[0:00:05] Welcome to the second Athens Innovation Summit at the Odeon of Herodotus Atticus. Here, at the very place where democracy was born, we could create a global platform to discuss the future of entrepreneurship and how to ensure that innovation makes the world fairer, more resilient, and more prosperous. By gathering here, we are standing at the very place
[0:00:31] Where some of the boldest, brightest and biggest ideas were born, molded and put into action. With incredible impact changing the way we live to this day. Today's high impact entrepreneurs follow paths that started in ancient Greece. And Athens stands as a timeless symbol of what is possible by humanity. We're hosting this night because if we believe that heritage can guide us to a better future.
[0:01:00] Then Athens has definitely a global role to play. Ladies and gentlemen, please welcome on stage the Greek Prime Minister Kyriakos Mitsotakis, the co-founder of Google DeepMind, Sir Demis Hassabis, and the co-founder of Endeavor, Linda Rotenberg. Good evening, everyone. I can't think of a more perfect setting or a more perfect pairing for tonight's conversation.
[0:01:42] We are seated at the southwest slope of the Acropolis, the birthplace of democracy, philosophy, and drama, to discuss one of the most dramatic moments in human history, the rise of artificial intelligence and the challenge it poses to all of us to get this moment right. I'm joined by Prime Minister Kiriakos Mitsotakis, who has been a strong advocate for innovation while also emphasizing the need for ethical frameworks.
[0:02:12] And Vicer Demis Hassabis, who has been at the forefront of this revolution while also advocating for educating citizens on the limitations of AI. I'm honored to be with you both and look forward to our conversation. So I'd like to frame tonight's conversation around three larger themes. One, what can we learn from past moments of enormous human change, the golden age in Athens, the Industrial Revolution, to shape our approach?
[0:02:42] Two, the current dialogue around AI centers on Silicon Valley versus China. And yet, as we see in the 50 countries where Endeavor operates, innovation is everywhere. We're big proponents of what we call the power of elsewhere. So question number two, what can we learn from a country like Greece and a London-based company like DeepMind that can help inspire founders everywhere?
[0:03:12] And three, with the prospect of general artificial intelligence or AGI, it poses a unique challenge to human civilization. So we'll spend most of our conversation in this area and ask you both to draw on your experiences to help us navigate the inherent tensions of the moment. So let's dive in. Prime Minister Mitsotakis.
[0:03:38] You and I met many years ago studying philosophy and social theory at Harvard. And there was even a class called Thinking About Thinking that began with a focus on ancient Athens. So we're here at the Herodian. As you think about the future of AI, what values from Greece's past should guide us? Well, first of all, it's a real privilege to be sharing the stage with Nemeth. Thank you so much.
[0:04:07] for being here and for supporting us in this fascinating journey to understand how we can use AI in a truly positive manner. It's amazing to be sitting here, usually we don't appreciate how this theater looks like from the stage, but I think it is sometimes worthwhile to think and draw lessons from classical Athens, although we need to point out that this
[0:04:37] Odeon was built during the Roman era. And to understand what was it that caused this amazing innovation in terms of organizing our societies and what forced Athenians to abandon old systems of political regimes in favor of democracy. And at a time when democracy is under challenge everywhere,
[0:05:06] Looking back to ancient Athens and trying to draw lessons from that unique experiment, acknowledging its limitations, I think is very much worthwhile understanding the true power of debate, the necessity to be able to convince citizens who will then vote all together about the collective actions of the city.
[0:05:34] And also to recognize, I think, what Fusilidis said in his funeral oration, which I always try to remember, that if you don't do well in a city that is flourishing, the city will find ways to support you. But if you do well in a city that doesn't do well as a whole, you will probably also end up not thriving. So connecting the benefits of the city.
[0:06:04] To individual flourishment is something which was very much at the heart of Athenian democracy. So I always like to look back and to draw lessons from this period. And I think these days, this is a very healthy exercise. Sir Demis, you have also spoken a lot of thinking about thinking and the need to draw lessons from the past.
[0:06:32] I know you're a student of the Athenian philosophers. And with regard to another great disruption, you've said that AI will be 10 times bigger and 10 times faster than the Industrial Revolution. So take us behind the scenes in the labs with your engineers and designers. Are you talking about Aristotle and Plato? And are you worrying about a new Dickensian London?
[0:06:59] Well, thank you. And first of all, I want to say thank you to the Prime Minister and the government and Endeavour as well for inviting me here. It's incredible to be talking here in this amazing historic place.
[0:07:10] And as you mentioned, those grades, I'm really proud of my Greek heritage on my father's side. And one of the things that I studied a lot growing up was the classics and Greek history and also mythology. And it really actually was very formative for me about the way I think about the world. And you talk about those, the grades, you know, Plato, Aristotle, Socrates, Archimedes.
