Rise of AI Conference 2026 · Berlin · Meta Stage

AI Made in Europe: Where We Already Lead

Oliver Stuke, Anna Gründler, Jörg Bienert, Irakli Beridze · Oliver Stuke: Capgemini; Anna Gründler: deepset; Jörg Bienert: Soofi; Irakli Beridze: United Nations

At Rise of AI 2026 in Berlin, a panel moderated by Anna Gründler with Oliver Stuke (Capgemini), Dr. Irakli Beridze (United Nations) and Jörg Bienert argues against the narrative that Europe lags behind in AI, presenting concrete projects instead. Stuke describes Spark, a German government lighthouse project applying legally sound generative AI to permitting processes such as fishing licenses and wind turbines, built with European partners including Mistral, Aleph Alpha and Fraunhofer institutes and named winner of a global prize for best use of generative AI in government. Beridze describes AI Pol, a UN and Interpol toolkit for responsible AI use in law enforcement, already piloted in India, Brazil, Kazakhstan, Nigeria and Oman with a goal of around 50 countries. Bienert presents SOFI, an open-source sovereign European foundation model trained with Deutsche Telekom and Nvidia, planned to launch in mid-2026 and scale toward 300 to 500 billion parameters. The panel discusses trust as a European differentiator, backed by UN research from Oslo, debates regulatory simplification and AI literacy mandates, and names internal fragmentation as Europe's biggest risk over the next 24 months.

Key takeaways

  • Spark, a German government AI project for legal permitting, won a global prize for best use of generative AI in government.
  • AI Pol, a UN and Interpol toolkit for responsible AI in law enforcement, is piloted in India, Brazil, Kazakhstan, Nigeria and Oman, aiming for around 50 countries.
  • Jörg Bienert's SOFI project plans an open-source sovereign European foundation model, launching mid-2026 and scaling toward 300 to 500 billion parameters.
  • UN research from Oslo found trustworthy institutions using flawed AI are trusted more by the public than untrustworthy institutions using trustworthy AI.
  • Panelists name internal European fragmentation, not lack of technology, as the biggest risk to Europe's AI leadership over the next 24 months.

Detailed notes

Passion projects: Spark, AI Pol and SOFI

Stuke highlights Spark, a German federal ministry lighthouse project he supports that applies generative AI to legal permitting decisions, from fishing licenses to wind turbine construction, built to be fully legally sound and explainable, developed jointly with European partners including Mistral, Aleph Alpha and Fraunhofer institutes, and awarded a global prize for best use of generative AI in government. Beridze describes AI Pol, a toolkit for the responsible use of AI in law enforcement built with Interpol and funded by the European Commission, launched three years ago, tested in 15 countries and now being rolled out to five large countries outside Europe, India, Brazil, Kazakhstan, Nigeria and Oman, with an eventual goal of around 50 countries. Bienert presents SOFI, an open-source sovereign foundation model developed with seven German universities and research institutions, trained with Deutsche Telekom's industrial AI cloud and Nvidia support, with a first model planned for mid-2026, a 120-billion-parameter model to follow, and plans to scale toward 300 to 500 billion parameters.

Building a European open-source ecosystem

Bienert compares the ambition to Linux breaking the lock-in of proprietary Unix systems in the 1990s, arguing that foundation models are becoming the operating system of the AI ecosystem and that sovereignty requires open-source models, not just infrastructure. He describes the Center for Sovereign AI, which has raised more than 100 million euros with partners including Siemens Energy, Audi and Deutsche Telekom, and an upcoming second OSFM conference bringing together European open-source foundation model initiatives. He warns that the deeper sovereignty risk lies not in cloud infrastructure but in the application ecosystems that Microsoft, Google and Amazon are building on top of their models, which Europe needs an open-source alternative for.

Trust as Europe's competitive advantage

Beridze cites a UN project with the Norwegian government and recent Oslo publication on public trust in AI, finding that trustworthy institutions using imperfect AI are trusted more by the public than untrustworthy institutions using trustworthy AI, and argues Europe's democratic institutions give it an advantage here. Stuke argues 'made in Europe' trusted AI could become a brand comparable to 'made in Germany' for cars, especially for critical infrastructure and government use cases, and frames Europe's dense regulation as a potential asset for 'law as code' applications, since AI systems can be built to navigate that complexity for users in areas from insurance to procurement.

