[0:00:04] Good evening, everyone. And welcome to this WCIT, the virtual company of information technology lecture in memory of past master Barney Gibbons. Firstly, a couple of thanks. Thank you to HSBC Innovation Banking for being such great supporters of this lecture. And this year, we have two extraordinary figures who have helped shape
[0:00:35] The Digital Age. Liveriman. Sir Demis Hassabis. And Liveriman Dame Wendy Hall. Our new members, you are both very welcome. Thank you. So I was thinking about how I start this. And every video of both Demis and Wendy have their whole CD read out.
[0:01:05] And I asked Mr. Gemini, what's the best definition for Demis? So, Demis, if anything is amiss here, then I would... I'm not to blame. So, Demis Hassidis, co-founder and CEO of Google DeepMind, and in my opinion, one of the world's leading research labs.
[0:01:30] I'm particularly excited as well that it's in this country. And your mission is to advance AI and to apply it to major scientific and social challenges. And he built a video game called Theme Park, which combined AI with economics, and you sold 15 million copies. And that's before you even went to school.
[0:01:57] After your university, which I'm not going to regale because it's a very long list of things you did, you built AlphaGo and defeated the world champion Go player and AlphaFold, which revolutionized protein structure prediction. You are multi-domain. You combine being a computer scientist, a neuroscientist, an entrepreneur, a former video game designer, and you bring insights.
[0:02:25] Multiple fields in AI research, and we are very pleased to have you here, Sir Demis. Dean Wendy Hall, one of Britain's most distinguished computer scientists, a global pioneer in web science, cross-disciplinary study on how web shapes society and vice versa. Regis, professor of computer science at the University of Southampton, and one of the first researchers to undertake major work
[0:02:57] in multimedia, hypermedia systems. A founding leader of the Web Science Trust, along with Sir Tim Berners-Lee, who is actually an honorary member here too. Yes, we both know him well. To advance the understanding of the World Wide Web and its social impact. The first woman to become a professor of engineering in Southampton and a former president of the BCS.
[0:03:24] And the ACM, which is Association of Computer Machinery. You are an internationally respected voice in AI, Internet governance, and data policy and digital ethics. And you are helping shape governments and institutions to think about how technology impacts society. Dean Wendy, you are very welcome. Thank you very much. Which leads me to my theme in my year.
[0:03:52] of the worship of the company, which is community. And if I say a couple of words on that, I chose the theme because I was trying to understand how AI and technology are going to impact our communities, our people, our industry, which touches the four pillars of this organization, which are charity, industry, education, and fellowship.
[0:04:20] And so today, I'm going to pose my first question on skills and AI literacy to you, Damwendi. Nearly a decade after your 2017 review, what does a serious national AI strategy look like now, from schools to university to mid-careers to reskilling our people? And I was just wondering, is the UK still missing or misreading
[0:04:50] This Skills Challenge. Oh, that's three questions in one. I can't believe it's 10 years. Demis was there when we did that. So I wanted to co-chair it with him, but he was already too busy. Do you remember? Yes, I remember. We tried to input into it. You did, you did. Right, but he wouldn't co-chair it. But anyway, that's fair enough. He was building deep mind.
[0:05:17] He's living it, living the dream. So, unfortunately, my answers aren't very positive. The UK did very well. In fact, we were...
[0:05:29] I like to say we were the first to develop a national strategy. In fact, Canada did beat us to it, but I think ours is better known. The whole presenting report was 2017, and then the national strategy was developed by what we set up out of that, the Office for AI and the AI Council, and the national strategy evolved from that in 21 during the COVID years. I remember all the calls.
[0:05:56] And I was AI skills champion for a while. There is no AI skills champion now. There's no AI council. There is no independent body overseeing what the 400 people in DCIT do in the AI unit. And I think the UK loses a trick in that. Many, many countries, and I traveled the world meeting governments around the world.
[0:06:24] They picked up on our strategy. Many copied it. You can see elements of it, words of it in other countries' strategies. Singapore, for example, followed our playbook and did it better. So they're now overtaking us in the league of...
[0:06:42] Developing AI for a tiny island, they bat hugely above their weight. And they're overtaking us in the league tables of, you know, it's very clear it's US and China. We were then third in the tables of AI strategies and adoption. We've slipped. And Singapore took...
[0:07:07] Really have done it incredibly well. And other countries are doing it too. Ours now, we do well, and we can talk about later the things we are doing well. It will come up later in the discussion. But the skills piece is very, very patchy. Unfortunately, several things came out of the whole percentage review. All those recommendations were taken forward. MSCs, PhDs, very little for schools.
[0:07:35] There's very little guidance for schools on AI. You see things coming out in quite a fragmented way. I don't think there's any oversight of AI skills in this country at the moment. Sorry to say that, but it's true. How about you, Dennis? What's your view on that?
[0:07:59] I'd probably suggest something much more radical. So I would agree. I agree with Wendy. We had an interesting head start in 2017. But, you know, 10 years in AI is like a century in most industries, right? Or maybe two, you know, maybe a millennia. So a lot has changed. I think we did a good job. I think I remember that actually we sponsored some of these masters. You did, absolutely. And I think that helped a lot actually we see today, London and...
[0:08:25] The UK is doing really well in things like startups. And I think we try to anchor that with DeepMind and brought a lot of talent into the country. And then now 10 years later, a lot of those people are now spinning out other things. And this is great to see. We need to do more of that. I'm sure we'll come to that discussion later. But specifically skilling in schools, I absolutely agree with Wendy. Now we need to look to the secondary school level, maybe even the primary school level.
[0:08:53] The horse is bolted in the sense they're all using the LLMs and chatbots in their spare time and for homework and things like that, but it's very unsystematic.
[0:09:02] And it's a bit like, I feel like government policies and education systems are a bit like, oh, I wish that would go away. But it's not. We're going to have to lean into it and help teachers and schools and parents and students work out, or how do we get the benefits out of it, not just probably the way that people are using it now and kids use it now. And I see my kids, you know, we'll just help them almost cheat on their homework or shortcut things. And we know that that will result in deficit of skills.
[0:09:29] Traditional skills, let alone AI and computer skills later on. But I think we've got an opportunity here. Maybe it's time for a new review for you to write, Wendy. If I was going to input into, this is what I would suggest, is we need to invert the classroom, in my opinion. So I would lean into, this is what I would do if I was in charge of education. I would say school, in school, should be about people skills.
[0:09:53] Projects, entrepreneurialism, creativity, judgment, all the things that you might expect the skills to be important in five, ten years time.
[0:10:02] When AGI has arrived, and I know we're going to debate about AGI, but even just extrapolation of today's AI, there'll be things AI can do well, and then there'll be things that will meet human judgment, taste, all of these things. But these are not taught in schools today. And no teacher training. So I'll come to that. So I think in class, it should be about human-human contact. How do you organize groups of people into projects? Like a Montessori school, but plus-plus.
[0:10:28] Or maybe you get at like Oxbridge and places like these sort of the top universities, these kind of individual supervisions. But you bring that into the classroom. Out of classroom is where you do the rote learning. That would be my opinion, aided by personalized AI that basically tailors the rote learning to the speed and the exact personalized to the exact speed of the student and what they need to do. It's crazy that we do a one-size-fits-all in class anyway on rote learning. It doesn't work.
[0:10:57] Because you have to cater to the average. What about the kids that are behind or the kids that are ahead? Good luck on both sides. So here with AI, you can actually personalize it so everyone gets exactly the tailored curriculum that they need. And then in school, it's all about projects. I think teachers would love that personally. And then you could have world-class teachers that are...