[0:07:33] Really, this is the birthplace of not just democracy, but also science and philosophy. And actually, my favorite piece of art is the School of Athens by Raphael in the Vatican and depicting that kind of golden era in an idealistic way. And I think we can learn a lot of lessons from that time. I think especially the way, as you mentioned in the beginning,
[0:07:55] Sciences and arts can be blended together and drama. And it's all one part of sort of being human and advancing human knowledge and understanding of the human condition. And that's how I've tried to work in my career. Even though I've worked in technology, I've also worked in things like designing video games, which involves artistic side of things too. So I'm very interested in how AI can help both with the sciences, advanced science, but also how it can help enable...
[0:08:25] More creativity too. Wonderful. So let's move on to the topic of innovation elsewhere. Demis, you were born in London. Your father was Greek Cypriot, your mother, Chinese Singaporean. And yet when you raised capital from American investors, they told you you had to move to Silicon Valley. You resisted and stayed in London. Tell us how you made that decision and what have been the pros and the cons.
[0:08:51] Well, sort of, you know, when we started DeepMind back in 2010, it's hard to imagine today, just 15 years later, but almost nobody was working on AI, even in academia, let alone in industry. So it was a very unusual thing to try and raise venture capital for. And we had to find unusual investors in the West Coast. But I was wanting to stay in London. I mean, I grew up in London. I love London as an international city. But the main things were, I felt there was incredible talent in Europe.
[0:09:19] And the UK that was untapped at the time for deep technologies like AI. And I felt that if we could set up an ambitious organization that was really trying to reach for the sky, those types of people would be attracted to work in a place like that. And we'd sort of have the field to ourselves in the UK at the time.
[0:09:41] And that's what transpired. And now, of course, hopefully DeepMind has been a good example for many other deep technology companies now to start up in Europe and in London. And we're proud to have played a part in that. But then the other really important thing was more from an ethical point of view, which is that even though we...
[0:10:02] 2010 seemed like a long way to go before we would make progress with AI. We planned for success. So at some point we imagined and we knew, and even before setting up DeepMind, that if we were successful in this endeavor, AGI and AI would be one of the most important technologies ever invented.
[0:10:20] So if that's true, and it was possible to build it, then of course that would also come along with a lot of important ethical questions and ethical challenges about how to make sure it benefits all of society, and also technical questions around how we keep that kind of technology under control and on the right guardrails. So we already had at the beginning of...
[0:10:42] We were planning for, if we were successful, then these ethical questions and tenets really would have to be built in from the beginning. So we've always had that in mind. And so where that comes with the Europe is that I think it's the AI and it's clear to everyone now, it's going to affect every corner of the world, every industry, all parts of society. So I think it's important for the world to have a say in how...
[0:11:07] This technology has developed, not just a small part of California. And I think Europe's voice in that, I think is really important and the value systems that we have in Europe. So I think you can only have that kind of influence if you're also at the leading edge technologically. And that's what we've tried to wave the flag for with DeepMind and do our part in that.
[0:11:30] I loved, by the way, meeting Yiannis of Reflection AI, a Greek founder in AI who worked for you at Google DeepMind. And I've been saying that Endeavor's all about starting the tech mafias and you have your like the PayPal mafia and you have your DeepMind mafia already. It's amazing. Mr. Prime Minister, you are working hard to make Greece a place that really matters in terms of innovation elsewhere. You and Endeavor Greece have collaborated on an initiative called Innovation Nation and Pharos
[0:11:59] AI is one of the first AI factories in all of Europe. We were talking earlier and you reminded me that you worked as a venture capitalist before you went into politics and have cared about this topic for a long time. So describe your local and regional efforts on innovation and globally, how do you plan to compete for talent like Demis? Well, first of all, thank you, Demis, for mentioning the important role that Europe can play.
[0:12:29] In this new emerging world where it seems to me that most of the discussion involves a competition between the United States and China. And the truth is that when we talk about Europe, we usually think of a supranational entity that is very good at regulating but not very good at innovating. And I think there is some truth to that. And one of the main missions that...
[0:12:55] We as European leaders have agreed to is how do we make Europe as a whole more entrepreneurial? How do we break down barriers between countries? How do we create a true single market, which really does not exist in digital services? I'm sure this is the experience also of a lot of Greek startups. And how can we actually also raise capital at the European level?
[0:13:18] To ensure that our startups or Greek startups don't necessarily have to be acquired by the big US companies in order to charter their own future path. So there is, I think, a big...
[0:13:31] European dimension in terms of fostering innovation and we have made good use of these European initiative and the European funds by setting up one of the first seven AI factories in Greece which is essentially a structure revolving around a very powerful
[0:13:51] A supercomputer that is in the process of being built in the city of Lavrio, but also trying to make sure that our startups have access to cutting-edge computing power and can actually develop AI solutions out of Greece. And I must pay tribute to the tremendous work that Endeavor has done in fostering
[0:14:16] What is an incredibly dynamic startup ecosystem in Greece. I was a venture capital many, many years ago. I set up the first incubator in Greece back in 2001 and never would have imagined that this dynamism would emerge out of Greece. And it's a complete...
[0:14:38] Change in the mindset of young people, thinking about entrepreneurship and actually taking the risk to set up their own company. So for us, I see AI as a transformative technology that can really allow the country to leapfrog other European countries in our effort to catch up with the rest of Europe. It's a simple.