What to fix in regulation

Asked for a regulatory wish, Stuke proposes making AI literacy mandatory across schools, universities and job training. Bienert calls for simplifying Europe's more than 111 AI-related regulations, arguing a startup's third or fourth hire should not need to be a lawyer, and suggests specific rules, such as mandatory labeling of AI-generated content as Norway and France already require for photoshopped images, would be more effective than broad frameworks. Beridze cautions that regulation should support solving problems rather than blocking innovation while waiting for perfect rules, but warns that a catastrophic AI incident could trigger a wave of restrictive regulation that sets development back for years.

What Europe needs to fix in the next 24 months

Stuke names fragmentation as Europe's biggest execution gap, with successful public-sector AI use cases remaining isolated islands by industry and country instead of part of a shared system. Beridze notes that about half of the UN's 194 member states still have no AI or digital strategy, though all European countries do, and warns that failing to close this global gap risks large-scale poverty and migration pressure. Bienert calls for pooling resources instead of competing between German states or European countries, arguing the real competition is with US big tech and China, and points to the German AI Association and the Center for Sovereign AI as steps in that direction.

Quotes

Made in Europe is like a sign that comes with trust, and if the applications are used by European institutions, this gives a good sign to the outside world.

Irakli Beridze

European trusted AI can become a brand like made in Germany is, or was, for cars.

Oliver Stuke

The LLMs more or less become like an operating system of the whole AI ecosystem.

Jörg Bienert

Questions answered in this talk

What is the Spark project mentioned at Rise of AI 2026?

According to Oliver Stuke at Rise of AI 2026, Spark is a German federal government lighthouse project applying legally sound generative AI to permitting processes, from fishing licenses to wind turbine construction, built with European partners and awarded a global prize for best use of generative AI in government.

What is AI Pol and where has it been deployed?

According to Irakli Beridze at Rise of AI 2026, AI Pol is a UN and Interpol toolkit for the responsible use of AI in law enforcement, funded by the European Commission and currently being implemented in India, Brazil, Kazakhstan, Nigeria and Oman, with a goal of reaching around 50 countries.

What is Jörg Bienert's SOFI project trying to build?

According to Jörg Bienert at Rise of AI 2026, SOFI is an open-source sovereign European foundation model developed with seven German universities, trained with Deutsche Telekom's industrial AI cloud and Nvidia support, planned to launch a first model in mid-2026 and scale toward 300 to 500 billion parameters.

Mentioned

Spark, AI Pol, open-source foundation models, digital sovereignty, AI Act, law as code, AI literacy

Keep exploring

More from Oliver Stuke

More from Rise of AI Conference 2026

More on Enterprise AI & Adoption

Transcript

Auto-generated transcript, not edited. · 7,368 words

Show the full transcript

Welcome everyone. I'm very very happy and honored uh to be hosting this panel today. Um we found three people that are very very passionate about the topic of Europe and how we lead in Europe. And they are not only passionate about it, they are driving it. And uh I'm I'm very very pleased to introduce my panelists today. We have Oliver. He's with Chip Capgeemini for 27 years now. Um no in the public sector for 27 years. He's been uh doing a lot of digitization projects and recently AI projects. He has been supporting the initiative that um state secretary Yatsenbeck mentioned Spark and he will talk a little bit about that. Um I'm very happy that you're here.

[snorts] Then we have um Iraqi Beritzia. Um Iraqi um has been building up uh the uh research center for AI and robotics at the United Nations and uh in particular at the United Nations I have to read this out. United Nations International Crime and Justice Research Institute. He has now 25 members in his team researching on AI and especially ethical AI. Um, and this is very very important when we think about what is the police doing and obviously we want to we want to use the AI the right way, right? and he will be talking about uh the initiatives he's been working on and especially AI Paul very happy to have you here and he's bringing a global perspective to our panel and then last but not least we have Jük Jük um is is a very popular figure in the German AI scene he has been founding the German AI association um and he is now um working hardly so that we have finally a European language model, a German language model, a European

one. >> I'm hardly working. Yes, >> you're hardly working on it. Now you're working on it. I see this is going to be a fun panel here. And um something that maybe not everyone knows is that he uh founded a startup and has lived in the Silicon Valley for some years. So he uh does not only bring in the German perspective but also the perspective of someone that has been living in the US and building a company in the US. >> Thank you. >> Okay. And maybe maybe I say that in the beginning. So the narrative around European AI is is very much a doomsday scenario. Uh we heard it already this morning in one of the talks.

