 Good afternoon. Welcome to today's press conference both through the journalists here on the floor as well as the audience and press who are joining us through our live stream. Thank you so much for being a part of this very pivotal moment in the advancement of global artificial intelligence. My name is Kathy Lee. I'm the head of the Center for the Force Industrial Revolution and also head of AI data metaverse. Today we are thrilled to announce the launch of an ambitious and also transformative initiative, the International Computation and AI Network of Excellence or ICON, as we say. I'm privileged to be accompanied today by a panel of esteemed experts. Joining me today is to my immediate left, Alexandre Fassell, State Secretary of the Swiss Federal Department of Foreign Affairs, FDFA. Next to him, Professor Christian Wolfram, Vice President of Wolfram Swiss AI. And next to him, Professor Bernhard Schoenckhoff, excuse me if I get the name wrong, Director Max Planck Institute for Intelligence System and Alice Chair. And next to him is Professor Keisira Mana, Board Chair of Data Science Africa. This initiative is a big step in rebalancing global AI efforts led by the Swiss Foreign Ministry and involves Swiss AI leaders such as the Federal Institute of Technology in Zurich, ETH Zurich, and the Federal Institute of Technology in Luzon, EPFL, as well as the Swiss National Supercomputing Center, CSCS, along with partners from the Alice Network. ICON is a unique collaboration. It brings together the government, private sector, civil society, academia to tackle the power imbalances and societal issues linked to AI. This initiative also shares a common perspective with the Water Economic Forum's AI Governance Alliance and aligning with IGAS mission and core working areas. Today's press conference marks the official start of ICON's incubation phase, a crucial step in their mission to promote large-scale AI research and development across different fields. This initiative goes beyond just technology. It aligns with the United Nations' sustainable development goals to ensure that AI progress benefits all of humanity. So my first question is to you, Secretary Faisal. Can you please provide some insights into Switzerland's specific motivations and also objectives for participating in this extremely important initiative? Thank you very much, Cathy. We all know that AI is amongst the key technologies of our time and that it will have a lasting impact on society, the economy, politics, and that it will have an important, probably determining role to play in tackling the global challenges we face as an international community, be it climate change, pandemics, economic inequality, and so on. Now, unfortunately, not all countries have the same access to the resources needed to implement AI projects. And so if we do not change that, if we do not address this problem, then potentially artificial intelligence becomes a driver of a new global inequality. At the moment where we are fast closing the digital gap, we must avoid opening AI gap. And so with ICANN, we believe that we can ensure that all voices are heard in AI research and that AI solutions are perceived as a global public good that are available to everyone around the globe. And that is our approach. And you mentioned the United Nations. So in this way, with that project, we are able to specifically support what the United Nations are doing, notably through its high-level advisory board on AI, to which the head of the AI Center of ETH, Professor Andreas Krause, has been appointed. And so our commitment emphasizes Switzerland's leading position in AI research, but is also an expression of Switzerland's constant concern and engagement for an equitable international order brought about notably through this leading location center of global governance, which is Geneva. Thank you very much for that. So let me go to Professor Wolfram. And if you may also speak, I think, on behalf of EPFL as well. So what factors led ETH and also EPFL to believe that this initiative is so essential to the AI landscape and research in general? Yeah, so let me echo what State Secretary Fassel just said. The Swiss AI initiative that is jointly led by EPFL and ETH is actually addressing exactly this point. AI is such a transformative technology and as we heard, it is probably impacting all the different areas of our lives. And this is very important that these kind of changes, both on a technological level, but also on the types of models that are generated and all the governance, the ethical implications are not just run by the private sector, but that the public partners, where actually quite a large amount of expertise is located, is driving this forward. This is sometimes a bit more difficult because universities are often distributed, but if we band together, like the ETH domain is doing this now, I think we have enough power to really shape the field and contribute to the development. And this of course also excludes access for everybody to these technologies, access to data, access to models. And this will in the end benefit us all because the global challenges will not just affect few countries, they will affect us all. Absolutely, thank you for that. Now let me move on to Professor Ashokov. Can you please tell us a little bit more about Alice, what the organization's goal and objectives are, and also just tell us a bit more about your motivation to be part of this initiative. Yeah, thank you. Alice stands for European Laboratory for Learning and Intelligent Systems. It's a grassroots initiative in AI with a focus on scientific excellence, innovation and impact. It builds upon machine learning as driving modern AI. So this is inspired by a model of human intelligence that's not pre-programmed but evolved and learned from data. Now virtually all of the progress that we're seeing in modern AI is due to machine learning. And it's important to understand that it's quite possible that we're still at the beginning of this development. So possibly much more is yet to come. It's for this reason it's crucial that the research in this field is done in a transparent way, including European societies