[0:14:59] And this means not just supporting our ecosystem of entrepreneurs, which is something that we do through incentives or supporting more venture capital, but also trying to use artificial intelligence at the level of the government. This was a topic of the discussion that we had today in terms of becoming much more efficient, but also in terms of identifying specific verticals where AI can truly be transformative.
[0:15:28] Healthcare, education, civil protection, just to mention a few of these vertical areas where we really want to be at the forefront of applying artificial intelligence for the benefit of all Greeks. So I'm very enthusiastic and very excited about the prospect of this endeavor that we have sort of started.
[0:15:56] To implement and really extremely bullish about also the future of the Greek entrepreneurial ecosystem. That's wonderful. And I would never have imagined that we would be able to fill this auditorium, you know, 45 minutes before Greece is going to play a very important basketball game. Thank you for being here.
[0:16:22] We will get you out in time. And Yannis is actually a supporter of Endeavor, so we love him. And I should say that everyone met our incredible chairwoman, Costanza, as well as our managing director, Panos. But it was your wife, an entrepreneur, Mareva, who first convinced us to bring Endeavor to Greece. Yeah, I should point out that, you know, I know Linda since we were classmates at Harvard College.
[0:16:51] I tried to convince her to bring Endeavor to Greece. I failed, but Mareva succeeded, so well done. I'm sure she's very persuasive. Okay, so those were the warm-up questions. Now let's dive into some of the tensions, starting with innovation versus regulation. So, Demis.
[0:17:12] You're part of an industry that is famous for its move fast and break things technology and philosophy, which translates to government keep your hands off. And yet you've said that perhaps AI was released from the lab too early. So now that the genie is out of the bottle,
[0:17:35] Are there opportunities for governments to regulate AI or is your colleague at Microsoft, Mustafa Suleiman, right that containment is no longer possible? So, you know, I think we've thought a lot about these topics over the years and I feel like...
[0:17:54] You've got to be at the cutting edge of the technology to understand what's coming down the road. So if in terms of it's moving so fast and it's changing so fast, we have to be careful about how we regulate it. It's so important. It needs to be regulated well. But we have to allow the innovation to flourish and be bold about that whilst mitigating against the risks. And actually, although, as you said, the...
[0:18:22] The tech industry is famous for its move fast and break things. I actually don't subscribe to that view with this technology. Of course, it made a lot of progress, and a lot of the modern technology we see today comes from that kind of relentless progress, and that part of it is good. But I think with a technology as transformative as AI, we have to be more thoughtful about it than that. Maybe in terms of the break things part of the move fast and break things, we need to be more careful and try and have as much force.
[0:18:52] So I would advocate instead of that philosophy to use the scientific method and to try and get as much understanding about these systems before they're sort of deployed really widely. And it doesn't mean we'll get everything right all the time, but because things are changing so fast, but we should try and do the best things possible to safeguard against these risks whilst enabling all the amazing use cases like...
[0:19:18] Advancing science and medicine and all the amazing things that I think AI will bring into the world. So we've got to get that balance right of being bold with the opportunities, but being responsible about mitigating the risks. And that's going to be a continual tension, creative tension, I think all the way to AGI. I love that answer. Mr. Prime Minister, what is your answer to the question of regulating AI?
[0:19:46] What can national governments like yours do? Or should there be some sort of supranational structure akin to what we did in the monetary system following World War II? Some are calling for a Bretton Woods of AI. I've been thinking a lot about this question and unfortunately there is no easy answer. I do believe that at some point some sort of global arrangement will be necessary.
[0:20:14] Especially an understanding between the US and China that there are real risks involved in this technology and that if this becomes a race to who develops first the most powerful model then there is really no incentive to regulate because there are commercial incentives and geopolitical incentives just to be the first to get your hands on this extremely powerful technology. And this of course
[0:20:41] It creates an environment where no one talks about regulation. And very few people are as thoughtful as them about the consequences and the ramifications of this incredible technology that we have created. In Europe, we have our own regulatory framework, which I think is reasonably well thought out. And I'm talking about the AI Act. But at some point...
[0:21:05] We don't want to be a regulatory black box in a world where no one is talking about regulation. So I think it is important to engage both with the United States and with China and certainly to engage with the big technology companies. And in my mind, we need to pick our regulatory battles in a smart way. So we need to set some clear priorities. For example, for me and for Greece, there are two areas.
[0:21:35] that concern me profoundly. The first is the health and the mental health of kids and teenagers. And the second is the impact of AI on our democratic discourse. And we were discussing this with Demis. We're going to have elections in two years. And I think the number of deed fakes that we will see by then is going to be a real challenge to the democratic discourse.
[0:22:02] To actually getting to the truth of anything if we don't know what is true and what is created by artificial intelligence. And especially when it comes to the question of mental health of kids and teenagers, there I think we need to be stricter also with the big technology companies in terms of either pushing for a design that really takes into account the relative age of the users.
[0:22:32] Or even considering more radical options such as banning the use of social media under a certain age if we cannot succeed otherwise in mitigating the negative impact of our kids having access to smartphones, which really have been built simply with a simple purpose of maximizing engagement. And again, this is not...