[0:11:17] Almost like superstars that do the lecturing and they're really good. Why do we expect every school teacher, every lecturer to be amazing at lecturing the same course? That's just a completely redundant amount of work. When I talk to teacher friends of mine, each of them has to teach GCSE maths and then create their own slide decks. Why are we in that world? It should just be one. And then the rote learning you do at your own speed at home with the help of AI.
[0:11:43] That's what I would do. But I don't know if the countries... Somewhere like Singapore will do this, in my opinion, or UAE. Can I riff off of that a minute? Because I agree with almost everything you've said there. I think the rote learning doesn't have to be done at home. We have a...
[0:12:00] There's no equality around this. A lot of kids can't do this at home, right? And we have to be... I definitely agree that we can use AI to stop the one-size-fits-all. One teacher, 30 pupils, they're all different. Teachers can teach much better if the children are receptive to what they're going to say and have already done the basic stuff, whether it's home or school or wherever. And I've always been...
[0:12:27] My first, I don't know if I've ever told you this, my first piece of AI code I wrote in 1984 to build an intelligent tutoring system on a BBC micro in prologue. You'll understand what that means. Not even a concept, just an idea that I got from my tutor at university. But I've always believed passionately that we can use computing and now AI to develop personalized...
[0:12:55] Tutoring systems for kids that will help everybody get to the same sort of level. But that takes a huge commitment from government to make that happen. And I'd love to help do that. And then taking one step forward, going to university, is it now become pretty redundant that we have a computer science degree? No, not at all.
[0:13:23] Of course not. Exactly. I mean, you absolutely need to lean into STEM and computer science because no matter, it's just a higher level programming language is the way you can think about what programming is going to become. You know, we used to program in machine code, then there was C, then there's Python, and now it's going to be, perhaps it's going to be English. But that's fine. Like, you're still going to need to know about architecting things and software, best software engineering practices.
[0:13:49] People who understand the deep technical, they'll be able to use these tools 10 times more effectively than people that don't have that technical knowledge. So absolutely, we need people to apply. But we also need AI across the board. So things we're doing at Southampton and no other university is similar. We're actually developing something called an AI driving license for all students who arrive in the university next. This is a good plug for Southampton.
[0:14:14] If you arrive in the university in October, this will be available to them and staff. Think of it like your highway code. This is for non-science students, really, to get them up to speed. Some of them will already be doing it, of course, with their own, but to get them up to speed how they can use AI. We have a policy on how to use AI in education, and we have also set ourselves a target that by next academic year, every discipline, every degree in the university, there'll be a module on AI.
[0:14:43] Customise to that, whether it's history, psychology, geography, physics, whatever. Customise. That's a huge task, by the way, but that's what we've set ourselves. And presumably you're going to have to tweak the curriculum as well. I was just going to say, I think you need to teach STEM and computer science and continue. That's very valuable training, but the curriculum has to change.
[0:15:06] There might be very different things that one should study. Absolutely. I don't know if you heard Mike Waldridge talk recently about the way it's going to change the computer science. AI taught as a part of computer science. I mean, we do an AI degree. A lot of universities are already. We've taught computer science with AI for a long time, but we now have computer science and we have AI. And I'm passionate about getting...
[0:15:34] Getting information about AI to non-science students as well. And particularly the risks and the other social effects of doing this, that will also have to be brought into the curriculum, presumably. Yep. Oh, sorry. And that's across the board as well. I think that was actually even part of...
[0:15:58] If I remember correctly, your original report of ethics being taught and challenges alongside the technical aspects of computer science. I think all computer scientists need to be trained in that. And by the way, I also believe that the time is now for humanities, like philosophy, economics. I think we really need... In the world we're about to enter into, post-AGI and these things, that's a new era, in my opinion. So I feel like...
[0:16:25] That type of thinking is going to be really needed. I was wondering how long it would take you, Dennis, to actually mention AGI. No, he's gone one further. He's mentioned post-AGI. I was just wondering if for the uninitiated audience here and listening, I just wanted to give us just a feel of how things will change with AGI.
[0:16:53] And the benefits as you see it, and also the time scales. I know you say 2030 and onwards, but a little bit of color on that would be very good. Yeah. I mean, firstly, my definition of AGI is a system that we can test and exhibits all the cognitive capabilities humans have. The reason that's important is the human reference point and the human brain reference point is the human brain is the only existence proof we have.
[0:17:19] We know of, perhaps in the universe, that exhibits general intelligence. And I believe that brain is a sort of an approximate Turing machine. And obviously Turing, one of my all-time heroes, he showed what universal Turing machines can compute anything. So that's computable. Maybe apart from quantum systems and other things. It's out of scope for today. But I think...
[0:17:39] I think that that's what an AGI system is, and that means it should be able to do the types of creative things and invention things that we know the best scientists were able to do. Today's systems are nowhere near that. They're not consistent enough. They have missing capabilities. They're very impressive in many ways, but they also, as we've all interacted with them every day, these LLMs, they're missing some components. Now, that may get solved with scaling, but it may also require one or two big breakthroughs still.
[0:18:09] So the jury's out on that. I think it's going to be given the amount of resources and brainpower going into it, you know, around four or five years is my expectation from where we are today, so around 2030. And then when that happens, you know, that becomes a new type of tool or system that we've never seen before and we'll need to
[0:18:38] Think about how to use such a transformative piece of technology. And it's going to change everything, in my opinion, economies, perhaps even the human condition. But of course, the positive part of that would be advancing science, curing diseases, advancing medicine. This is why I spent my whole career working on ARN, what we call AGI technologies. We helped popularize that term.
[0:19:03] But obviously there are challenges with systems and we already see the beginnings of that with the Gen Tech systems that become more and more autonomous. How do we make sure they stay on guardrails that we set and they do the things that we want? Then there's questions of things like, beyond that, of as the machines get more capable, consciousness, other things, which maybe we'll get into. But the first step is to get these tools right.
[0:19:28] And then, Reni, you have an interesting... Well, I don't agree with Demis on all this. I respect what he does, right? And if everybody was like Demis, we'd be all right. But they aren't, okay? Yeah, they aren't, Demis, right? So there are people that are just in this for the money or the power and will release anything and say it does everything.
[0:19:53] Without any testing, any evaluation, any safety checks, any guidelines. This is happening in the States at the moment. I loved that, ironically loved that you did a...
[0:20:04] Debate with Dario at Davos. I call it the Demis and Dario at Davos debate. Okay, the four Ds. And I use this with my students because it was Demis saying the things he's just said here. And Dario, who was out there to get the money for his company and is now going IPO, saying, we're going to reach AGI by the end of this year. And you were trying to pull him back in that, right? Because, well, actually, there was a bit...
[0:20:30] Let me roll back. The big thing for me, if, and I don't accept it, but if we reach some form of something you accept as AGI by 2030, which I personally am not on that side of the fence, but if, then that gives us four years to sort out global governance of AI. Because otherwise, we've got machines that are...
[0:20:54] Cleverer than us in some aspects. And the big thing is that if the software we're building starts to decide to do things on its own, the sort of, I'm sorry, Dave, I can't do that moment, right? And this will inevitably happen. You know, Stephen Hawking in 2014 said, if we build machines that are cleverer, I'm not using the word intelligence, by the way, here.
