Andreas Stavropoulos: A very warm, “GenAI Summit” welcome to the Prime Minister. It’s such a pleasure to be here with you. I’m looking forward to having a good discussion.
Kyriakos Mitsotakis: Thank you, Andreas. Good to see you. And we’ve known each other for a long time.
Andreas Stavropoulos: A long time. But more recently, one of the ways we got to know each other a little bit better was about a year ago when you set up the Ηigh-Level Advisory Committee. We’ve heard mention of it during the day today by a couple other colleagues in the government. So you set up this High-Level Advisory Committee on AI for Greece and I’d love for people to hear what really prompted that decision? What were you hoping to accomplish through that?
Kyriakos Mitsotakis: First of all, let me point out that I do believe that the AI revolution can be truly transformative for our country. We’ve suffered a lot over the past decade emerging from a deep financial crisis. I think the country is clearly on more solid footing, the economy is growing, our public finances are in order.
But when I try to think about truly transformative steps that the country can take, especially when it comes to the business of government, οbviously AI should present us with a tremendous opportunity and if I take into consideration also the fact that we have many global experts in the AI field, everywhere in the world, it seemed logical to me to bring some of these brilliant brains together to advise us on how AI could be a true opportunity for Greece in various areas. And I’m sure we can discuss them in more detail.
So the idea behind this Ηigh-Level AI Committee was relatively simple. Present us with a document that is not just theoretical but highlights specific opportunities for fostering an AI innovation ecosystem in Greece, for using AI to improve the business of government and for hopefully turning Greece into a leader when it comes to smart AI regulation.
I know that regulation is a word that sometimes people treat with a certain degree of scepticism, but it’s clear to me that some regulation will be necessary. But let’s think about regulation as an opportunity and not just as a barrier. So I’m expecting the report is more or less ready, I expect it to be officially presented to us within the next week. And then of course the challenge will be to make sure that we have the proper governance to implement many of the very interesting suggestions that will be presented in the report.
Andreas Stavropoulos: Well, we organised on the Committee, we organised ourselves against four different areas, right? One of them actually I wanted to start there was the legal and regulatory framework. And today in the prior discussions today there was a lot of discussion around data collection, governance, use of AI, access to data, use of AI more generally in the public sector. How do you think Greece can play a leading role in this domain?
Kyriakos Mitsotakis: First of all, we need to be aware that we exist within the European legal framework. So we’re also bound by European regulations, not just the Digital Services Act, but also the AI Act which will come into implementation. But you know, a broad European regulation sets the broad boundaries.
So I think the challenge that we need to sort of address is whether we have regulatory opportunities within the broader framework of the AI Act to actually foster development of critical AI technologies in Greece. Whether it’s regulatory sandboxes or whether it’s an understanding of how we treat anonymous data without sometimes the very strict constraints of data protection mentality that may come from the previous century, but at the same time to also focus on those areas where we feel that we can actually play a leading role as Greece.
One area which is to a certain extent related to regulation has to do with the topic about all of the ethics of AI. If Greece, as the birthplace of classical philosophy, cannot contribute to this discussion, I can’t see which country will be able to do that. So making sure that when we talk about these complicated questions, we actually have a voice that can be heard.
We can become a global centre for these types of debates. We can bring leading thinkers around the ethics of AI to Greece. This seems to me to be particularly important and when it comes to specific sort of possible regulations, to have a pretty good understanding, which I think is also going to be the product of a vibrant debate like the one that is taking place here.
I mean, what does AI for good mean and what are the real risks of AI, in particular when it comes to two topics which I obviously care a lot about. The first is, you know, the broader discussion about AI and democracy and the impact of AI on the public debate. And the second, which is slightly more narrow, but still I think of great importance, and that is the impact of AI on children and teenagers and in particular their mental brain, but also the way they learn and absorb information.
I think these are topics which I think we can be leaders also within the European Union in terms of smart regulation, without stifling, of course, the innovation that is so necessary to further develop AI as a transformative technology.