[0:22:59] This is probably a question relevant to social media, less relevant to artificial intelligence. But of course, as these algorithms become smarter, they will also become smarter in terms of capturing our attention. And when we're talking about the brain of a child or a teenager, we're really running a massive experiment with the mental health of the next generation. And I don't think we are fully cognizant of the risks that we are assuming. So I would really like to engage
[0:23:25] with the big tech companies and to make them understand that they need to be part of this discussion and they need to be bolder in thinking about this challenge. And if regulation is necessary, then we should not hesitate to regulate that space smartly. One last point. I think Greece has led probably Europe in terms of
[0:23:53] Technologies around age verification. So we now have what we call a kid's wallet, which is a platform that allows parents to set up the phones of their kids in a way that suits their own preferences and choices. It's a first step in the direction of at least making parents and kids aware that being sort of addicted to the phone
[0:24:23] And being addicted to constant scrolling is probably not the best thing you can do as a kid or as a teenager. Yeah. And I'd love, Dennis, to have the chance to respond. Let me just add that one of our classmates two weeks ago announced that one of their children, who's in their 20s, had taken their life, left no note.
[0:24:43] And the parents discovered that they had been conversing, not with Google, but with one of your competitors, AI, and had used that as their therapist and friend and decided to take their own life. And it's just, I'd love the chance for you to respond because this is a real thing that parents are worried about. Yeah, there's a lot of complexities here that I think have to be navigated very carefully. And I think the prime minister is absolutely right about that.
[0:25:06] We at Google, of course, are thinking very carefully about those things. And I think we should learn the lessons from social media, actually, where, you know, this attitude to maybe move fast and break things went ahead of the understanding of what the consequent second and third order effects were going to be.
[0:25:21] And I think with AI, we have to get that right. And in fact, we also need to think about the way AI interacts with social media. And I think it can help with some of these attention issues in terms of summarizing things and helping with finding the things that are useful and helpful for people without hijacking their attention. But it can also, if it was misused, it can make the problem worse. So I think we've got a lot of...
[0:25:48] I also agree with the Prime Minister that ideally we would have some international cooperation or understanding or minimum standards around the deployment of AI. I think that's where the rubber meets the road is where there's the research.
[0:26:07] And there's interesting research challenges there about AGI and autonomous systems and how to keep control of those things. But actually how when it affects the public, it's through the deployment in products and services. And I think that's where maybe regulation in the near term can focus on because it's more tangible as to what the effects are that governments want and society wants to see. But again, enabling all the good use cases of AI as well and increased productivity.
[0:26:36] So again, it's going to be about getting that balance right between innovation and mitigating the risks. Another area where there is dissonance between political leaders and tech leaders is whether the future will be shaped by scarcity or abundance. Mr. Prime Minister, I suspect or can imagine that a lot of your energy is focused on managing the real world consequences of climate change.
[0:27:06] And yet we hear from the futurists that AI will solve all these problems, ushering in an era of radical abundance, and all we need to do is shift our zero-sum mindset. So as a leader, how are you balancing the short-term anxiety with the long-term promises? Well, Keynes famously said, in the long term, we're all dead. So I think it's a very valid question. I think that if one were to...
[0:27:36] Try to imagine the world in 2050. I think a lot of the challenges that we are facing today, such as, for example, nuclear fusion, unlimited cheap energy, these problems will probably be solved by then. But the question is, how do we get from where we are today to where we will be in 25 years? And there is no doubt in my mind that AI can create
[0:28:04] Tremendous productivity gains and hence create tremendous wealth. Then the question becomes, okay, if there is a lot of wealth to be created, how is this wealth going to be split and how is everyone really going to benefit from this amazing technological revolution? And we're discussing what happened during the industrial revolution.
[0:28:30] And this transformation is going to be much more significant. And there was a lot of social upheaval. And we should have no doubt in our minds that a lot of jobs will be threatened and will be destroyed as a result of this revolution. And we're not just talking about blue collar jobs. We're talking about white collar jobs. And that is going to cause a lot of anxiety.
[0:28:59] It will put our pension systems under a lot of stress. So you will have productivity gains for industry or for capital, especially the big companies that are developing these technologies or the companies that are actually using it in order to enhance their productivity. So wealth will be created, but you will also have a significant displacement of jobs. And I think it's already what I'm describing. I think it's already happening.
[0:29:25] In the United States, if you look at the labor market, and I don't think we are ready as a society about this reality, which will hit us relatively soon. So there was always this discussion regarding the destruction of jobs and the creation of new jobs. And every time in every big technological revolution,
[0:29:53] People were very, very pessimistic only to realize that maybe more or different jobs or better jobs were created in the future. But there is going to be a period of transition. So unless people actually see benefits, personal benefits to this revolution, they will tend to become very skeptical. And if they see, sorry for being so blunt, obscene
[0:30:22] Wealth being created within very few companies. This is a recipe for significant social unrest. So I would very much urge us to engage in a more substantive discussion with the technology companies to at least acknowledge the fact that these changes in the labor market may actually happen sooner than we think.