[0:21:23] I'll just say one more thing about that. He said that was the end of the human race. That's one of the last things Stephen Hawking said in a BBC interview. There are threats. We absolutely have to have a global governance in place. I mean, I was on the UN High-Level Advisory Board. It won't be... Not enough people... The UN isn't what it used to be.
[0:21:47] And it won't be of the UN, but it could spin out of the UN. There's the global dialogue starting in July where some governments are going to come together and start really talking about what this means. And just this week, Trump went back on his word and said, we have got to start.
[0:22:06] What did he say? Not regulate. What's the word he used? The companies and the check, evaluate the software that's coming out. I can't remember the word he used. Because when he came into power, he did a, what did the president say? Executive order. He overturned Biden's rule and he did an executive order that said no regulation at all. And he stopped the states.
[0:22:31] Regulating. Well, he went back on that this week, or his government did. And actually it was because I think the tipping point was the Mythos software from Anthropic. Yeah, it's also not just that. There's a lot of people inputting. We're spending a lot of our time inputting into those things behind the scenes. We don't shout about that from the rooftops. Mythos has been a very interesting, good trigger because it worried the banks.
[0:22:58] And Besson, which is always the beginning, you know, that's when people take things seriously, right? They're talking to HSBC people here. You know, when the banks get worried, then government wakes up, which is, you know, maybe it's a good thing. And that means there's a window of opportunity here, I think, to do something. And I think that's why we saw the executive order eventually get through. But it's not enough. And there's more. And actually...
[0:23:23] We're in, I mean, as you know, Wendy, we're in total agreement. I do think there needs to be kind of an international standards. You know, my view is, and maybe where we disagree is timelines on AGI. But my point is, is that there's a, I actually talk about it as a 50% chance of 2030 with some obviously margins of error. But the point is, it could happen depending on things like reincursive self-improvement, if that works. So there's a lot of things, people are working a lot of things that could speed up the timelines, right? If those things don't work, then it might take...
[0:23:52] But that's great.
[0:23:54] Then we'll just have more time. But my view is we don't have the right institutions fit for purpose to deal with something as consequential as this. Unfortunately, this is happening at a moment where the geopolitics is as complex as it's been in 30 years. It's more fragmented than ever. The UN is not what it was, as Wendy was saying. So maybe one could imagine, in fact, I did imagine 20 years ago, maybe it would be some form of technical subcommittee of the UN or something. That's unthinkable now.
[0:24:23] So I feel like there's not a lot of time, even four years, let's say it's 10 years, that would be better in some ways if Wendy's right, if it takes a bit longer because then we have more time to get ready and society to get ready. But look, we got to take it seriously is what I'm explaining to people who care about this, that it might be sooner.
[0:24:42] So that's the point I'm making is the urgency. It's absolutely urgent now. We've just seen mythos, cyber. That's just the beginning. I'm more worried about things like bio risk coming down the line. Obviously, this is something that I work very closely on to try and cure as many diseases as possible with AlphaFold. But those kinds of tools in the wrong hands can also be misused. And so there are two things I worry about. Misuse of general purpose technologies by...
[0:25:08] Rogue actors for harmful ends. That can happen. Not what those tools were intended for. That even affects things like open source, which is generally good. But how do you restrict bad use cases of that? That's a big problem we don't know how to solve. And then there's the technical question that Wendy alluded to, is making sure the systems, as they get more and more autonomous, more powerful, more smart, how do we make sure that they stick to the guardrails we set them? And then even if we solve those technical questions, then there's the economics questions. How does all of society...
[0:25:38] Broadly benefit, not just a few handful of companies, a handful of people. That's not a good way for society to go. And then beyond that, even if we get that right, there's the philosophical question about meaning and purpose. So there's a whole stack of issues, each one more complex, in my opinion, and we've got to get going yesterday on these things. So I've started naming it with others. We're going to create an international AI agency. Let's just call it something.
[0:26:08] And it can spin out of the work the UN doing. It won't be of the UN. It needs to be independent. And we need to get... I say we, I mean the people that care about this. There's no particular organisation or group of people. And it's got to involve the tech companies, it's got to involve governments, and it's got to involve civil society. But it's got to be something that the companies and governments can sign up to. Someone's got to fund it too. But that's like...
[0:26:38] You know, we cannot just sit back and say someone else is going to do it. We have to. So I'll be in Geneva in July, you go? I'm not sure. Oh, Sam will know. Yeah, exactly. There's the first dialogue coming up.
[0:26:57] Global dialogue about this is happening in Geneva. That is being run by the UN, but it doesn't mean what comes of it has to be of the UN. I mean, that's why, by the way, the AI summits that were started in the UK with Sunak's government, and we were highly supportive of that and helping push that. I think those summits have been good. They're not enough, but at least they're getting the international community together to try and talk about some of these subjects. But nothing came out of India, Demis. Sure.
[0:27:26] There are 250,000 people at a conference. How long have people been trying to do climate change COP meetings? We can't go that route. There has to be a different route. Talking shops are not going to be enough. And I just don't think the UN has the power to do anything currently. So it's going to need to be new institutions.
[0:27:49] That's a tricky needle to thread. A lot of us are working, and I'm working, trying to work behind the scenes on this. I appreciate Wendy's working front and center on these things. I think both are needed, actually. Behind the scenes advice and maneuvering and then head-on tackling of this. We haven't mentioned China either. We will come to China soon.
[0:28:11] But in terms of the big tech companies actually taking part in this, is it not like Turkey is voting for Christmas, though? No, of course not. In fact, they want it, actually. That's the thing, actually. I mean, I know all the lab leaders. They're all characters, let's put it that way. But nobody wants catastrophic things to happen. That's no good for any companies. It's no good for any individuals, any reputations, or for business. So it's sort of...
[0:28:38] Some of them are genuinely worried and scientifically worried, and they always have been. I know their motivations from a long time back. People like Dario, we used to chat about these things when we were both postdocs. Others have come lately, and maybe they're more interested in power and money, but it doesn't matter what you're interested in. You don't want catastrophic things happening, and nor does government, even the most capitalist, shall we say, or market-driven governments.
[0:29:02] They don't want that. Or authoritarian, actually, frankly. So there is hope here because actually it's in everyone's interest to come to some sort of agreement. The problem is the usual tragedy of the commons, prisoner's dilemma issues, which is why we need some kind of governance and government involvement because the companies, it's always in someone's interest to make money by defecting from some agreed protocol.
[0:29:24] That's the problem. And individually, I think all of the good news is I would say all of the researchers and lab leaders and company leaders I talk to in this space do want something to happen and do want some guardrails and some standards. The question is what those are, how that should operate, and who should run that. But me and others, we're making proposals behind the scenes to some of the main governments.
[0:29:52] That's great to hear. Thank you. When you said the big tech companies, you meant the ones on the west coast of America. I did. So don't you care about the big tech companies in China and what they're doing? We do. China would love to run whatever agency we might set up.
[0:30:14] They would do it because their support, you know, they are, as Demis said, they are as keen not have a catastrophe in AI for many different reasons and the same reasons as the West is. But actually, they would love to run whatever we're going to set up. And the other thing about China, I'm sorry, I'm going into it because I don't think you can avoid it, is they largely produce their models, you know, they're producing open code and they're much more open. And there's good and bad about that.
