the things it does. It doesn't stop me from using it. I just don't think this human presumption that we should be afraid of everything we don't understand is a new phenomenon. We've been coexisting with things we don't understand for a long time. Lila, how do you see it?
In fact, if we look at DeepMind, I mean, a lot of this work was started by you and Demis. Yes. In fact, while Noubar was speaking, I was thinking about AlphaFold as an example of a great 50-year challenge and our advanced AI system, AlphaFold. Now we have it available to millions of researchers around the world, with a search as simple as on a map, to be able to search and understand how a protein is predicted, a three-dimensional structure, which is extraordinary because it tells us what the basic component of life is and what it could mean if it folds incorrectly. And we've been able to work with researchers, like in Spain, and take a drug and apply it to Chagas disease, an FDA-approved drug that affects 8 million people worldwide. Something that, as humans, we hadn't thought of, but asking the questions that helped us establish that correlation between the vaccines. So I think the opportunity here is not just in discovering new things, but in how we can apply them differently. In fact, there are 180,000 researchers right here in India working on the problems the country faces and advancing other types of work to understand diseases and the drugs that accompany them.
What are you seeing, Kiran? I mean, what does it mean for Biocon? So, do you expect your company to change in the next five years?
Yes, of course, I think drug innovation will accelerate, but I also expect affordability or affordable access to accelerate too, because you're shortening the timelines from lab to market. That said, I think even though right now we're all focused on the discovery process, which will definitely accelerate and shorten, the next challenge is how to shorten the most expensive part of pharmaceutical innovation, which is clinical trials. And that's where I think AI can play an important role in helping us think about what we want to extract and collect from clinical data that will allow us to predict whether they're safe and effective. And I think that's the kind of thing that can really shorten timelines and get drugs to patients faster. So for me, that's the excitement of AI, and that's what we're looking for as a company, saying: how do we approach AI not just to shorten, but also to keep affordability in mind for countries like India?
And I'd like to come back to the speed, but especially the affordability of it. Do you think AI will make drugs cheaper?
No, as I said, I think it's a matter of time and money. And if you can reduce that time, of course there will be a cost associated with using AI itself, but hopefully the overall benefit is that you can get drugs to patients more cheaply and quickly. That's my point.
Noubar, in that sense, you know, you can have a drug discovery that's faster, but when will you do it again? You know, will patient treatment be faster?
Yes, I think there are many components in which AI is helping and will help a lot. I think the hardest thing will be to do what's out there in terms of speculation these days, which is to simulate clinical trials and avoid them completely. I think that requires an understanding of human biology and physiology that I'm afraid we're quite far from. Well, I think we've digitally modeled many complex systems, but I think most people who don't work in biology and medicine don't understand how complex these systems are. I suppose our colleagues at DeepMind who worked on physics and biology problems learned what Schrödinger learned 80 years ago—that biology is not just physics. And as the scale increases, the complexity of the problem increases exponentially. In fact, the folding problem is probably one of the most deterministic parts of biology. Everything else, through interaction effects, becomes very complex. So it's a long way of saying that I think there will be short-term milestones. For example, which patient do you want to include in a trial? Can you train across their electronic medical records and find a subset that's more responsive—not just based on diagnosis, but also based on their molecular profiles? In the case of cancer, I'll give another example: there's great interest in screening tests to find cancer early. The problem is that those tests generate 10 times more false positives than true positives. What are you going to do with those people? If you include them in the diagnostic system, costs will skyrocket. Can you use AI to improve triage and greatly reduce the number of false positives? I think those are things that, as we start to address them, will begin to have an impact on clinical care, and over time we could imagine shorter trials and clearer endpoints. We're working on these things, but I don't think this notion of a digital twin of a human that we use for testing is something we should expect soon.
But what are you talking about? What timeframe are we talking about? Four or five years to solve the bottlenecks?
I think the impact is being felt now, and it's hard to say in what order, but what I was talking about is this massive disruption that's drastically shortening clinical trials. I wish it would happen—I've been doing this for 39 years. There's nothing that addresses the cost of new drugs like the complexity and time of clinical regulation, but I think it might be a step too far to present it to society and then disappoint people. I mean, there's already enough with AI causing disappointment and fear, so I think going out and saying we're going to cure cancer, as some of the experts in leading labs say, as if it were an imminent goal, is really hard to say and then realize that in five years you'll still be curing. I mean, we've been curing cancer for 70 years. The joke in the biotech industry is that if mice had a cancer problem, we would have solved it completely because we've cured every type of mouse cancer. Humans turn out to be radically different from mice in their physiology, so we have to be more humble. Don't let me say anything else. I would say that what AI has taught us in recent years, the new versions of AI, generative AI, is that we now know there are more things we don't know than we thought. In other words, as a fraction of what there is to know, we've now gone backwards because we now know it's accessible if only we could make measurements; if only... So I think the struggle between the unknown and the known will remain interesting for a while, and that means that, I think, as humans, we have to be quite humble in the way we do it.
Lila, how do you see it? Yes, that deserves applause. Why not?
I'll go back to what I said at the beginning. I think it brings us back to asking better questions. Asking questions we perhaps didn't have access to before. As Kiran mentioned, we can now take an interdisciplinary approach. Like in the past, where perhaps we zoomed in a lot to be able to observe patterns that might have been inaccessible. One of the things I've actually done, you know, I was in India last week, and one of the areas that has excited me a lot is learning about how we've been collaborating here locally, and part of the local innovation is in the agriculture area. Well, if I can shift for a moment to the human aspect, we're working with an organization called ICRISAT, based in Hyderabad, and one of the biggest challenges we have is that sorghum, which 500 million people depend on as a food source or to feed livestock, has only been studied in the West—the Western varieties, not the native ones from here. So one of the challenges India faces is how to really accelerate knowledge about crop genetics and how to improve them. Well, now we're working with crop breeders to study sorghum, or jowar, as I learned this week, to highlight, to be able to map all the genes and make them freely available so that innovation can emerge from there. Now, as my colleagues mentioned, there's a lot of work to be done after that. But if we can achieve the breakthrough and get the inputs from the crop breeders, the inputs from the farmers upfront, and accelerate decades of work and start processing it all, imagine what this could mean for a more drought-resistant crop, something that can better withstand diseases. And this is something quite personal for me because my mother researched sorghum in Africa over 20 years ago and won the World Food Prize for one of the first advances in drought-resistant sorghum. But India and Africa still face many challenges, and it's really amazing to see now what might be possible with some of the opportunities to apply AI in areas that have traditionally been underrepresented and understudied.
And I imagine the adoption speed in that space is going to be much faster than anything else in...