and values. So to ensure that this takes place, we need to attract and retain top talent in our institutions and we need to build ecosystems that also support the creation of startups. And we view the ICANN initiative as a crucial cornerstone in this endeavor. Thanks. Professor Mena, tell us a bit more about your organization and the motivations as well. Okay, thank you for that. The organization I represent is called Data Science Africa and we like to think of ourselves as a grassroots organization that tries to leverage this very powerful technology of data science, artificial intelligence to solve problems that are relevant in our African context. So we bring together academic industry players and we've evolved towards, we initially started as a capacity building organization. So one of the major gaps that existed when we first started was capacity, just the expertise in this field of data science. It's an emerging field and AI as well. We really consider this broad spectrum of emerging technology. And the expertise, there was a huge gap on the African continent, we were aiming to plug the gap and we've evolved towards where we actually help researchers actually build solutions, whether it's be in agriculture because there is a huge need to solve these problems in agriculture, the climate crisis and also things to do with language. So making sure we can leverage these technologies to the benefit of people on the African continent. That's really the reason data science Africa exists. Thank you for that. Let me go back to State Secretary. During the last 12 months in particular, there's a proliferation of different AI networks and research groups and many other types of alliances and associations as well. Tell us why ICANN can be set apart from the other similar initiatives. Yes indeed, AI is a very crowded space, as you say. But when you look at this space, then you immediately recognize the specific mission and unique position of ICANN. When you look at Europe, for example, with the European Lumi and Swiss Alps, we have two of the most modern and powerful supercomputers in the world. But in developing countries, you do not have such capabilities. And again, that can create a new equality and open a new gap. And the mission of the ICANN is to precisely fight this gap and prevent it. Thank you. Let me go to Professor Wolfram. What would you say, what distinguish ICANN from other projects, especially the ones that usually engage with a more commercial model? Yes, so I think one big aspect that has been touched upon in some of the comments is really the aspect of transparency and openness. There's a huge discussion on the dangers of AI and how to make it equal distributed. And I think one aspect in that area is really that it is open. That means it must be absolutely clear what data was used to train the model, how the data was obtained, how one arrived at this model, what are the basis. And this is often not the case in the models that are provided by the private sector. And this is, of course, the chance for us, especially as we are a Swiss institution, to provide a completely open and transparent system, especially when we generate models that might be helpful on a global scale, and thereby also enable other people to utilize the knowledge that we generate in the context of these models and apply them to their own aspects. And thereby, we again would leverage the field a little bit and make sure that a certain sense of equality is restored. So if I may ask a follow-up question, the data will mainly be public sector data or is that something that's to be worked out? I think this is something that has to be worked out. Public data sector is one thing. Maybe it's also data that is collected as part of an ICANN initiative. We are discussing this. So I think we'll have to couple this. Existing data might not always be sufficient. So I think one aspect will also be to discuss how do we get the data to achieve the power to train a model that does what we wanted to do, for example, in the African space that we discussed previously. Right, so obviously you need to figure out how to deal with and leverage the low resource data as well. Let me go back to Professor Shokov. What do you think, what kind of contributions that your members can make to this initiative? So in a nutshell, what we can contribute is scientific expertise. So the members of ELIS include essentially all the top AI or modern AI places in Europe. So that's, of course, ETH, EPFL, but it's also places like Oxford, Cambridge, the Paris Ecosystem, the Max Planck Society. So overall we bring a considerable heft of high-powered international expertise. And we like to think that this is crucial for such a network because we can contribute to our expertise in deciding how the resources, the compute resources best should be used. Thanks for that. And Professor Mena, let me ask you this question from the demand side. What do you hope to gain from working with ICANN? Well, what does success look like to you? I think there's things we can gain, but there's also other things we'll be able to contribute because there are certain unique challenges that we are facing within the African context. So we face certain, you know, problems that we tackle using these technologies, AI. And then we have challenges based on compute resources, but we have emerging expertise in the certain problem domains that we have to focus on, like I was mentioning, agriculture, language, and climate resilience, require a set of expertise that's emerging within the African context. And our network includes academics and industry players who have that expertise, which I think we can then contribute to this ICANN network. And there can be a lot of mutual benefit between members of Data Science Africa who have their own unique expertise and expertise within ELIS and within Swiss universities, EPFL and ETH. And I think success looks like a lot of fruitful collaboration between people on both, on all sides, really, to tackle problems that are now becoming common problems, like how to do very efficient agricultural production leveraging technology. Absolutely. You're not only the recipient of