[0:30:44] But they are much more welcoming of these ideas than, say, the US government is. Okay, since we're talking about China, I was just wondering whether a lot of the investment and a lot of the stuff that's happening in AI is in the US and China. And are we going to just be, as a country,
[0:31:12] Are our companies just going to be left behind and be watching and be users of the technology rather than, like what you've done, Dennis, is being an inventor of such technologies? Well, we can't be. I mean, I think it's important that, and I've told this to all various European leaders, that you can't just be, I told this to Macron back in around 2017, and he tried to do something about it. You cannot be world champions at regulation only.
[0:31:39] That's not a thing. Exactly. So to have a seat, I literally told him that. And I said, you've got to also be leading in the technology and have some leading companies and leading academics so that you can have a seat at the table. So first of all, you see what's coming down the line.
[0:31:56] and you also have a sort of moral authority to have a seat around the table and be listened to. And there was a time when I was trying to encourage a sort of trilateral thing with Britain, France and Canada because I would say those three countries were kind of
[0:32:14] The next best at research after the two superpowers, US and China. So actually as a three, if you look at citations, papers, researchers, it's pretty powerful. Like those three countries together now, you could include some other European countries and also there's Singapore and so on. But those three could create a kind of moderate bulwark between the two superpowers. Didn't quite happen for politics and other reasons, but it's still an opportunity there. I think the UK is still uniquely positioned to take advantage of this. We've tried to do our bit.
[0:32:43] Obviously, we're part of Google, but I kept DeepMind headquartered here. We brought thousands of top researchers here. They did some great work at DeepMind. Many of them then, after five, six, ten years, went and started other things. And once you embed yourself in a location, you tend to stay there, right? Family reasons, other things, you'd like the city, and so on. And I think we're seeing that. And all ecosystems need anchors like that. Silicon Valley was not Silicon Valley until Hewlett-Packard and Intel, all these kind of companies seeded it.
[0:33:13] But they also came from strong universities nearby, Stanford, Berkeley, and so on. And we have that here, the Golden Triangle, Cambridge, Oxford, Hampton, UCF. We've got so many strong... These things take centuries to build up. We have those assets. We have very smart people. We just need to have more ambition and utilise them here, and hopefully government encourage that, or at least not get in the way of that. And I think we're doing pretty well here, but we could do better. And then the next stage is...
[0:33:41] is financing the growth stage of companies, which is missing, which I think maybe some of the financiers in the room can think, how would we do that? I think the stock exchange needs some updating. I don't know why companies don't float on the London Stock Exchange anymore. They will go to the US. There's obviously some problem there that needs to be addressed. And I just think we could take advantage of this moment. We have all the right ingredients. The other big thing is we've got to address energy costs.
[0:34:10] And our energy dependence, partly for security reasons. So perhaps there's going to be a confluence of good things here. But we have the most expensive energy in the Western world. And the problem with that is that's going to directly translate into intelligence. It's literally watts to dollars to tokens. So you cannot build our industry out. That's the new industrial base, if you believe what's going to happen with AI, is data centers, information, it's tokens. How much can you inference per unit cost?
[0:34:39] It's vital now. It's vital for all our industries, but it's going to be vital for the industries of the future. And then that will pay itself. I don't know if we're going to get to climate and environment, but obviously there's a worry about that with the data centers and things like that.
[0:34:54] I just think that's a bit, to me, that's a red herring because if we build these technologies right, it will be AI that will help us deliver the technologies of the future that will actually solve climate change. So we're working on fusion, material design, better solar power, all sorts of things are going to be advanced in this new renaissance era I see coming. We could be at the forefront of that. And that is the thing that will then completely change the climate equation, not trying to shave off 1% or 2%.
[0:35:24] Can I come into that? Because I'm riffing off of that. I think in the future we will, talking to our engineers at Southampton, we will all want a data centre in our city or our town because we'll turn it round, we'll flip it, so they become the generators of power, electricity in the future.
[0:35:45] It's where it's going to go. The other thing I want to say that we are good at here and something you're going to see quite a bit about over the next few weeks. Firstly, we have, thanks to what Rishi Sunak did, the best AI security institute in the world. They are the best and they are setting the benchmarks for how you evaluate what's coming out of the tech companies. I wish they'd do China as well as America, but anyway, that's another whole issue.
[0:36:14] And we are part of the network of those around the world which are now called the network of AI evaluation measurement and science or measurement science. They've changed the name because Trump didn't want them to be called security institutes. And we also have funded a new center for AI measurement at the National Physical Laboratory.
[0:36:40] And this is driving the whole science of AI, not AI for science, but the science of AI, which I think is something we're very good in this country at leading on these types of new sciences. I think that's a huge potential for the future. And it all comes under the umbrella of what is being called AI assurance, which we're also very good at doing this type of stuff in the UK. And you're going to see an announcement.
[0:37:07] in London Tech Week next week from the government about growing that as a career, not just for technical people, but for people who are going to run the AI assurance industry. And we'll be able to lead on that here in the UK if we keep going. And I think this government's got that right.
[0:37:27] I 100% agree on that. Just again, to call the praise out of the AI Safety Institute, as it was called, Security Institute now in the UK is the world leading. And we are extremely good at that. And I think that's something our academic base can feed into too, because you actually want non-company.
[0:37:44] Organizations doing that type of work, really. And so I think that is leading the way. And I think that's one way towards having the technical capability for whatever standards body there is. Exactly. You need the highly technical folks, and those could come from these security institutes, networks. Super. Thank you. I'm going to just switch to a slightly different topic now. I'm just going to ask them this.
[0:38:11] Hemis, I wanted to ask you, you mentioned about UK and the power of the brain that we have here. We have got a good talent pool. I was very keen to understand, A, why did you set up DeepMind in the UK? You still run DeepMind in the UK, which is phenomenal in London. It's great for us as a company as well. But what we want to, I wanted to ask you, how do we find new British...
[0:38:40] Breakthroughs and create global businesses out of them. Well, look, I started DeepMind in London for a few reasons. One is that I was born and grew up here and I love the country and love London and I love where I studied at Cambridge and so on and the things it gave to me and I wanted to give back. Second thing, but on the commercial side, as a personal reason, the commercial side of things, there was a lot of talent in the UK and actually in Europe that was untapped.
[0:39:08] So the story I used to tell my early investors was, imagine you're a physics PhD from Cambridge, world class, but you don't want to stay in academia in physics, but you want to stay in the UK and you want to do a job that really stretches your mind. The only option for you would be to go into finance and hedge funds. Excuse the finance people in the room, but I wanted to give a different option.
[0:39:32] Right. Other than that, like, what if you wanted to stay in something scientific or deep tech? So I just felt this is back in 2010. This is like the, you know, the dark ages that no one was doing any AI, deep tech, anything like that in London. And so I had the whole field to myself. So I was hiring these amazing people for like 50k a year.
[0:39:50] You can't even get an intern for a day for that these days in AI. And so that was incredible on a shoestring budget we had. We raised a couple of million to just try and get going. And I could hire world-class people, not just from the UK, but all around Europe, EPFL, ETH, and the top French schools. And there was just nothing that was rivaling that. So that was an amazing, to begin with, and it lasted for about...
[0:40:17] Six, seven years. And then big tech sort of realized there was a lot of talent here. They've all set up their satellite offices here, which is good for the UK, actually, because it brings more and more talent and investment here. And you see that all around King's Cross. The whole of King's Cross is now this amazing place because of all the tech investment that's gone in. And then, again, those people stay for a few years, and then maybe they go and do startups. So that's the other thing. And then the final thing was, I think our...
[0:40:45] A bit of an underdog story is I kind of wanted to prove that it could be done.