So we are lagging behind. Uh we don't have the infrastructure. we don't have the investment. But uh we hope that after this panel you say this is not true because we have a lot of very beautiful examples and we want to flip the coin tell you what's already there and um yeah um find out what are the strengths we can build on and uh with that let's get started. Um my first question uh is um actually a little bit of a personal one uh to all of you. So in the when we uh prepared this panel uh there was so much passion in the room. So I my first question is actually what is a European AI in initiative you are very passionate about or something that you are yeah that made g gave you a wow feeling recently and I want to start with you.

>> Uh thank you. Um I would underline what uh Thomas Tombbeck introduced uh to you earlier today um the Spark project and I understand why he's rightfully proud of that because the BMDS uh has been awarded the the the global prize of best use of Genai in in government. The reason why I am amped uh uh due to this project is for two very distinct reasons. First of all, it's the first AI application I'm aware of that is 100% legally sound. everything explainable and I think you understand what it means on an engineering perspective and the second one is it was a joint effort from the European AI of our god mistral um supported it decided alpha three frowner institutes and a few more and and that really is what gets us forward isn't it to really collaborate on one distinct project and show what's possible within Europe >> maybe you can tell us briefly what Spark doing exactly to the people that were not there in >> Yeah, you're totally right.

So, um it's a BMDS lighthouse project that they um started with a few years ago and that really kicked off roughly one year ago that applies AI to um and in a law is code context to support application and permitting processes from fishing licenses to construction of wind turbines to fusion reactors. if that becomes a a thing um by again applying legally sound the German administrative law corpus to whatever it's being shown and it works and was awarded the awarded the global prize >> nice thank you so much Iraqi what is your passion project when it comes to European AI >> right Anna thank you very much thank you and uh first of all I'd like to thank uh Fabian and Veronica for inviting me here again over the 10 years it's my third time being here and it's really fantastic to see how much has grown this initiative and how many valuable conversations we've had over the last 10 years.

Now um the at the United Nations one of the top priority right now is the governance of AI to ensure that AI solves problems and it does it in a responsible way and this is probably the declared one of the most urgent issues of our humankind and UN fully recognize that se so we don't have a global governance structure and sectorally we're attempting to govern the use of AI in a respons responsible way. My center leads the governance of AI in the field of law enforcement and we've been doing that for last 8 to nine years now. Together with Interpol, we created something what we call the toolkit for the responsible use of AI my law enforcement.

That's a governance. It's not a principle. It's a governance guide book which is which can be or is applied now to the countries and the entire project was financed, funded and led together with European Commission. So this is an European product which is globally applicable. The toolkit itself we launched it three years ago at the science conference at Interpol. It was tested by 15 countries and is applicable to all 194 UN member states. Last year we launched the second phase of the initiative which is a quite a massive global initiative which is called AI pole and AIOL is meant to translate the toolkit into the standard operational procedures of the large countries which are outside Europe now.

So bringing European product outside Europe for the governance of AI and the law enforcement which is probably one of the most sensitive areas of AI application because everybody's scared about use of AI by police and rightly so at the same time police should be using AI without that you won't be able to solve problems but it should be done in a responsible way. So this is what we're doing. AI Paul now has five large countries where already we are implementing this. This is India, Brazil, Kazakhstan, Nigeria and Oman. It was deliberately chosen the five countries across the world so that they can serve as a multiplier effect that other countries will participate. So after that we envisaged around 50 countries will be part of that.

So we'll be implementing this right now. We already have done the organ entire readiness assessments AI readiness assessments in all five countries and we're already implementing this Europeanmade guide book Europeanmade governance instrument into into this very sensitive application. So this is one of my passionate projects. I have a lot of them our center but but this is something which really gives a good example how something which was created with European values embodied obviously with the whole universal principles of human rights and uh and and UN declarations but it is still driven with the European sort of made flavor to it and is being exported everywhere else in the world. So you're basically bringing eur European values into the world >> and I think this is extremely important in that sense so that um that type of frameworks work.

We see that and countries are very passionate to uh adapt it and this is a really good example of collaboration of European Union, United Nations and Interpol and the recipient countries in the world. >> Wow that's really really impressive. Okay, York, now to you. What's your passion project? >> Well, on one hand, I think we we have a lot of great activities and and and startups in in in Germany like for example, Black Forest Labs. Yeah. We always forget that uh the team of Black Forest Labs really invented uh the image generation. Yeah. Which is all over right now. Uh and they are quite successful. Yeah. Unicorn startup. We have other unicorn startups like DeepL still the the leader in in terms of uh AI based translation.