Andreas Stavropoulos: Specifically, you mentioned ethics, and obviously a big part of the world has had elections recently. There’s all these questions about the impact of technology on the democratic process more generally. And there’s also frankly a lot of concerns about the role of AI. You know, it can be a great empowering tool, but it can also be a tool that potentially could thwart democracy. Some people are concerned. Do you have any high level thoughts about what it is that we could do as a country or even as Europe, in terms of either regulating or thinking about principles, whether they be explicit regulations or not, to make AI more of a boon to the democratic process?
Kyriakos Mitsotakis: Well, even before the significant breakthroughs of AI that have taken place over the past years, we knew that the big online platforms had a model that promoted engagement by essentially allowing people to communicate with other people who shared the same views with them. And, of course, that sort of more polarising content was rewarded in terms of engagement compared to sort of a more moderate discourse. So this notion of virtual sort of echo boxes where you only listen to people who reinforce your biases is a reality. And this is unfortunately only getting worse.
So again, finding ways for people to be exposed to different ideas and to have some sort of meaningful dialogue within the digital space, it’s not an easy suggestion, but we need to find a way to break this feedback loop where we only, at the end of the day, get exposed to information which reinforces our own biases. And I don’t have an easy answer, but at least I know that this is a real problem.
And then, of course, you add to the mix the way that AI can be weaponized by malign actors. You know, the possibility that deep fakes can be created with great ease and they can really catch up like fire. We had an incident today where we had a simple post which supposedly was done by my Minister of State stating something which he had never said, very basic stuff that anyone can do.
And until you actually realise that this is happening and you contain it you know the damage may be done. And of course the more we realise that someone like myself who has been filmed numerous times, has been photographed numerous times, if one were to create a fake video of myself, there’s enough data to actually do it. And then until you realise what is happening the damage actually may be done, so these are real threats.
And of course my fundamental belief is that we need to also have this honest discussion with the big tech companies. They also need to understand that they have skin in the game and that they cannot sort of simply hide behind the fact that they’re platforms and that the content that appears on these platforms is none of their business. So I think this discussion is beginning to happen with the big tech companies and hopefully something good is going to come out of that.
Andreas Stavropoulos: You mentioned before switching gears for a second, because I think we can come back to regulation and more let’s say philosophical questions if we have time, about what is the nature of even knowing something if now you can fake all kinds of things. But I don’t want to get too philosophical, at least not yet.
You mentioned children before and the ability to learn and having good learning outcomes. So education and research was one of the other major subcategories of the committee where we worked on a number of different thoughts. Clearly in your administration you’ve liberalised higher education through non-public, nonprofit university programmes and education has been loomed large in your agenda for both of your administrations so far. Specifically, when it comes to AI, do you have thoughts about how it could further promote your agenda around education?
Kyriakos Mitsotakis: Well, first of all, we clearly know that we will need more young people who are trained in the technical skills that will be necessary in the AI industry and the other thing is that up to the universities themselves to offer these types of programmes, but also to understand what upskilling in AI may mean for people who may already have a technical background. And there we will also be looking to work not just for the academic community, but also with the private sector to understand what are the skills that we really need to offer around the broad AI space.
The second very interesting topic is how AI is going to transform the educational process itself. And there I think the jury is still out because of course the idea of having a personalised tutor to help children with those specific needs is in principle very, very intriguing. But what we learned during Covid, is that there is no replacement for the classroom and the real interaction between the teacher and the children and of course I always see technology as an enabler, and not as a replacer of human activity. It will happen in certain jobs and we can talk about that in a bit but surely in the educational process.
For example, we add a smartboard which we’ve done now to almost all Greek classrooms from 5th grade on, it is a transformative tool in terms of empowering the teacher to go out of his comfort zone and to find new material and to make the lesson itself much more engaging.