this initiative, but you're, you know, equally contributing to it as well. Let me see if there's any questions on the floor. I do have more questions, but just wanted to see if there's any immediate questions, both online, offline, or waiting. Let me go back to State Secretary. What is this, a Swiss initiative? It is an international project, and the Swiss initiative indeed. It is the concrete expression of the Swiss digital foreign policy and the Swiss science policy, and those two elements aim to capture the excellence we have in Switzerland in terms of science, research, and technology, and capitalize on this excellence to make it fruitful and useful for the international community in tackling global problems. And so we have this excellence, and we are happy to team up with international partners who also are based in excellence, so to team together. And so from the outset, from the design phase of the project, this initial Swiss initiative becomes an international project with the AI network. Alice, as we have said, with the Swiss components, EPFL-Losanne, ETH Zurich, Data Science Africa, Lumi Consortium. So you see that there is this constant drive from Swiss diplomacy to make the world benefit from our excellence in science and research and technology, and team up with friends who are in the same fortunate position to offer that to the global commons. Indeed, the Swiss government has a great track record in terms of international collaboration and coordination, so we're very excited to see this. Speaking of being the fortunate, I'm gonna turn to Professor Wolfram. So ETH is fortunate to have access to one of the best performing supercomputers. How do you benefit from sharing it with others? So this is maybe a point of clarification while the CSCS is part of the ATH domain. It is a national infrastructure and it is paid for by Switzerland, right? So, and should therefore also be used as such. And I don't think we are giving anything away. We are gaining, and this is what Professor Mena said. It's not just gaining in part of a collaboration, but also the knowledge gained from these questions. Let's discuss, for example, too little water for the farming. This is of course a problem in Africa and the models might be used to inform farmers how to treat certain crops, but even in Switzerland we started to have droughts. In Italy it's even worse. So I think these problems will come here as well, so we can utilize even the knowledge that we gain from these models and apply this to a problem that might not be that evident in Switzerland. So it's really a given take in both directions, not just on the level of a very fruitful collaboration, but also from an exchange of knowledge between the countries where problems, global problems get swapped. And I think this is why it's important that this resource is used because ultimately it will benefit as well. Absolutely, so it's a win-win situation. Going to Professor Schochoff, is that what you're thinking when you started this discussion and conversation around the synergies between ELIS and ICANN as well? Yes, very much so. So we are in ELIS, and I believe the same is true in ICANN, the way we conceived it. We're committed to ensuring that the major scientific and social issues surrounding AI are tackled jointly and transparently. And if one thinks about how doing it to do this, this requires not just top talents, the scientists, but it also requires experimental equipment. Now physicists need telescopes and particle accelerators, biologists and medical doctors, they need imaging technology, et cetera. AI and machine learning scientists need large-scale computing. And so this is what this initiative about. And now of course we are committed to ensuring that Europe, together with other open societies, create these conditions. And we're thinking of something like a CERN or an EMBL for AI, so with shared and importantly sustainably operated compute clusters. And with institutes where the next generation of algorithms is developed, because we're, as I said before, we're at the beginning of this development. So I think CERN and EMBL, I'm mentioning them because they are examples of what is possible if politics and science pull together with a common goal. And I think to remain fit for the future, we need such an effort in AI and state secret remains before the high-level advisory body. They have just published their interim report. And we're very happy to read that they are very much pushing in the same direction and suggesting to have a distributed CERN or EMBL for AI internationally organized. Absolutely. I'm pretty sure that the International Society is very happy to see such effort already with so many significant partners putting their effort and share scene. Going back to Professor Mana, the AI research, the collaboration landscape recently have been really tempered by also the geopolitics as well. What's the ideal future that you wanted to see in terms of collaboration, AI research collaboration between Africa and the West? I think there's a lot of emerging expertise on the African continent in the field of AI and this emerging technologies. And what I would really want to see is fruitful collaboration with centers which have a rich history and excellence. And I think that's one thing we can really take advantage of. Yes, our institutions, for example, my university, Dedankimath University of Technology is a young university. We are talking there at 12 years old. So when, but there's enthusiasm among the young people on the African continent who are going to be providing a lot of the workforce for the future. But this expertise needs to be nurtured and we need to tap into, I think, countries or nations which have already had a long history track record in excellence. So for me, I feel there's a huge opportunity to have very collaborations which have mutual respect and understanding that expertise exists on both sides and we need to work with as broad a coalition as possible to ensure that we can take advantage of this important technology and not have any blind spots. Because if you just focus on one joke and the other, you run into the risk of not fully exploiting