[0:40:50] I just felt we were punching under our weight in this area and I think we proved a point with DeepMind and hopefully that inspired, and I think it has inspired, the next generation of founders to start amazing things here in hard technologies, deep tech, whether it's like quantum or fusion or indeed finance and banking. But I think we could go further. We now need to produce our first, let's say, trillion dollar company here.
[0:41:19] that remains independent and grows here. That's probably the next challenge for the UK ecosystem. But I believe we'll do that. And there are pockets of that like Daniel X trying to do this in Stockholm with Spotify and then he's inspiring a next generation of entrepreneurs there. And you just need a few things to happen and seed and then the ecosystems with the government's help can build on that. Yeah, but the key is the government's help.
[0:41:44] That's where we don't get it. There are people in this room who are working on that very thing. You know that. And it's treasury and you just can't get through, change the way they do things. So they're missing opportunities, I think. We've got the brains. We've got the universities. We've got, as you said, Demis, we have to repeat that, all the talent here.
[0:42:10] Amazing talent here. Other countries, I mean, if you look at what's happening in America and how this government is stopping funding research in America, it's very scary. It's a good opportunity for us, though. And, of course, you've got China.
[0:42:25] Sorry to keep mentioning China, but you have. But we have a fantastic opportunity to ride high now, but we do need the government to sort out how we get the investment into the companies. We can get the unicorns. It's getting it to the next stage. Yes, the next level now. And just one more question for you, Demis. I was very keen to understand why did you choose protein folding as the first?
[0:42:53] Yes. Well, look, I spent my whole life working on AI because I thought it would be the ultimate tool for science. So I've always been interested in the big questions.
[0:43:02] You know, of the nature of reality, how do things work? And for me, my expression of that was to build the best tool to help scientists. Now, protein folding and protein structure prediction, I came across that as an undergrad in Cambridge. A lot of my biologist friends, one of specifically was obsessed with that and would talk about it all the time. So I listened very carefully. I tend to try to listen very carefully, especially about people who are very passionate about things.
[0:43:29] I thought it was an amazing, first of all, an amazing scientific question.
[0:43:33] It was amazing. It's like the ultimate puzzle. How does this one-dimensional genetic sequence fold up into these exquisite three-dimensional little nanomachines? Amazing. And it was also what I call a root node problem. If you cracked it, it would open up whole new branches of discovery. Drug discovery, fundamental biology understanding. And I'm always attracted to these. It was described to me once as Fert-Matt's last theorem equivalent for biology. So who can resist that? And people have been...
[0:44:03] We're trying to solve it for 50 years. And then the other thing was, even back then, this was in the 90s, of course we had none of the technology that we have today, but I could feel it was the right sort of problem.
[0:44:15] for AI to perhaps help with. So I just filed it in the back of my mind for like nearly 20 years, really, 15 years. And then until the day after we had the AlphaGo match and I came back from South Korea and not only had we won the match, but AlphaGo had played an original move, this Move 37. So it made a sort of discovery, obviously in the limited domain of a game, but still kind of unbelievable that it could do that. And that was what I was waiting to see, having a system that could
[0:44:45] Not only perform at that level, but do those kind of things. And the day I got back is when we started the Alpha Thrill project.
[0:44:52] I hadn't heard that before. Good. That's good. And now we continue with isomorphic labs. Why isomorphic? Because I feel there's an isomorphism between biology and information sciences. I think biology can be thought of as an information processing system. Because I'm an algebraic topologist, right? So my PhD was in all this sort of stuff. I'm fascinated. Yeah, you'll like the name. So that's building on what AlphaFold did, which is protein structure.
[0:45:22] But that's only one, it's an important, but only one component of the drug discovery process. And that's a field, I think, that needs to be completely reimagined. And that's what we're trying to do with Isomorphic. Yeah, it'll be interesting to see what happens in the next three to five years in the pharmaceutical industry. Yes. Very, very interesting times. We'll be watching what you guys do.
[0:45:44] Because you're doing something with material sciences as well. Yeah, that's a separate thing. So we're doing that at DeepMind. So we're applying AI at DeepMind to many areas of science, but those are earlier stage.
[0:45:53] They're like pre-alpha fold stage. But material science, we're helping with fusion, containing plasma and fusion reactors, mathematics, weather prediction. We work with the Met Office. We've got the best hurricane prediction algorithms in the world. They're also way faster than the traditional fluid dynamics systems. So that was actually used by the US Met Office.
[0:46:16] It's a pity we couldn't predict the RMT strike today. Yes, that's a lot harder. It's a lot harder because it's about people. Exactly, they're the most complex things.
[0:46:40] My Clark is giving me really bad looks thinking she wants to start the questions, but she's smiling now. But I just want to touch on something that you said earlier on, which is equality. And I'm very curious, a very short answer on this, please. Are we going to open doors for people regardless of gender, background, privilege, or does...
[0:47:05] Technology risk actually taking us down a different path. Well, I mean, you know, I have to say, people, I really was reluctant to use the word obedience because when I got married, I refused to have the wedding ceremony that had the word obey in it, right? But I obeyed here, so... But so I am passionate about... I'm going to start with...
[0:47:34] Gender thing. I've been passionate about the lack of women in computing since the 1980s when I started in computing. And the scary thing is it's got worse rather than better. And in AI it's far worse. You look at Silicon Valley and the leadership is all one monoculture.
[0:48:04] You know, I see the younger generation of me and Sue there sitting in the audience. We've been trying to do this for so long. Coming in and looking at with all new... We can change this. We can change this. You know, the Anne-Marie Imbafidians of this world. We haven't shifted the dial in the last 40 years.
[0:48:24] And it's so important. It's even more important in AI. My mentor, one of them was Karen Spark-Jones, and she used to put on her T-shirts, computing is too important to be left to men. You know that one? This is not to denigrate men. This is to say it's for all of us, right? We need everybody involved in this industry. And it's not just gender diversity.
[0:48:51] I now say AI is too important to be left to men, right? It's actually more important because of all the things that if we're going to get to this holy grail of AGI, then...
[0:49:06] We have to think about what type of society we're building or we want to build. Now, I know that someone like Demis understands this has got to be done, but there's a lot of people out there that don't. They glibly talk about AGIs. It's all going to be the same. We just have these very clever machines and we'll just carry on life as we are now. It won't be like that. So we have to think through what type of society we're trying to build. I mean, you have... There's only so...
[0:49:36] So long you can say to people, right, your jobs are going to go, your kids aren't going to have a career or a job, and anyway, this could wipe out the human race. There is the beginning of a resistance movement to AI, and I think that's very dangerous. I remember, some of you here will remember, when people were saying, no nuclear bombs in my town, and there's all the Greenham Common women and everything.
[0:50:04] Fighting the nuclear bombs because people feel powerless. In America, they feel totally powerless at the moment. They don't understand how this government got elected really. We feel powerless because once you have elected a government, then you have no control over what they're going to do. I love saying this, my PA, her 18-year-old daughter, she's a very bright young lady going to Bristol University to study something clever.
[0:50:32] And she told her mom the other week that her class have decided their classroom is AI-free. Right? But that's what the end of the people, students at Southampton, I'm sure in other universities are saying, because we're telling them how to use AI in their studies. And they're saying, we don't want to do that. If this stuff is all the things the companies are telling us, we don't want anything to do with it. Now, we don't want that. We want people to grasp the...