[0:30:50] Because at the end of the day, if AI is a productivity tool, it enhances clearly the productivity of workers, but it will also replace human labor. And every corporation would have an incentive to do that if at the end of the day, the return on its investment is positive. We will get back to the question of jobs. But Dennis, I want to focus a little bit on this notion of radical abundance.
[0:31:20] It does feel like we're being told daily that a cure for cancer, limitless clean energy, lifespans to 150 with interstellar travel is like weeks away. And you won a Nobel Prize, shared a Nobel Prize for AlphaFold, which correctly predicted the structure of 3D proteins. And that is amazing.
[0:31:42] But actually, skeptics are now comparing this era to the dot-com bubble. And The Atlantic went so far as to call AI a mass delusion event. So are you worried that your colleagues are overpromising and underdelivering and that people aren't feeling the benefits of all the wonderful things you're speaking of? Yeah, so I've no doubt that I think that a lot of the...
[0:32:08] Progress in AI has been incredible and almost at light speed, but it's not a silver bullet in of itself. And I think what I often say is that it's being overhyped in the near term, the very near term, the next couple of years, you know, AGI, I think is at least five to 10 years away. There's still some big breakthroughs that I think are required, but I think it's also underestimated, underappreciated quite how transformative it will be when it does come, even still now. And just the way of quantifying that, we're talking about industrial,
[0:32:37] I think we can learn a lot of lessons from. But I expect AGI, the advent of AGI, to be maybe something like 10 times the impact of the Industrial Revolution, but maybe 10 times faster as well. So kind of 100x. So in a decade, over a decade instead of a century. So if you think about that, that's pretty monumental.
[0:32:56] I think that's what's coming down the line. And we have to use the time now in governments and technology companies and society needs to use the time now to prepare for that and what that means. And I think there'll be these amazing things. There will be this kind of notion of radical abundance in terms of an alpha fold. I think it's just the first example of that that we'll look back on in a decade. And hopefully there'll be many more examples of these big breakthroughs. Working on cancer. Yes, on things like fusion and material design, optimal batteries.
[0:33:25] All sorts of incredible new technologies in both energy, in physics and in medicine. And I think that could lead to a world where we have this notion of radical abundance, where we're not limited by the amount of resources. Maybe we're also, because we have this unlimited sort of renewable free energy, we can make space travel easier. These types of things, they sound quite fantastical perhaps today, but I think in 20 years, five years time, it will be perhaps maybe...
[0:33:55] But then the question comes is how do we make sure all of that...
[0:33:59] And I think that's more a government and society and social science and economics question rather than just a technological one. And how do we do that internationally? And what are the right governance bodies? And probably they don't exist today, the right types of institutions to deal with those types of questions. So maybe we need to do some institutional building.
[0:34:28] And then we can sort of deal with that next era. So I think it's going to be very exciting, but there will be challenges. And I have a lot of confidence and faith in the adaptability of humans. You know, we're incredible. The human brain is the only evidence we have maybe in the known universe of general intelligence being possible.
[0:34:52] And look, we're sitting in, look what we've, humanity's created with our minds, modern civilization, modern science, modern philosophy. And I think we're infinitely adaptable. And I think we'll do that again for this next era. So you brought up near-term jobs and he's saying the part of the near-term allocation of equitable distribution of resources and jobs is up to leaders in government like you.
[0:35:22] What steps are you taking to ease this uneasy transition? How are you preparing Greece for the disruption of traditional jobs? And how are you preparing Greek citizens for the new jobs of the future? Well, first of all, I think it is important to align our education system with the needs of the labor market. This is a challenge that is faced by
[0:35:52] by all governments. As the requirements of the labour markets change, so do also the expectations from our educational system. There's a lot of talk about reskilling and upskilling. It's a difficult exercise to implement, but it is probably the only solution I can see in terms of ensuring that we can offer people maybe
[0:36:21] Slightly different jobs from what they had, which will be better adapted to the new reality. And of course, I do need to point out that Greece is a country with a healthy manufacturing base, but it is primarily a service economy. And I think service economies, probably in the short term, are better protected than countries that rely more on manufacturing. Or I can see...
[0:36:47] The double impact of AI and robotics because we have not spoken about robotics yet, but I think that is another completely transformative technology to significantly displace a large number of jobs. Look at Europe, for example, and the importance of the automotive sector. Electric cars require fewer components. The supply chains will be shrunk. We're faced with a lot of competition.
[0:37:16] already from the US and in particular China. So the countries that are dependent on auto manufacturing unless we radically rethink the role of the European auto industry are going to be threatened in terms of significant job displacement. In our case for example in tourism or in jobs we discussed it that require still human interaction we are in the short term
[0:37:43] Better protected. I don't yet see in the near future a robot being able to serve a customer who comes to visit Greece. But this does not mean that the job displacement has not started. And we need to be very monitoring the labor market very, very carefully and try to adapt as well as we can. And of course, also educate our children.