This is not sort of rocket science technology, but it does really change the way the educational process is taking place. So I think we need to monitor this space very, very carefully.
But let me just point out to a risk I see, we need to be aware of. This is a broader concern that I have. First of all, we know that sort of social media algorithms at a young age are addictive, they’re designed to be addictive. And we know that this addiction has negative impacts on the mental health of children. I think this is an established fact.
So we need to do something about it. Whether it is educating our parents about the parental controls that already exist in the applications, we plan to do such an event because most parents who buy their kids their first smartphones are totally ignorant about the fact that already, as we speak, there are things you can do to restrict the use of a mobile phone and the applications by a parent.
But of course, there’s also another topic which is of even greater concern to me. If writing in itself, I mean, is replaced by nice LLMs that can write better than we do, how will the kids learn how to write? Is it important for the kids to learn how to write? Is it important for the kids to memorise if they can find any information they want online? How important is it to have this information, you know, possess this information if it is all available?
When we were at university -and we were at the same university, Andreas was two years younger-, we still did our research in libraries with the big drawers and looking for books. And I’m not saying, you know, I’m not a Luddite, I believe in the progress of technology, but I have a concern. The human brain has evolved over millions of years to process information in very specific ways. And suddenly we throw at it a completely different reality, which is a virtual reality. And we have no idea whatsoever, no idea what this will do to brains that are developing, children and teenagers.
And the temptation is there. We wrote so many essays in universities and will the universities, at the end of the day, you know, have a capability to really distinguish between something that has been written by ChatGPT or by us? I doubt it will be so easy to actually do that, because the models themselves are also getting smarter. So if the kids don’t need to memorise, if the kids cannot, will be tempted not to write because someone will write for them. I mean, what is left at the end of the day?
So, this is a topic I know may sound slightly esoteric, but I think it is something which is of great concern to me. That is why when we think about education, especially children and teenagers, not so much higher education, I would urge us to err on the side of caution.
Andreas Stavropoulos: Well, let’s think about how you turn some of that into an opportunity, right. Greece punches above its weight when it comes to academics, especially like some of the panels mentioned before, especially in this area in applied mathematics that has to do with AI, in specifically AI area. I mean the head of our AI Committee, Costis Daskalakis, is a famous AI computer scientist, mathematician, and we have many examples like that around the world. A very active diaspora. I believe that Greece in NIST, one of the leading conferences, Greek authors were the third most quoted when judged by the names of the people who were contributing. So we have, let’s say, riches when it comes to academia and the ability to operate and teach human talent in these kinds of things. Can we turn some of these concerns into opportunities?
Kyriakos Mitsotakis: Well, we’ve spoken about the risks, let’s talk a little bit also about the opportunities because I see many. First of all, we have a very vibrant ecosystem of start-up companies active in technology and in particular, active around AI, in Greece. This was not the case, I think, five, ten years ago. We have venture capital firms that have raised significant amounts of money. And it seems to me that what’s happening in Greece is already being noticed also by people, certainly by investors from the U.S.
So one challenge and one of the briefs we gave to the Committee is what are some smart ideas to help foster this ecosystem? And what is the infrastructure, what is maybe the public infrastructure that we need to make sure that this ecosystem does much better than what it has already achieved.
And some of these ideas, maybe, I’m sure you’ve discussed them, the ‘Daedalus’ supercomputer, our application to set up to be one of the seven AI factories in Europe, which I think is very well grounded and very well supported. This will give, I think, an additional impetus to the ecosystem to further develop, to have a sort of more companies that emerge out of Greece, more talent returning to Greece.
And of course this also brings us to a question which is very relevant to Europe as a whole, because one of the topics we discussed at the European Council is this competitiveness gap between Europe and the US, to a great extent attributed to productivity related to technology. And of course we know that unfortunately now the natural sort of exit for a startup company if it were to grow would be either to accept usually VC money from one of the big American investors or even move to the US.