Uh we have a lot of other unicorn and and and leading technology star startups like Netn or Hastake and others. uh so uh there is a good basis and what I'm really personal uh passionate about of course is is my my own project that I'm leading called Sophie sovereign open source foundation models where we are developing with together with uh the seven leading universities and research institutions in Germany uh an open-source uh sovereign uh foundation model a large language model because uh We honestly believe that this is the basis for AI sovereignity. Yeah. the LLMs more or less become like an operating system of the whole AI ecosystem and uh we are doing this uh with this large team but I are very clo working very closely and and well together also working together with the teams of do telecom because we're doing the training on the industrial AI cloud and do telecom and with a great support of Nvidia uh so there there are meetings every few days between these teams.

They're working very closely together and for me uh after a long time this feels really like as we say in Germany things are growing together that belong together and that we are really creating some kind of momentum in order to achieve uh sovereign uh sovereignity in the AI space uh for Germany and Europe. >> And when is the model live when can we use it? Uh we are planning to launch the first model um later this year in June or July and one or two months later uh a second model with 120B and are also in conversations regarding uh further investments uh to extend the model size to uh 3 or 500B. So this will be really leading edge models and um I think that can be uh a very good basis for what is also always called industrial AI because most of the industrial AI applications or applications that we have just seen from uh uh from finance informatics uh uh most of the applications are based on on large language models even agentic

systems you can't if if you're on it if you are not able to build these applications on uh sovereign and and open-source foundation models. You are not sovereign at all. >> Yeah. And talking about open source, so you have now the new model out there. So Oliver can use it. Where does he find it? Is it on hugging phase or where where will you provide? >> Yeah, it will it will be on hugging phase. We will start with some pilot customer. We are getting huge interest from the industry from from the uh from a lot of companies. uh and after the pilot phase it will be also available on on hugging phase >> okay and if Iraqi now wants to use spark when where can he find spark because we heard this morning it's also open source or partly open source right >> yeah exactly um so it's a it's it's a longer period of time and they're are currently um putting one module after the other on on centers the center for digital sovereignity

>> and and open code um I think the um the most valuable um modules will be available from July 1st around that. >> Yeah. >> And and given that you asked, we are already in discussions to apply them in Portugal, in um Norway, in Netherlands. So, uh it would be quite fitting if the United Nations used them as well. >> Mhm. No, certainly. I mean uh at the at the UN we have numerous platforms where we do bring different technological solutions specifically designed to contribute to the UN sustainable development goals and other impactful issues. So we certainly happy to facilitate such things. >> And now um one now we talked about a lot about the things that work well from your perspective.

Maybe what is something that we can learn from others? So I know you've been in the Silicon Valley. >> Mhm. >> What do you think we in Europe can learn from others? Um I think especially in regards to Silicon Valley we can learn to have a brighter clearer vision and to follow this vision in regards to technology and its appliance. Uh also with with some braveness. Yeah. This is what what we are missing here and and with some risk takingaking. Yeah. Sometimes it seems to me when we are looking at the future or developing strategies for the future Yeah. We are we are looking at the status quo. Yeah. And then maybe for the next three years looking at a linear progression but the progression in this technology is exponentially.

Yeah. And I I think we all uh politics and and corporations should consider more what's coming in in terms of technology development and in terms of uh being able to compete in a global economy. >> Yeah. and and then develop strategies and follow the strate strategies uh uh with with with more investments with more willingness to innovate and and this is something there where Silicon Valley is is far ahead and we definitely should catch up. >> What do you think Iraqi would you agree? Do you have a maybe a more global perspective? Well, I mean from the global side, what UN does and what the center for AI and robotics which I'm uh heading does is that we create numerous platforms for collaboration and learning from one another from different parts of the world.

For example, six ago six years ago I launched a project which is called AI for safer children. This is designed to use AI to combat sexual exploitation and abuse of children online and directly help law enforcement agencies to solve the particular crime area. Six years ago when we interviewed law enforcement agents who are working in that uh in this particular area, none of them knew how to use AI. Right now we have 128 countries part of the global hub using 100 applications or more which is there by the way half of them are being developed in Europe and and we are directly supporting the um training of these officers to use AI on daily basis to in two weeks time we're hosting a training in Madagascar for four 12 African countries directly supporting them to use AI to solve this crime.

This is very practical, very efficient. Uh something what works. Over 3,000 officers been trained by now for last two or three years since we launched the training part of the of this. And there's a huge collaboration happening around the world. The developers who are developing these applications, the AI tools and I I mentioned it half of them are in Europe have made a pledge that they would make the AI tools available for free for the running cases for the countries who cannot afford it upon request and we facilitate uh those things. So this is enormous collaboration in the world and something which is needed because the sexual exploitation abuse of children online is a hugely growing area.