[0:38:11] In terms of using these new technologies. Because at the end of the day, I prefer to acknowledge that our kids are already using artificial intelligence without naming specific companies. I prefer to train them to use artificial intelligence in a smart and efficient manner than to have them...
[0:38:38] Go and ask their own questions and then just present them as their own work at school. Because this is a big temptation. You should have no doubt about that. My girls will have oral exams and in-class essays. It's difficult to explain to, I don't know, even a 15-year-old or even a college student that they actually have to think and write if artificial intelligence can do this for them.
[0:39:07] So we should also talk to neuroscientists about the development of the brain and what are these basic skills that are required for us to become fully developed, well-functioning humans. Because our brain is a product of millions of years of evolution. But if it starts losing the capacity to do things which it has been doing for many, many years,
[0:39:37] What could that mean for a generation of kids who may have difficulties just writing if artificial intelligence can write for them? What are the basic skills that make us truly human that we need to protect at all costs and to use artificial intelligence?
[0:40:06] Maybe in a more sort of Socratic method to challenge kids because it can be truly transformative. Every kid could have a private tutor, but we need to understand that right now it's probably used by kids to help them avoid to do things that they should be doing on their own. Yeah, I totally agree with that. I think that's why...
[0:40:30] I mean, the educational system has to catch up, I think, in terms of kids are using these systems today, but maybe not in the most beneficial way.
[0:40:38] train the next generation to use these new tools in ways that are complementary to human thought and help advance their productivity. And if we do that, I think they can have 10x the kind of productivity and output that we have today, just in the way they've used computers in a smart way. People used to worry about this when the advent of home computers, also mobile and the internet, but they can be used both in positive ways to speed up things like learning and understand more than one subject area.
[0:41:07] Connections between those things. So for example, multidisciplinary thinking, I think is easier than ever before now because of the availability of that information. But only if you go and do the hard work to go out and get that and actually absorb it properly rather than just on a superficial level. So I think that this has lots of implications for our educational systems. But I think for the next decade or so, I would still advocate the core learnings in STEM, mathematics, physics. I think if you have a better understanding
[0:41:36] Understanding how the technology is built, you'll be able to use it in more productive ways. I think that'll still be true for the next phase. Yeah, and I'd love to hear just one level down. I mean, there's a lot of debates as to which short-term jobs will be most at risk. So where do you come out of this? Who's most vulnerable in the short term? And longer term, what would you recommend? What skills, what tools, what knowledge should kids today, students today, lifelong learners be developing? And if you were back in colleges.
[0:42:03] What STEM classes and what humanities classes would you be taking? Well, look, I think it's very hard to predict the future, like, you know, 10 years from now. In normal cases, even harder today, given how fast AI is changing, even week by week. So what I would say is the only thing you can say for certain is that a huge change is coming.
[0:42:25] So if you have that as a premise, then we need to start thinking about with students of today, what are the meta skills, I call them, that you need to learn in addition to knowledge that you need to have about mathematics and sciences and so on and humanities? What about the meta skills of learning to learn? Learning about yourself, what conditions do you learn best under? How do you optimize your own learning rate on a new subject? Because one thing we'll never show is you're going to have to continually learn.
[0:42:54] I think throughout your career.
[0:42:56] The opportunities I think will appear is actually the biggest ones will be connecting two different subjects together and finding something that is analogous between the two or combining them in some way. And one of those subjects could be AI applied to many things. But I think there'll be lots of low-hanging fruit, both scientifically and in business, in finding interesting connection points between two seemingly disparate fields. So that speaks to this sort of idea of multidisciplinary learning. And I think AI can actually...
[0:43:25] It can actually be a very helpful tool, personalized education, personalized tutor, to allow you to go much more deeply into many subjects than perhaps you could before. Wonderful. So perhaps the area of greatest tension as we move toward AGI is around this issue you brought up, Mr. Prime Minister, about the essence of what it means to be human. And if we go back to ancient Greece for a moment,
[0:43:52] What made this hillside famous wasn't just the thinking about thinking, it was the architecture, the theater, the creativity. As the mother of an art historian and a poet, I'm especially interested in what happens to these areas of our shared identity.
[0:44:10] Come from an artistic family and have an artistic background. You've also said that culture is the output of our collective brains. So how do you believe that AI is going to reshape culture? Well, I hope that what we're doing at Google and DeepMind is trying to build new tools using these technologies that help.
[0:44:33] Empower the creative process. And we have huge respect and I have huge passion for the creative arts. As you said, earlier in my career, I was working on video games and designing and working with artists. And I saw firsthand doing that in my teenage years.
[0:44:48] How powerful it can be when you combine cutting edge technology, graphics, physics, and AI in those days in the 90s, along with artistic and creative talents. And a lot of magic can happen when you combine those things together and their technology enables new creativity. So that's the way I'd like to see it go. And we have a lot of collaborations with top...