And one of the reasons why companies are moving to the US is because there you have one regulatory environment, whereas in Europe you have 27 regulatory environments. One of the suggestions that has been put forward, very much supported by me, is that we need sort of a 28th European startup regime by which, if you are part of that regime, you don’t need to go through various sort of approvals and regulatory hurdles in all the member states in order for the single market to really work.
Think about it, we have so many, you know, almost 400 million people in Europe, but when it comes to digital services, we don’t operate as a single market. So there’s also a European dimension to what I am describing. But certainly I’m very much looking forward to the specific suggestions of the report. If I look at, you know, the talent, the ecosystem, I mean, your presence here, I was very surprised to see how many people are actually involved in this space.
And again, allow me, sorry, I’m not an engineer. So when I try to understand how exactly an LLM model works, I really struggle. I was good at math at school, but that was about it. Then I started doing other things. But there is a breadth of talent in our universities, but also abroad of people who would want to return to Greece and work out of Greece. And I think it’s not that difficult to create sort of the proper environment. It’s already happening to a certain extent, but we can do more.
Andreas Stavropoulos: I’ll tell you on that, there was a study, actually, I just read it, that’s why it’s top of mind. This think tank in the States called “Deon Policy Institute” actually did the largest study of its kind with Greek professors of the diaspora. And I know organisations like “HIAS” have done similar and asked the question, would you be willing to come back to Greece and teach for at least part of the time, under what conditions, etc? And it was a surprisingly high percentage, and to me, surprisingly low, because he actually asked, what compensation would you need? And I was both surprised by how many were willing to do it and by how much of, let’s call it a pay cut they would be willing to take to be able to partake of the other advantages.
Kyriakos Mitsotakis: And of course, when we talk about our public education system becoming more open, more flexible, more eager to engage in these types of partnerships, I mean, I think we’ll announce very soon many interesting joint degrees. And I’m not talking about the nonprofit private university space that is opening up. I’m talking about our public universities and how eager they are to engage in particular with top US Schools. I can see lots of potential. I mean, we have something that is difficult to beat.
First of all, for all Greeks, there is always a sense that we want to come back. And of course, I think what the country now offers is a basic sense of stability and that we’re moving in the right direction and that there are more opportunities today. And that is why I think we are also able to bring back talent. I mean, the brain drain has, to a certain extent, been reversed.
And of course, one may not envision making the same amount of money that one can make abroad, but if one can have a good enough salary, a good enough prospect, and of course, the advantages of -I’m not saying that Greece is a perfect country, but it offers certainly many advantages of living in Greece-, this is becoming a compelling proposition.
And every time I meet people -also many of you may come from the diaspora- that tell me “I’m really considering for the first time to come back to Greece”, I think this is a very good sign.
Look, there’s a global war for talent. Talent is a scarce resource, and that’s why we’ve also given tax incentives to people, Greeks who want to come back. So we’d be foolish not to take advantage of the fact that we have so many skilled Greeks abroad, especially in the AI space.
Andreas Stavropoulos: You mentioned jobs. And, you know, one of the things we did with a group here is we had an email address that people could send questions in, and, you know, you could find that they clustered around a few subjects that some of them, I think, will be no surprise.
One has to do with this question of jobs and will AI take our jobs? Will people’s, let’s say, meaning of their professional aspirations, the meaning that’s attached to that, is that in danger in many ways by AI? On the other hand, as you say, we have a war for talent and unemployment levels that in general have been dropping. So it’s a little bit of a conundrum. But what would you say to people who are concerned about AI taking over our jobs?
Kyriakos Mitsotakis: The history of technological breakthroughs, you know, has resulted in jobs being lost and jobs being created. On average, more jobs have been created than have been lost. Granted, AI is a truly, truly transformative technology. But for me, the real challenge is, you know, AI as an enabler to improve productivity.