[0:51:00] The great opportunities that AI will afford us. So that's something we've got to tackle. And it's not part of the safety, it's part of... And the tech companies have got to get involved with this, and I don't know where governments fit in, but we have got to have this discussion and think about what type of society we're trying to build where we've got these very clever...
[0:51:25] Potentially intelligent. I use that carefully because I think, I'm going to say one other thing, then I'll shut up, which is that I think there's a difference between machine intelligence and human intelligence. And we might get AGI, but they're not going to be human. Thank you. I've got a question from the floor for you. It's from the policy chairman of the City of London Corporation, my very dear friend, Chris Hayward.
[0:51:52] It's a question to you, Sir Demis, if I may. And it's in two parts. And the first part is, if there is one policy decision that the UK needs to take in the next 12 months to lead not just in building AI, but in deploying it responsibly at scale, particularly in financial services, which is the area I represent, what would it be? And the second part to the question is, where do you see the...
[0:52:16] Greatest opportunity for collaboration between organizations like DeepMind, other leaders in the ecosystem and institutions such as the City of London Corporation to help accelerate safe adoption. Right, okay, so it's two good questions. I think the first one I probably would talk about two things actually, which is one is the energy situation.
[0:52:42] I think that needs to be addressed anyway for security, but I think it's doubly important given the AI economy coming down the line. We need to be part of that. We need to build data centers. We need to have that first party stuff here on the ground and take advantage of that.
[0:52:55] The second thing is what we also talked earlier about sorting out, how do we get from unicorns? We do very well getting to unicorns. How do we get to the really big global influential companies? We need some of those here. That's the next challenge. And I think government has to be part of that, stock exchange, the whole ecosystem, your world needs to sort that out, I think, to help allow for that to happen. So I think it could happen with the right ecosystem.
[0:53:22] And then the second thing on the, you know, what can we do more and collaborate? I mean, I think these kinds of forums are good. This is why I was delighted to participate in this. I think it's great. And it's always fun chatting with Wendy too. And these kinds of forums where we bring together, but also maybe not just technologists, but we need to, this is the amazing thing about London specifically. We actually have world-class things in many domains. So the US, as you know, that's split across different cities.
[0:53:51] And the government's in DC, then you've got New York's financing, life sciences is Boston, and so on. And then you've got Silicon Valley, and then you've got entertainment in LA. We have all of that pretty much to a world-class level in one city. And I have always been a huge believer in multidisciplinary.
[0:54:10] I'm even more of a believer in that now and I think AI can enable that because it allows you to get expert quite quickly at some other subject area. And I also think what we need to do to solve, like actually I totally agree with Wendy's earlier point about that's what I was meaning about philosophy. It was just a highfalutin way of saying it. We need to work out intentionally what sort of society do we want to have and build in a post.
[0:54:34] AI, AGI world. Forget the term. The term's irrelevant. We're not disagreeing about the term. Forget the timeline. Who cares if it's 5, 10 years, even 15? It's soon. It's in our lifetimes. It's in our kids' lifetimes. It's going to happen. So just deal with that. That's is. That's a given. The horse is not going back into the stable. It's happening. We need to then work out what do we want to do about it. The future is not written. It's for us, society, to write that. But only if we...
[0:55:03] Approach it without fear. It's a challenging thing. It's scary. There's uncertainty. I think no one has the answers. None of the big tech people have the answers. No scientists have answers yet. But it doesn't mean there aren't answers out there. It's for us as society to decide that. And I think that requires all parts of society to be involved. And I'm meaning the humanities as well as the sciences and the tech and industry and government.
[0:55:30] The UK and London could be a center for that, actually. But only if we're successful enough economically and technically to have a seat at the table. That's the prerequisite for then having a voice that others will listen to. Okay. Thank you. The next question. Did you want to just add a bit? No, that's fine. Okay. Next question, Deputy Master, Gary Moore. Master, thank you.
[0:56:01] Interesting you use the words, the future has not been written. In Frank Herbert's Dune universe, the prohibition of thinking machines became a foundational background of that civilization. It stemmed from a catastrophic interplanetary war where artificial intelligence and sentient robots enslaved humanity. Nice positive start to the question.
[0:56:28] So you've touched on this already, but as we move to AGI and perhaps thinking about regulation, do you have any reflections in a world where some of this stuff's already open source, how we manage rogue actors, not just those who submit to regulation? Well, we have to pass laws that have teeth.
[0:56:54] Well, we've got to, that's my, I mean, we've got to actually, I mean, take, for example, deep fakes, right? We don't have laws to deal with these properly. Sorry, I've got to, we, you know, if someone takes someone else's image and does something awful with it, that should be illegal. And when, if and when they get caught, there should be teeth in however the sentence is. That's simple. There's going to be harder stuff to deal with. I'm going to, you take over. Yeah, sure. So,
[0:57:24] I mean, LeanFace is a good example of that. And there are technologies like invisible watermarking that we've developed and others that it's going to become an industry standard now.
[0:57:33] Would allow if a government wanted to, and I think this will happen because it seems obvious. It's obvious. Outlawing of that, but also making sure that any model that generates media has to have watermarking as part of it and also a detection system that say government can have access to or journalists or eventually maybe the web should just automatically.
[0:57:56] Flag that up, right, if this is a generated image. Seems like that's, to me, no regrets, no brainer thing to do. But look, as to Dune, which, you know, that's what could happen if things go wrong. By the way, I've watched the film, I've read so much sci-fi, but I didn't read Dune, maybe for this reason, because it banned AI. But there are many other, you know, but I watched the great movies.
[0:58:24] We might not get the chance to recover that if AI was built wrong. I don't know how it resulted in the Dune universe. But if things go badly wrong, there may not be potential to kind of just walk that back. So we've got to get that right first time. What I would say is in science fiction, there are actually, since we brought it up, there are many great examples of universes where...
[0:58:45] Humanity did get this right, and we're kind of like maximum human flourishing to the stars. So I'd point to Ian Banks as one of my favorite sci-fi novelists and the Culture series. And that depicts a future world, maybe a thousand years from now, where humanity's spread to the stars. And there are also AIs as well, but they sort of coexist in this largely peaceful way.
[0:59:15] There are many actually interesting visions of what that could be like. Maybe we need more science fiction authors to produce positive stories about the way it should go because actually a lot of science fiction does affect
[0:59:30] What scientists then think to do, right? I mean, this has been proven through history. So I think when I speak to a science fiction author, I tell them, write some sort of hard sci-fi that shows some realistic view of where we could get to, where we should be trying to aim for in the next 50 to 100 years. I think we need a big creative explosion of ideas of what sorts of societies should we build, assuming we get some of these things right, like the technical.
[0:59:57] Aspects correct. And then we might be in a post-scarcity world, but what does that mean? How should we economics work in that sort of world if we solve fusion and material science and some root node problems? I like to talk about teamwork. I like to think about AI becoming part of teams. I think it will become... It all seems to be happening so fast.
[1:00:23] I mean, the technology might be moving on. Adoption is not happening anywhere near as fast. And I think it's not that all the jobs are going to disappear. I'm not, you know, the Elon Musk, no one will need to work rubbish. Well, I think it is anyway.
[1:00:43] Jobs are going to change and I think we will get used to AI in the workplace. It will replace, I mean, some of the grunge work which is done that people have to do, AI will do that and all will be the better for it. Humans can, we can be more creative. But it will be, take a, you know, when you're diagnosing someone's illness.
[1:01:04] You need the human doctors, but the AI will come along and say, I can tell you from the data this. You'll have all the personalised medicine. It'll be teamwork. And as I said before, I think we have to recognise that machine intelligence and human intelligence are different and that we should rejoice in that and use it for the betterment of society.