[0:45:09] Film directors and artists and musicians around the world who come in and use our tools when we're still forming them and give us feedback about what they would like to see, what would be helpful for them in terms of outputting and making their creative process more powerful or simpler for them or quicker for them. And I think that's still in the early days of exploration now of how that will be enabled. And actually just going back to the jobs question, I think there's still this part of like,
[0:45:38] The special element, almost like the soul of the art and the culture that the human narrator or creator will bring to the art itself. And I think that's never going to change. And I think that's the heart of drama and human drama and all of these things. And I think AI tools will just be enablers of those things. And with jobs, it may be that sort of counterintuitively, we may revalue certain services or certain expertise that may change, but that may not be a bad thing. So for example,
[0:46:07] One counterintuitive example I give is that we may end up valuing nursing or caregiving more than being a doctor because of the different way AI will affect those different professions. But perhaps we are undervaluing and underpaying caregiving today. That might be the case. So we'll have to see how that develops over the next decade or so. Great. I want to come back to this issue of creativity and what it is to be human. But Mr. Prime Minister,
[0:46:34] You've said that Greece should contribute to the global ethics of AI and you've also argued for a human-centric approach. So what do you mean by this human-centered approach and what should Demis and his colleagues keep in mind as they're pushing towards AGI? Human-centric approach to AI means that AI should be a tool towards human flourishment and not
[0:47:04] A replacement for human activity or human thinking. So it's a question of how AI can help us become better, to put it in very, very simple terms, which, of course, sounds like a very general statement. But if you dig a little bit deeper, it raises a lot of questions. You asked the question of
[0:47:32] A basic philosophical question of whether we can treat, you know, how do we approach AI? Does AI have a conscience? Can a system of artificial intelligence express emotions? What if a 15-year-old starts interacting with a system of artificial intelligence and that system becomes, that bot or whatever you call it, becomes his or her friend?
[0:48:02] And how do we program these algorithms in terms of dealing with what is a fundamentally human prerogative, and that is emotions? Because we've spoken a lot about thinking and emotions is what distinguishes us to a certain extent from all other species. So we know that AI can do a reasonable amount of thinking and it can speak in natural language. So speaking, which is also what distinguishes us
[0:48:32] from all other species has already been done by AI. But how are you going to program these algorithms to ensure that if a 16-year-old starts engaging with the system, and this is a real example, and starts asking questions about suicide, that the system steps in and offers
[0:48:58] You know, those answers that one would expect it to offer. And there is no easy answer to that. But this is a topic that really concerns me. And I would be very, very interested to hear Nemes' thoughts on this. Yeah, this is where I think we are really talking about philosophical issues here in some sense, not just technological ones of what it means to be human.
[0:49:28] What is special about the human condition? What do we want to protect about that? And I think I would advocate, although we're only just one part of the field, but with our kind of leading role of building AI in the first instance very much as a tool.
[0:49:45] That enhances and helps humanity flourish. And I think if we do that in the right way and build it in the right way, I think we can have sort of this amazing new golden era of maximum human flourishing where we understand the universe around us a lot better.
[0:50:00] Human health is enhanced. And perhaps we're also, as you mentioned earlier, traveling to the stars and spreading consciousness to the universe. So I think that could be the future, an amazing golden future for humanity. But on the way, I think the first step will be to build these incredible tools and then maybe use those tools to help us understand our own minds better. What is consciousness? Can we define exactly what it is? That's still in the philosophical realm. Can we measure it?
[0:50:27] Do these systems have any semblance of that? I think today, currently, they clearly don't. And other things that we think of as uniquely human, from creativity to emotions to dreaming, we would like to understand, I think, from a neuroscience perspective and psychology perspective, what those mechanisms are. And then I think we'll have to make at the next stage the decision about what are future AI systems, what should we build into them?
[0:50:52] But I think for the very first step, we should very much think of them as useful tools. And that's what humans are in the end. We're tool-making species. That's what separates us from the other species is our ability to make tools, both physical and now with computers, sort of digital. I'd love to dive into this because I heard a wonderful podcast you did with a mutual friend and Endeavor Global board member, Reid Hoffman.
[0:51:16] And you and Reid were talking about these new AI tools and natural language prompting, what some are people calling vibe coding, and how if harnessed correctly, creatives, film directors, game designers, and other creatives could become superhuman. And tying that into our conversation about Elsewhere, two of the most exciting companies in the Endeavor network now are Eleven Labs, the leading voice AI company in the world founded by two Polish entrepreneurs.
[0:51:44] And Runway, which is pioneering video generation AI founded by two Greeks, two Chileans and a Greek. And so I'd love you to tap into this notion and expand on how AI tools can super power us or make us superhuman and not just the pets of AI. Yeah, so I think these kinds of tools you're talking about, and we help build a lot of the models that are underlying some of the things that some of these tools use, is...
[0:52:11] I'm going to be empowering in a couple of slightly different ways. On the one hand, I think it will democratize creativity in the sense that everyday people will be able to create interesting things quite easily, much more easily than they can today, using some of these tools, video creation, audio creation, and so on. But at the high end, I think for the professional creators, artists and directors and so on,
[0:52:35] And I think...