Unfortunately more kids use internet more predators use internet. So a lot of exposure is happening. Plus what is added now is that AI generated videos and images which in some cases are very difficult to distinguish from the real ones and uh and officers have hard time to actually find out about it. So what you're saying is what we can learn is collaboration or we should even do more collaboration to bring the technology to more people. >> Yeah. Yeah. Right now I mean we have a sort of quite a dangerous digital divide in the world and that type of collaborations are trying to fill in these gaps and from the UN perspective we we try to facilitate this type of interactions as much as possible.

>> Yeah. >> And this is so much needed. uh we run numerous AI literacy projects for children for others and uh and European involvement is tremendous in that so we do need to recognize that and when we mention about Europe being behind I mean I can comment a lot about this it's a I think quite a sort of a European self-declared myth and uh and I don't know whether it's done deliberately just for tactical reasons or really believing u people really believing that but uh but the game is not over yet game has just started So, how can it be lost when uh we're just starting the game? >> Oliver, when you hear that being behind is maybe a tactical myth or a myth, what's your what's your view on this?

Would you agree? [sighs and gasps] >> Well, first of all, yes. Um secondly, what what just became a lot clearer to me when you were presenting your your your cases where we can learn from from from others is that just the other day I read in the the newspaper that due to an interface survey um if Germany is the fifth most favorable place to work in AI. Um so maybe just maybe the systemic challenges aren't that big. Maybe our challenges are more on the confidence and communication side and that is something we can learn from others to I I know we're German. It's on our DNA to be a little discon but honestly maybe maybe that's a thing we can learn from others to focus on the bright side which is the motto of today's conference isn't it?

We are already leading in many places uh and to work with that. And the other thing is um we are aware of this 50 billion investment of Google into anthropic and the next 100 billion investment will be in the newspapers next week and so forth. Maybe the thing we can learn from that is that we don't have to compete in the same race. We can find our own playing fields where we are strong and make up our own rules like like you're already doing with with your own sovereign language model like you're doing with your work. We don't have to be in the same game. Nobody forces us to do that. >> So then let's talk about that.

What is our playing field? So we talked about trust. I heard a lot of trust um sovereignity. So what what is it? where because we want to find out what is the strength that we have to that we can build on what is it maybe maybe you >> um I think we have a chance and we have a good chance if we are really joining forces in Germany and Europe >> uh because we don't have the deep pockets like US big tech um but if if we are working more closely and and very well orchestrated together there is a chance and if we are establishing some kind of open source ecosystem. Uh and then there's a good blue blueprint.

Yeah. In the '9s uh there were um several vendors for hardware with their own Unix operating system. Yeah. And so if you as a company choose uh one of these platforms, you're more more or less already stuck into the whole IT stack. Yeah. And then there was then there Linux came up and and democratized the whole area. Yeah. And and today it's it's a given. Yeah. And [snorts] maybe if we achieve this with the foundation models uh as operating systems of the AI ecosystem. I think this really could be a good approach and um we are heavily working on on on both aspects. On one hand uh we founded the center for sovereign AI in order to join forces uh from uh research and and institutions.

We are getting funding of more than 100 million euro yeah to even support and launch open-source projects uh working closely together with uh corporations like Seammen's energy, Audi, EABG, the Deutsche Telecom and Matana. still some spots left. So if you want to join us, yeah, just just contact me um we're look looking forward to working together with you. And uh so this is one initiative that we are actually uh doing and uh forcing. And on the other side, we are really trying to bring the uh open source ecosystem together. In one month, we have the second OSFM conference for open source foundation models where all the European initiatives that are building open source foundation models join for an exchange uh of of ideas of experiences um uh and with the vision to establish something like a Linux for uh foundation models.

>> That's impressive. Thank you so much. And hopefully you also think about the everything around the the open source model the orchestration and everything that is that is that's also part of that right >> definitely definitely yeah because we see uh I think the uh the threat of uh for sovereignity like in the cloud services yeah it uh in the cloud services it's not the infrastructure of the service but the whole ecosystem of applications and the whole environment that uh Microsoft Google and Amazon built on top of And we see this happening right now. Yeah. If if you look at the weekly or daily announcements of Anthropic and and Google and Open AI, they are starting to build a whole ecosystem of applications and features uh agents around their systems, around their base systems.