[1:01:32] And I guess in the older times, in the 60s and 70s, people talked about computers in the same light, I guess. I always use the example. My father was an accountant. He finished his work and ran a finance, well, it wasn't called finance. He ran the accounts department in a company. And they did everything by hand. They had no calculators, no computers. He could divide pound-sheet and pence by pound-sheet and pence in his head. It was all ledgers. They copied the ledgers. That's what they did.
[1:02:01] And those jobs have all gone. But there are more jobs now in the finance industry because of the calculators and computers can do so much that human beings can't do. And this is what Demis said earlier. AI will do things, or you said it, I think, AI will do things we couldn't even dream of doing before.
[1:02:22] And we need to relish that opportunity. But we've got to, alongside it, get the safety and the guardrails and whatever you want to call it, sorted. And we can't sit back on our heels and assume someone else is going to do it. We've got to do it now. So part of our livery company is, one of the pillars is charity. And we have a question on charity, which I think could be quite interesting. It's going to be the chair of the AI panel.
[1:02:53] Paul Excel. A panel is like a committee which helps members discuss, improve, etc. Paul. Thank you, Master. So my question is, across our AI for Charities community, we've got over now 90 charities who are doing some absolutely fantastic fundamental work using AI, unleashing real benefits to the communities they serve. But they're generally very concerned and worried. I mean, you've got very vulnerable,
[1:03:22] People working in these, that they're supporting in these communities. So given that we, to your point, we have a chance, all of us, to shape the future that we, our children, our grandchildren live in. Where would you say things like from a responsible AI, ethics, trust, empathy, safeguarding, if it's fundamental to prevent harm and deliver positive impact, what one to three responsible actions?
[1:03:51] Would you advocate to ask us as homework, everyone in this room, governments, civil society, to ensure that the global solutions that both serve the needs of the society, but also can deliver security, sustained growth, opportunity for all, all the good things, almost like a Goldilocks. Maybe that's too much of a positive vision, but surely it's possible for us to grasp here in London and the UK.
[1:04:22] It's quite a hard question to answer. I'm going to start off by saying my number one would be diversity. I'm sorry, but that is, when I was asked somewhere, what keeps you awake at night, it's not any of the things that Demis has been talking about, or I hear that it's the fact that we don't have enough, we don't have diversity in this world. And I think it's because of that, it'd be harder to build an egalitarian world using AI if you haven't got diversity.
[1:04:52] I'm looking, I'm thinking at it through a gender lens, but I do mean diversity in its widest sense. I'm getting older, we all are, that stupid tautology. I'm beginning to feel, you know, there are things I can't do like I used to be able to do. And we really need to think about how we, you know, the elderly, like me, are going to get the benefit of AI.
[1:05:20] School kids who have no access to technology. There's so many things where we've got to try and create a level playing field for people. And that's exactly where my theme of community came out. Exactly what you've just been articulating. Okay. We only came up with one. What's your top one? Well, it's a tricky one. I'm thinking mostly about what needs to happen at the...
[1:05:49] The sort of global governance scale, which obviously will affect everything. But I think the bottom-up grassroots aspect is critical too, in the way Wendy said. But we've touched on a lot of themes, education, making sure people are re-skilled. These are all community level.
[1:06:05] The curriculum is updated. I've told you what I would do. Maybe you agree with that. Maybe there's some way you can influence that. I'm not going to have time to do that. But someone should be doing that. So it depends what the interests are of your members. But I think helping charities get up to speed on these things and take advantage, actually, of these tools.
[1:06:27] It should be great for a lot of things like back office things, stuff that charities have to do and spend money on so that they can spend their money and time on helping the community people themselves. A lot of resources that I understand in charities are to do with fundraising in the back office, but it'd be better if it was frontline. I think that those things, we're going to need both things and hopefully they'll converge.
[1:06:51] The sort of global governance scale with the frontier labs and the superpowers and then community scale. And again, maybe this is something we can lead the way on here in the UK. We have a civic project in Southampton where we're taking...
[1:07:05] We talk about making Southampton an AI city of excellence, but we're doing it from the grassroots up. So, for example, in the charity sector, we do pop-ups. Actually, I do pop-ups. And we take some experts in a sector and we have a pop-up workshop and we invite people from, so we have one for the charity sector. And it's a sort of come along, you know, what was that Woody Allen film? Everything you want to know.
[1:07:33] Was that? No, it wasn't. I shouldn't do that. Go there, was it? Everything you wanted to know about AI but were afraid to ask, okay? And seriously, we encourage people to come who feel very vulnerable or frightened or scared, and we talk to them.
[1:07:51] As a community of charity workers. And we get examples of best practice of people who are working in charities and they're using AI to spread the word. We've done that in three sectors now. We're doing more of it. It's about awareness and education. Super. I have another question from Liverman Nicholas Bale. Nicholas.
[1:08:20] Thank you very much indeed, both of you. We're going to have a situation, I think, in a few months' time where all the Western big AI developers are publicly listed companies with the responsibility to their investors. And the big global investors are the only players that operate at the scale of trillions comparable to the companies.
[1:08:48] They clearly have a big common interest in having none of their investments blow up. So I guess my question to both of you, perhaps particularly Demis, is what can the investors do to help encourage the responsible behavior because they may be able to help the game theoretic issue rather effectively? Yeah, I think that's a great point actually. So I think that investors...
[1:09:16] could exercise a lot more influence than they probably do. I think the problem is it's so lucrative to get into those funding rounds that the terms are basically there and it's take it or leave it as I understand it. And you can't really miss out because how many of these rocket ship companies are, literally in one case, but how many of them are there? Not that many, maybe just a handful.
[1:09:45] That, I think, skews the balance a bit. Also, a lot of them have unusual control structures, obviously, where, in fact, they're not beholden to the shareholders, even when they float. So, maybe. But, I mean, look, the average investor, though, just to be fair, mostly cares about their returns. So, I don't know if that we want any, you know, even further just purely economic incentive. I think that's already part of the problem.
[1:10:08] Perhaps there's some way around that. And maybe when, what I'm hoping is the market will self-correct a bit in the agentic era. Because if you imagine that one of the leading frontier model labs wants to provide agents for a financial service company, for example, I would imagine the financial service companies will want some guarantees around the guardrails around those agents, what they'll do, what they'll do with their client data. You can't have one of them putting a trade on that loses you a billion dollars the next day.
[1:10:37] Because of some misspecified goal function. So what I suspect will happen is maybe some sort of error like that will happen with the companies that are providing things without the right guardrails. And then the market will do the correct thing and incentivize actually...
[1:10:58] responsible behavior, especially in the regulated industries, which are also somewhat the most valuable. So I think it could be good, actually. That's the kind of silver lining with these technologies becoming more commercially useful and viable, especially in enterprise, which is what we're seeing now.
[1:11:15] Kind of for the first time. We sort of saw the consumer part of it, but there's also now enterprise. And that may actually then line up incentives in a way that's actually really good for responsible deployment. So that's, I think...
[1:11:29] Pretty likely to happen, actually. And I think investors could, and also enterprises that are CTOs and CIOs of big banks and other companies, you should be using your power and your purchasing power to support providers. Obviously, you're not going to compromise on quality, but support providers that you believe are acting responsibly.