For example, we have not discussed this, but where can AI help us when it comes to the public administration? For example, and we’re already doing this, we use AI in order to review legal contracts. A process that can take an hour can be done in 10 minutes because most of the preparatory work has already been done. So we’re not replacing anyone, we’re just making them much more productive. And I do think that the first step clearly will relate to that.
And of course, I think it is the obligation, especially of companies, and I saw that you had quite a few companies participating in this gathering today, to use AI in order to improve their productivity. Again, maybe some jobs will be lost, new jobs will be created, but I don’t believe that we are faced with a great sort of jobs extinction, at least in the short to medium term.
And again, the challenge is how can the technology improve productivity when it comes, for example, to healthcare, there I see probably the most benefits and there Greece can play a role. Because if you have well curated healthcare data, and we have lots of it, and I think there we have the potential to actually leapfrog other countries, one can offer this data and come up with very, very sort of fascinating, I presume, and intriguing solutions to existing problems.
So I think there’s one thing which is developing innovative solutions and the second which is applying existing solutions to processes, whether they’re related to the public sector or to the private sector, and improving productivity. It’s not that we’re going to solve all technological problems in Greece, but if we’re smart adopters of technology, we will deliver better public services and our companies will become more productive.
Andreas Stavropoulos: I remember one of the tasks you gave us when you formed the Committee was “think of vertical areas where Greece could have, arguably could have a competitive advantage”. You just mentioned healthcare potentially being one of them. Not because, you know, our doctors are necessarily smarter than other people, but because in some ways coming from behind can actually be an advantage if you don’t have to deal with legacy systems, you can leapfrog a whole generation. I mean, if you think about what happened to the cell phone companies we all knew when we were growing up, none of them survived the internet and the smartphone movement, it completely shifted. So I wonder if this… Are there other verticals that you think that Greece could play a, this is a “gr-english” translation, protagonistic role?
Kyriakos Mitsotakis: First of all, let me take a step back. I think the fact that in 2019 we created the Ministry of Digital Governance was truly transformative in terms of the governance of offering AI solutions for the broader public sector. Because what we did at the time, and nobody really noticed, all the big databases were essentially integrated within one ministry. So we broke down the vertical silos.
And as the Ministry started offering, you know, interesting solutions, all the other Ministries wanted to team up with the Ministry of Digital Governance because Ministers understood that there was something in it for them as well. So we created a positive loop for offering, you know, the first level of digital solutions, which is what you see with gov.gr and now updated. But this is sort of the basics, this is 1.0. Now we’re moving to to the “smart” state at a higher level.
And sometimes the solutions are already there. For example, we have a very sophisticated electronic prescription system in Greece. It didn’t take rocket science to figure out that by doing some basic cross checks and using simple tools, we can find the doctors that actually overprescribe. But it was not done for other reasons, not because the technology was not there.
So sometimes you need the will. If, for example, we want to control illegal buildings in Greece, it’s not that difficult using spatial mapping and satellite imagery to see what exactly is being built and understand whether you have a permit for it and whether the footprint of what you see is what has been permitted. These are things which we can do and which we are doing.
And one area which is of great interest to me is climate adaptation. Because there, of course, when it comes to wildfires, predicting how a fire will spread, when it comes to advanced flood prevention models, where it comes to more sophisticated weather prediction systems, there I think we are already seeing very interesting solutions and I think we can do interesting things already.
And the other sector is also defense. Defense is becoming much more important. In Greece we had a tradition of buying expensive defense systems from the big, be it the US or France. And of course you’ll always need your planes and your frigates. But there’s so much happening in the defence space that is small, decentralised and that can really be developed by the brilliant talent that we have in Greece, that it would be a pity not to develop it, not only for our purposes, but also to actually become a country that, why not, exports technology.
For example, you know, the anti-drone system that we have now on our frigates was developed by Greek engineers. Αnd having a flexible model of providing sort of initial seed financing for prototypes and then seeing if that works, I think it’s an obligation for us to do it. So I see significant opportunities in defense. Βecause we spend a lot on it and because our whole philosophy is changing and it has to change to incorporate the lessons of, for example, the war in Ukraine.