And we have to really build an alternative uh uh in regards to this ecosystem not only the base foundation model but also being this uh uh establishing this as a basis for an open source ecosystem of other applications agent systems and whatever will come. >> Yeah. [clears throat] Iraqi would people pay for that if they had a trusted ecosystem sound solution that comes from Europe or would they say no I still go with the others >> look I mean a lot of advantages have been mentioned this morning infrastructure data talent and all of this is true I want to focus a little bit more on the trust side uh we did research the issue of trust quite a lot we are having a current recent project funded by the government of Norway on how to ensure that public trust the use of AI by law enforcement.

I mean we are working in the area of that. So so we focus on this but but on the trust side uh we last week in Oslo we launched uh recommendations and quite a large publication actually over that it is available on our site. first time UN agency together with the UN member state tried to examine thoroughly the issue of trust and and come up with the practical recommendations how to ensure that public trust because end of the day all these applications which are being developed all the AIs are developed is for public to use it to solve certain problems and if public accepts that or or doesn't accept that that's a gamecher and trust issue is tremendous in this especially in the sensitive parts.

We talk a lot about the data breaches and an anonymization and all of that stuff but but end of the day what we found is that there are two main aspects there on from from the trust side although I cannot summarize entire 100 pages of of the publication but two main things one is that the um trustworthy AI or developing the trustworthy AI from the design side is not enough it's also the institutions who are using the AI is very important So um the the study found that even if the trustworthy institutions using poorly or bad AI is still more trusted by the public rather than on the way around if not trust trustworthy institutions using trustworthy AI.

So on the both sides a lot of advancements need to be made. public needs to uh trust the use of AI by the institutions and I think that in that case Europe has quite a lot of advantage of democratic institutions who are developing and using AI and public trust the use of AI and I think this could be a quite a big advantage in the future and uh kind of a a model which other countries may want to apply or other other parts of the world may want to apply. So basically made in Europe is uh like um a sign that comes with trust and if if the applications are used by European institutions this gives a good sign to the outside world as well.

Certainly and this this does create some sort of examples and models where others can also follow and and uh in some years to come I believe that that part will be very much of an advantages portion of how public will be trusting the because there will be many applications there will be numerous applications we're going to make many mistakes as well and uh and the game changer may be in the issue of how much public will trust the user of AI by especially the public institutions. >> Okay, Oliver, that's your business case. Right now you can go scaling Spark. >> Um so as you just said can become a sign. I really think that European trusted AI can become a brand like made in Germany is or was for for cars and whatever but for for two or three main reasons that are not so easily copied by others.

Well, first of all, they don't prioritize that. But secondly, well, and you made the point, even if if the data center was in in in Germany, it would be only half of the solution if the whole large English model would be trained on well pamphlets from the Chinese Communist Party and Fox News alone. That well, I assume you do it differently with Sophie. Um, thanks. Um, so um if if you start too late, then you cannot achieve the whole trustworthiness you're you're aspiring to. So and if we focus on that, if we operationalize that, then it can be huge selling point. Not for every application. Obviously there will be million applications where scalability and afford the price uh is is is the key driver but there will be very important critical ones around critical infrastructure where in government where that is imperative that's the one and secondly um I I get it that all of these regulations are tedious we have in Europe and they they keep us from being fast but sometimes It's not

about being fast, about endurance and being resilient to changes. That can be a huge selling point. So if you if you take this AI system, you can stick with that for the next 10 years. I know that Lage models are evolving on a daily basis. So you might want to change that. But the whole system itself is resilient based design and as with with regards to Spark because I think that's a very important notion to to be aware of. um the the the the whole maze of regulation can be overwhelming. I get that. But if you have use cases where the maze is the context and the maze is the um is the key use case like for law as code, it can be huge asset what we have here in Europe that this whole behemoth of regulation.

um if if you if you put AI into that to to find its way and to automize permitting processes, they know their way. It's it's it's it's like an operating model for AI, including guidelines and including um traceability and that is something that can be an asset as well for all of these laws code applications that are out there from insurance to procurement to governmental services and so forth. >> Okay. So regulation plays a key part in that. I know um we have mixed feelings about regulation in this group. Um now it's the 10th birthday of Rise of AI. Let's make a birthday wish. What regulation um would you change if you could? Um yeah, let's start with you.