[1:11:51] That's what you can do. And I also say that to the average person in the street even. It's like, well, even as a consumer, you could vote with your dollars and vote with your feet and use the market forces for good, to support companies and labs that are doing things in ways and with leaders that maybe reflect the values that you would like to see in the world. And I think as these systems become more commercial, I hope that's what investors and enterprises will do.
[1:12:21] I think it could be tied too to whatever organisation we collectively managed to set up to do governance of AI across the world because you could, I don't know if it would be charter or treaties or whatever will come out of that and the investors could say if your company doesn't sign up to this then we're not going to invest in you, for example. I'm going to go to the one last question and this question is from the CelStyle oldest.
[1:12:49] Analyst in the UK, Liverman, Richard Holway.
[1:12:52] Yes, indeed. Yesterday, I wrote an article celebrating Computer Center entering the FTSE 100, only the second tech company to be there. Indeed, in the list of the top 100 tech companies by market capitalization, the UK only has one, and that is Arm, and that is actually listed on NASDAQ. Now, if DeepMind had actually gone on for an IPO, I guarantee that you
[1:13:22] You would currently be in the FTSE 100. In fact, you might even be leading it. And indeed, you might have acted as a catalyst to many other UK companies to follow the same route and you would have built the infrastructure and ecosystem around it just as they have in Silicon Valley and whatever. And I wonder what your views were about that. I think everybody knows what mine are. Well, the alternate history, yes, sure. Well, look, I don't know where you were in 2014.
[1:13:51] I don't know if you would have backed us with several billion dollars back then. Maybe you would have done. But I couldn't find...
[1:13:57] So back in 2014, which is another world in AI terms, we were still the only people doing AI. This is before OpenAI was set up, before we'd done AlphaGo. We just had the Atari games. And people were like, you're playing Space Invaders. Why is that important? But Google understood the importance of that, to be clear. Larry Page, specifically. So I give him a lot of credit on that, because he himself always was interested in AI. He saw Google as an AI company. The reason I sold was I knew.
[1:14:26] It would require $7 billion of investment in order to take on, try and become one of those companies. And I was barely able to scrape together $10 million rounds at that point. And the problem was is that we'd done enough, the Atari thing had alerted enough of the Silicon Valley titans. They were coming after my people. So I had to find, and we had no business. We didn't even have a product, obviously not, because I was trying to make the core technology. And so that thing I was telling you is I was paying.
[1:14:53] I was paying myself zero and people were being paid like you know 100k a year and then they were getting offers of 10 million a year so I could sort of I was sort of thinking and then they could if that didn't work why they could they could afford to come back and say here's 50 million a year so at some point
[1:15:09] That was going to go wrong. And at the time, this is before SoftBank and some of the big VCs were doing billion-dollar rounds. I was slightly unlucky with the timing because maybe if we'd held on 2015-16, I could have done a fundraising round like that. Wouldn't have been British investors, unfortunately, but I could have got it from some of the more out-there investors that had the firepower, but that just wasn't available in 2014. But the other reason I don't regret it is I wanted to...
[1:15:38] You have to get back to the research and the science. It's very hard to fundraise for a startup, worry about product, and then do something like AlphaGo and AlphaFold. I don't regret it for a moment because for me it was always about the research and the technology and what it could do rather than the money or the power and things like that. I appreciate it being better for Britain, maybe for my bank balance, but I wouldn't swap my Nobel, for example, for any amount of money you could give me.
[1:16:08] Literally any amount. And what I've seen, 3 million researchers around the world have used AlphaFold in their important biomedical research. A leading pharma friend of mine said, there probably won't be a drug invented in the future that hasn't used AlphaFold somewhere in their process. So I'm very proud of that. And I think it's something built here in Britain. I think we should be proud of that scientific fact. And who knows? Maybe I'll complete the job with isomorphic.
[1:16:38] That's kind of part of the plan, fully headquartered here. But we need to fix whatever the problem is with the stock exchange right now. Why are companies not floating on that? I don't know what... Yeah, so we need to fix that. I'll come to you in a few years' time and we can talk about it. So I'd like to just say in praise of Demis that I think the UK has done hugely well out of what Demis has done. It's obvious. And I don't think we should explore...
[1:17:06] The UK benefits from what you've done with DeepMind. I go to California, I'm out there next week, I go to Google. We ride high in California because of this man. You're starting out in Southeast Asia now. Billions of dollars of investment have come into the UK because of this partnership and high-powered jobs and talent and other things.
[1:17:34] But what's actually emerging is that whilst it feels like all the companies, and they're all doing the same sort of technology, they're actually emerging as different types of players. I've always said this. So you see, well, I don't know what's happened to OpenAI, but the Microsoft, we're getting all that through our...
[1:17:54] They're just using Office. And Anthropic's got its particular specialities. And we're going to see this type. So I don't like to make it... My money has always been on Google, with or without DEMIS, because I like their values, and they've had the search engine, they've got the data, and it's all... I love Gemini, actually. It's my fave. And it's always been where my money is. But I'm not just saying it because I'm sitting next to DEMIS.
[1:18:22] But it isn't a race that there'll be one winner of, actually. And there will be many different types of AI companies that emerge. I find that whole thing really exciting, what we're going to do with it. Well, before I say my thanks, I'd just like to invite Emily Turner, CEO of HSBC.
[1:18:47] Innovation Banking for a last few words. We've got to thank you for the power banks, have we? Thank you. Thank you. I'm an academic. I'll always take a freebie. Well, if you want more merch, I think we have some in the back. We can work that out. Please join me in thanking our esteemed panelists today for what was an amazing discussion. So, Sir Demis and Dame Wendy, thank you so much. I obviously also want to thank the worshipful...
[1:19:25] Thank you to the whole company of information technologists for hosting us today. I'm sure everyone, like me, felt the great juxtaposition of coming into a beautiful building like this and the history of the organization along with the topic today. Just an amazing piece. For those of you who don't know us, just very quickly, HSBC Innovation Banking, we are a UK purpose-built bank to serve the innovation economy. We work with 4,000 innovators, entrepreneurs, and investors in this country.
[1:19:55] As everyone talked about today, as everyone here knows, this is an amazing time to work in the innovation economy and importantly in the UK where we have the confluence of talent, of diversity of thinking, of diversity of domains and expertise, as Demis mentioned, of ambition.
[1:20:11] As you clearly articulated, and it's an amazing moment. In the first quarter alone, 74% of VC dollars in this country went into AI companies, which is a massive increase on just a few years ago. I know there was this sort of theoretical question on what would have happened if DeepMind had stayed here. I will call out that over 10% of the funding in the first quarter alone came from companies in the UK started by people who left DeepMind. So I think the story of DeepMind and the team.
[1:20:38] In importance on this economy and will the next not just unicorn but decacorn come here, probably will have its fingerprints linked to DeepMind. So just wanted to call that out.
[1:20:50] Certainly for me, but I'm sure for everyone else here, it's been amazing to be part of the discussion today, to hear from people who got us to where we are today in this space, but importantly, are playing such a huge role, all three of you, and where we go tomorrow. So please join us for drinks. Let's continue the discussion. And if you didn't have questions answered, hopefully people will be around to continue it then. So thank you all, and thank you to our panelists. So all that remains for me to say, Liverman, Sir Demis Hassabis,
[1:21:26] And Liverman, Dame Wendy Hall. I hope this is not the last time we see you in our company, in our livery. But please feel free to join us at any of our events. You'll be more than welcome. Ladies and gentlemen, can I ask you one more time to give Sir Demis and Dame Wendy a big hand. It's always fun.