[0:52:59] Both of those things paradoxically will be true with these new tools. So you'll get a lot more tinkering around and fun things that are kind of throwaway bits of creativity. And I think you'll also get, enable the top creators to create even more meaningful and more plentiful versions of those things that deeply connect to us. And I think there'll actually be a desire, a strong desire from that, from society to have more deep and meaningful art.
[0:53:28] That's hopeful. Okay, before we turn to the last question, Mr. Prime Minister, I would love to hear your thoughts on a topic that we haven't explored yet, which is the subject you began with, which is AI and democracy. To many of us, democracy is feeling fragile now. And do you have any hopeful thoughts to offer us?
[0:53:57] And then we have a basketball game. I think democracy obviously faces numerous challenges, but I do believe that the prevalence of fake news and the new reality that is emerging of AI being able to easily sort of create content that is what we call deepfakes.
[0:54:29] It's obviously deeply disturbing to us. But I think a lot of the damage has already been done by the sort of old traditional social media companies in terms of them only being in touch with people who think along their lines, because that is exactly how you drive engagement. And also clearly, you can drive engagement in a more...
[0:54:56] Effective manner if you talk about negative things and about positive things. Societies are extremely polarized and people who are completely unable to listen to the other side of the argument. You see that in the US with the horrible assassination that just took place and this complete inability to have debates where the
[0:55:24] The position of the other side is respect. We may disagree, but we're not enemies. We don't kill each other. Democracy was invented and also is a way to manage our differences in a civilized manner. But if we start picking up guns and shooting at each other, then the whole notion of democracy collapses. So I would envision AI as maybe a gentle tool to challenge our
[0:55:54] Rather than reinforcing our beliefs and putting us in touch with other people who think along the same lines, because this is what drives engagement, what would it mean to use AI to, in a smart way, say, well, maybe this is not exactly how you think it is, and to challenge people to think differently. Maybe there we can think about opportunities, because clearly the traditional social media companies had no interest
[0:56:24] In developing this sort of interactive dialogue that is at the end of the day at the heart of any well-functioning democracy.
[0:56:34] Did you want to respond? Well, maybe I can say a couple of things about that. I think, first of all, we see democracies are sort of under threat across the globe and our value systems. And I think the way to protect that is you have to protect that from a position of strength. And I think that means technologically and economically. And I think that's why AI embracing that in all its forms, the opportunities is critical for Europe to do, UK to do, and the West to do.
[0:56:59] And so that's part of why it's so important to be technologically leading in this new space. In terms of what AI can do to maybe help the discourse, I'm hopeful that we can use AI to do exactly what you're saying, Prime Minister, of broadening the debate. And I have this notion of if we get AI right, maybe this idea of technology working for you as the individual.
[0:57:25] So you have a personal assistant that has your best interests in mind and maybe then surfaces things that are perhaps nourishing or beneficial or helpful to you, not just what the kind of social media algorithms do today, which is mostly optimizing, you know, a single sort of optimization process of your attention, grabbing more and more of your attention, but not necessarily in a way that's beneficial to you as the individual. So I think with AI, because it's going to be a much more sophisticated
[0:57:54] Algorithm, if you like, than today's very simple statistical algorithms, it can potentially take into account much more things than today's sort of social media, narrow social media algorithms. And of course, you as the individual can speak to it in natural language and effectively program what you would actually like.
[0:58:13] Your assistant and your technology to do for you. And then have it go out and almost negotiate this torrent of other algorithms to bring to you or be helpful and beneficial to you and helpful in your daily life, including exposing you to more interesting ideas. So that's my hope for what AI will end up being used for. And I think the capitalist and market dynamics will hopefully select.
[0:58:39] And users and consumers will select for those types of AI systems that do have those attributes. So one minute for each of you remaining. What do each of you hope that the 5,000 citizens and leaders in the audience take away from today's discussion? Would you like to go first for a second? He'd like to go last. Well, hopefully everyone can take away the thoughts of...
[0:59:08] How much complexity is here? And there are people thinking about this who are in the technology area, but also how all parts of society that are represented here have to come together and have these kinds of debates and also on an international level. And I think Greece has a really important part to play in that as a cradle democracy and science and maybe helping to lead the EU and Europe to think about these things.
[0:59:35] Well, I would hope that we think more deeply and more constructively about the challenges, the opportunities and challenges of artificial intelligence. And I would also hope that this
[1:00:00] The dynamism that is exhibited in the Greek startup ecosystem continues, strengthens, and that we can actually manage to use this transformative technology for good purposes while mitigating to the best of our ability.
[1:00:28] The risks that are clearly there and that I think we discussed today. And as I look up there, I would also hope that Demis will, as a Brit who traces part of his roots back to Greece, would also...
[1:00:52] Help us in our effort to bring the porthenon sculptures back to where they belong. On message. Thank you both for this extraordinary dialogue. And thank you to the Prime Minister's office, to Google DeepMind, and to Endeavor, particularly Costanza and Panos, who curated this evening. Sitting here, well...
[1:01:27] Sitting here in this extraordinary venue, we're reminded that the future isn't just for computer scientists and political leaders. It's for philosophers and artists and citizens, and that multidisciplinary dialogue is the way forward. So thank you. Thank you all.