>> Okay. Um well, I'm not the biggest fan of regulation for the sake of well keeping me from doing things. Um but if I can make a if I can make a wish. Um well we are all aware of this article 4 discussion from the AI Act and around AI literacy and I get the point but it's not using the big leverage that there is why don't we make AI literacy mandatory in every school education every university studies and every job training to ensure that not not only people working in a bank or whatsoever are aware of that but that the the whole population how knows how to deal with it because AI can't decide anymore what's real what's not and I I work with AI and and the other thing is to really apply it and to make it our nature so why don't we make it mandatory for everyone >> okay if somebody from BMDS is still there um one suggestion here do you have a suggestion Iraqi >> you know from the

United Nations perspective uh I don't have any particular regulation to sort of pick on or or or choose but the the main essence here is at which I always say that AI should be used. >> AI is something which solves problems and AI should be applied as much as possible but at the same time done in a responsible way with the human rights compliant manner. Now those are big words. How to do it? Of course there are practical ways to do it. You will make mistakes but at the same time you'll be solving problems. um there is a fine line there of hindering innovation and at the same time not solving problems. So countries have experimented and there are sort of places where there are less regulation.

There are some which probably is over overregulated. I think we need to figure it out as quickly as possible how to ensure that that we will not solve we will not leave problems unsolved because we are waiting for the regulations or for some sort of loss. But at the same time we need to ensure that the fastly developing field of AI which is becoming stronger and stronger is not going to reach a stage where we will not be able to control it. And at the UN when we working on the governance issues, this is one of the top priorities to ensure that we will still be able to govern this powerful technology which is every day changing and becoming stronger and it's not going to govern us and and this will not be the last technology we ever invented and so on and so forth.

So I don't want to get into the catastrophic part but at the same time uh we all do realize that in our not only in our lifetime but last 10 years 15 years so much has changed. It's totally different world we're living now and I was asked a question some time ago that um how do you see next 10 years and I said look at what was 10 years back and then multiplied maybe 10 times and this is what's going to be or even more so completely unknown so we are in an kind of unchartered territories and that's why uh uh there will be many many hurdles and variables there whether on the global level because we don't have any kind of a global charter on AI at the moment uh which at one point will be definitely needed and there's a lot of work done on that on that sense but nationally also so that uh regulations should support the innovation and should support the solutions of the problems and uh and I think

that's a baseline >> solving problems is a good uh a good point for you what's your view what would be your birthday wish on on the people that regulate >> I think uh uh I would wish for more simpl ific simple simplification of the whole regulation ecosystem. Yeah. And in Europe we have as far as I know more than 111 uh different regulation law packages 113 are up to come. Um and this cannot be handled by anyone. Yeah. And it doesn't make sense that the the third or fourth employee of a of a startup has to be a lawyer. Yeah. In order to figure out what's what's possible and not. And uh um in terms of if if if you talk to a lawyer that is very strict and very risk averse, he would advise you it's not possible to to develop and train a large language model in Europe.

Yeah. Because of copyright, uh data security, personal data issues and so on. Yeah. So I think we have to be far more flexible, far more simpler in the regulation ecosystem and far more flexible. Yeah. because uh we are we are too slow. Yeah. The uh I I was uh observing and uh partially taking part in the whole process of the uh AI Act from the beginning and it it it's it was really a nightmare especially for engineering people. Yeah. This what are they talking about? Yeah. And uh we we need we need to have a framework where we can really adapt quickly to uh the recent technology developments because they are coming much faster uh than we can and we are lacking behind with the with this way of doing regulations.

Yeah. And um we have 111 uh uh regulations but I think the one of the most uh uh one of the biggest threats right now which is deep fakes and AI generated content. This could have been in easily put into a law just saying you have to identify AI generated content like uh Norway and France already did years ago with uh photoshopped content for social media platforms. Yeah. And this is something that we have to be more pragmatic, more more quick uh more faster and more flexible about and and uh in in a way that everyone understands and can apply even in a startup. >> Yeah. Yeah. I mean I can uh briefly speak from the perspective of a startup.

We were also struggling with it. Now I I think we found our way. But what I found quite interesting, I was in the US some weeks back and there I learned that they are saying, "Oh, we all want to adopt the EU AI Act." Because at least they have something in the EU and there it's every state is different and um the people that um build AI applications, they don't know what to do because it's different everywhere. So yeah, but maybe a more pragmatic approach. Yeah, you want to add to this? >> Just as a present, you can also use Spark to check whether your your your your business model or whatever is compliant with all of the regulation around it and you can do it for free.