Andreas Stavropoulos: While we’re in climate, there was a question that actually a lot of people had, and it was actually posed today as well, about energy efficiency in general. And, you know, some of these AI systems are big consumers of energy and one might think, well, you know, Greece could have an advantage to green energy, but are there things we should be thinking or doing in light of our natural resources? Nuclear? I mean, the US large companies are even investing in nuclear energy again. What are your thoughts in general about energy efficiency around AI?
Kyriakos Mitsotakis: Well, first of all, I think there’s a new interesting story about the geopolitics of data centres and Greece has been attracting data centres and we’re doing it because we think that there is a spillover effect. But of course, we also need to be sure that we address the energy issue and we may actually come to sort of ambitious projects where you have really big data centres, big consumers, which are powered by dedicated energy sources which are off grid. So this is actually what’s to a certain extent is happening in the US.
So, two thoughts on this. First of all, on nuclear technology, I’ve been making the case that when it comes to sort of bringing down our emissions, we need to be technologically agnostic. So we shouldn’t say, “this is the way to go necessarily”. But there’s so much investment in clean tech taking place, that probably in five years we would know much more. And I don’t want to place our bets or put our eggs in one basket and realise that at some point there’s another technology that has been developed.
For example, what’s happening with SMRs, small modular reactors, is of great interest to me. We don’t have sort of a nuclear background in Greece, but certainly we would be interested in pursuing this technology and seeing where it can lead us. We have islands, for example, that need decentralised systems. Not necessarily all of them will always be or can be interconnected.
And of course, the second thing is, how can AI help us with a system of energy production which is so random? Because renewables are random and making the system smarter seems to me to be a very complicated optimisation problem to which I’m sure AI would have interesting solutions to offer, because we will always be very dependent on renewables and of course storage will be part of the solution and we will need some baseload backup. But as we increase the penetration of renewables, having smart stochastic systems that make this complex energy market actually work, that’s a good problem to try to address.
Andreas Stavropoulos: And there happen to be some startups that I know have Greek founders around the world that deal with those kinds of things. Very interesting. Last question. Time is up. We have a next panel coming. But last question. That also I hear from a lot of, let’s say, the more sceptics out there. Like, Greece is a small country, AI is a big global race. Can we really play ball, so to speak, at the Premier league level, as a small country?
Kyriakos Mitsotakis: We can punch above our weight. We certainly can because we have the talent. And I think we also have a government that is willing to try to educate itself on how it can use AI to be more effective. So certainly we can do this. And when I look at what other countries are doing, I don’t necessarily see, you know, them being so much more advanced than us when it comes to our thinking.
But again, our biggest assets are talent. So if we can bring, develop enough good brains around AI, but certainly leverage those people who are active in the AI space abroad and bring them to Greece or to Greek related projects and build the right infrastructure, definitely we can do this.
And after all, who knows where the next big idea is going to come from. I mean, sure, there is a question of scale. You need hundreds of billions or trillions to invest in this huge infrastructure. But that’s not necessarily what AI is all about.
Andreas Stavropoulos: For me and we’ll close it there, this kind of commitment, having you here today at an event like this -we have a lot of the elements in Greece, as you say, we have academia, we have talent, we have a lot of elements. The story has not yet been told fully, but it is through things like your presence at events like this and committing the kind of work you’re doing with this Committee -and hopefully seeing some of these, you know, to come to light, some of the initiatives-, is that kind of commitment that I think gives us a chance.
Kyriakos Mitsotakis: Well, I almost invited myself to be here because I really wanted to communicate this message.
Andreas Stavropoulos: And this is what I’m talking about. So thank you. Thank you for being here.
Kyriakos Mitsotakis: Thank you very much.
Andreas Stavropoulos: You honour us with your presence.
Kyriakos Mitsotakis: Thank you.