this technology. Absolutely, because one of the things about democratizing access to AI is really to prevent the relegation of global South countries merely as an endpoint in the AI value chain. I know earlier we discussed about the challenge of having access to quality data and also models which this initiative is aiming to address but also the talent issue I think is definitely, in terms of ranking, I would say it's probably even number one. I just wonder if any of you could comment down from the providing training or talent perspective, is there any strategy put in place or any early thoughts that you could share? So maybe one thing to share is there's an initiative that several of us have been involved in has been going on for a while for machine learning summer schools and we have now moved to a model where we also stream these machine learning summer schools to different places. So last time we ran such a summer school we had two locations on the African continent in collaboration with Data Science Africa with large groups of very enthusiastic and highly talented students and this is a model that we want to expand within this initiative. Anything else to add? No, I think this is a perfect example. We are here starting but I think this will be an integral component because it's not just transfer of data or knowledge but it's also transfer of knowledge and talents, right? And that should be part of this initiative and will be otherwise we won't be able to develop this together. Absolutely. Now let me ask a hard question. Like I said, this is, I can guarantee you so welcome in the international society but we all know the obstacles as well. Because it's such general purpose technology there's always potential for misuse. There's also always risks of technology transfer as well. Therefore the governance mechanism really need to be put in place and I know it's a daunting task. Again, any early thoughts in terms of governance mechanisms will be welcome. Can I start by saying that such a project does not aim to create a general governance or aim to prevent the use of AI for lesser intents or say, intents that are not geared towards the global public good but what we must do is to ring fence the use of AI for the global public good and by creating, launching, doing, implementing projects of that kind and by doing it then of course you need to answer all the traditional governance questions, dual use and so on and so forth. You need to apply to the work within the project so that it can safely develop and through that then you also acquire the insight in governance questions that you then can scale up and contribute to the general, to the global debate on the regulation of AI at heart because you've gone through the motion yourself to realize a concrete project. But as I say, it is protecting the space where AI can be co-created and used for the global commons by everyone. Right, anything to add? So clearly AI is technology that can be used for positive and negative purposes such as almost any other technology. Now we could say, we think this is dangerous so we keep it in the hands of a few. Right now this is in the hands of a few large companies. All we could say, we think this is potentially dangerous so we need full transparency and everybody has to be able to look at it, train these models, attack these models, test their vulnerabilities, et cetera. Personally, I think the second approach is more promising and this is very much what we're trying. So we're trying to increase access and I think if you look at the development of humankind it's a series of giving more access, democratic access, giving access to water, giving access to electricity, giving access to energy, giving access to the internet and now it's about giving access to compute power because this is what fuels artificial intelligence that will augment our own methods. Well said. Any questions from the floor before we conclude? Nothing, let me go back to each panelist just to say very quickly within 10, 15 minutes. 10, 15 seconds. Second. Seconds. I'm trying to prefer. What is your wish and what would you like to accomplish during the next incubation phase which I'm sure they will run about to support? You know, the professors running the show will tell you my wish, Cathy, would be that you invite us back next year so that we can report on the progress made. Absolutely, you can count on us for that. So I think my wish is that we really move concretely forward and we have done, I started on this today, is that we take use case, real use cases and start implementing them and use this to work on the governance model in parallel because as we develop our use cases we have the chance to address all the governmental issues that we will have to face that were mentioned before and I hope that this will speed up things instead of thinking a year about how a governance could look really start with the projects. The computation is available from March onwards so we actually now need to use it. We cannot wait a year and I think this is my wish that we move forward very concretely in the next six to nine months, I would say. Absolutely, it's urgent. Yes, so I completely agree with the previous speakers and maybe if I can just add, in addition, of course we are hoping for broad buy-in both from the public side, we're open to all sorts of partners but also from the private side where right now a large amount of the compute resides so we are imagining a world where also private companies will say we are willing to share a certain percentage of the compute that we have for the common good because we have problems that we have to solve together as humankind. Yeah, I'm just hoping that we can have very nice concrete examples of outputs from members of this network and also new partnerships that emerge because now different members have joined. I would like to express my gratitude to all of the panelists and to the press in the room and also online, thank you so much for your support for this extremely important initiative and we cannot wait to report back on the progress in a year's time. Thank you everyone, enjoy the rest of your day. Thank you. Thank you.