>> I was just going to say that now you have AIS which can actually help the uh startups and founders to comply easier than before. So you do have >> I mean a general idea of having a risk based approach is something I think very meaningful. >> Yeah. And now we just have start start working with it. It is meaningful because we have experienced lot of things in the like in the UN. We have experienced many different occurrences when we were at one point in in my career when I started um earlier I was working in the field of disarmament on the on the in in the organization for the prohibition of chemical weapons and I was in charge of the chemical terrorism related issues and until we had some use cases where real chemical weapons were used some of the risk adoptions or some of the sort charters were difficult to adopt it and sometimes you have to see something bad happening to do it and and that's why it's so important to look

at the risks to ensure that nothing bad will happen so that this is going to hinder the entire sort of ecosystem. Imagine that something catastrophic happens with the use of AI soon this may actually create a whole hurdle of new regulations which may sort of put a lot of development behind as well or or or stop it for four years. >> Yeah. >> So this is something that you also want to avoid. >> Yeah. So I'm looking looking at the time. Okay. We have six more minutes now. Okay. Final one or two questions. We were talking about the beautiful things that are happening already. The great projects that you are all working on. Um and that trust is like the our export um yeah uh sign uh that that uh really can also help us to create additional revenues.

But what is the one thing Europe needs to fix in the next 24 months? What do we need to fix now? And if we don't fix it, the window closes our leadership window. >> I have an idea. Well, >> okay, then go first. Um I I I think our biggest biggest gap in execution is fragmentation. Um to talking about the the successful AI use cases that are there in the public sector alone, they are still not part of a system. They're all small islands and that's one industry and one country. If I look at Europe as as a whole, maybe you can confirm that everybody is still occupied with their own agenda. And so if you're if you're doing well single individual races against bigger competitors, maybe that's not the right approach.

But given that the BMD is uh celebrating their first birthday today, tomorrow, maybe that's something that they can um help us with even more. >> Yeah. Or also something where we can learn from the UN because you have to collaborate all the time, right? So um >> okay we work with all countries in the world and I give you an example um right now uh UN has 194 member states out of that 50% of the countries don't even have any AI or digital strategies right we are in the 2026 and half of the countries none of these half of the countries in Europe so in Europe all countries have some form of digital or AI strategies is individually and collectively as well.

So the countries which don't have it may actually end up in a quite a difficult situations because what AI strategists do is that they identify what countries have it where they want to be so and how to get there and how to allocate budgetary resources to fill in these gaps according to their national strategic sort of priorities and and and this is a vision and this is something where it is applied. So some speakers said that yes in some cases this vision exists but um countries don't do anything about it or or slow in that yes I admit it probably it happens but at the same time once this vision is there and you have the whole infrastructure ecosystem so you can actually push this implementation but if you don't even have that vision it's a quite a dangerous situation from the global perspective what we see is that if this is not corrected we may end up living in the world where four billion people will be living in a full poverty and and

and uh where they going to go right uh we had this kind of situations in Europe where uh migrant crisis created political issues and these were very small crisis compared to imagine that you will have one one billion people coming to the European door or any kind of door so so what you going to do about it right and this is that's why it is very important that that type of development, that kind of advancement, knowledge advancement, AI or sort of technology literacy or AI literacy is applied across the board so that people will be able to build their own solutions and develop more or less evenly so that uh we will not have that uneven dangerous world where potentially four billion people will be coming here or anywhere else.

So >> Yak what do you think? What do we need to do in the next 24 months? >> Yeah, I think as as mentioned earlier, we have to join forces. Yeah, there's there shouldn't be any comp competition between ministry A and B or University C and D um or Bavaria, Northr and Westfailia or France or Germany. The competition is in China and uh in the US in the US big tech and uh we are spending all over Europe a lot of money. Yeah. But we're spending it not effectively and wisely. So if we need we need clear strategies and the willingness and also ambition yeah to move things forward together and I think then we have a clear chance.

Uh there are things moving into this direction but not quickly enough and uh as as said initially we we we are trying to to help with this with the foundation of the German AI association now with the foundation of the center of sovereign AI and we would like to everyone to join us and join others in order to bring uh bring all capabilities together and really to to be able to build a sovereign ecosystem. system. >> Thank you so much. What a nice ending. So, um, thank you so much for bringing in your perspectives and I hope all of you learned a little bit about the initiatives that we have in Europe that work and that maybe this narrative of Europe being behind is a narrative that we tell ourselves and we should be brave.

Go out there, spread the word, spread our technology, um, work together to make Europe an AI comp continent. Yeah. Thank you so much. >> Thank you, Anna. Thank you.

Rise of AI 2027 · Berlin · May 2027

Be in the room for the next one.

Ten years of talks started as conversations in one curated room in Berlin. The next edition brings 300 selected leaders together in May 2027.