International Relations and Defence Committee
Wednesday 28th October 2026
Skip to latest contribution
This debate has concluded
(began 13 hours, 41 minutes ago)
Watch LiveSkip to latest contribution
This debate has concluded
10:35
Lord Robertson of Port Ellen
-
-
Lord Robertson of Port Ellen
Thank you. And well, welcome to this, the fourth Public Evidence session of the International Relations and Defence Committee of the of the House of Lords and our investigation into the multilateral aspects of the government, the the governance of artificial intelligence. I have to tell you that this is live on Parliament TV, and there are some people who are interested enough to listen to these, to to these things. I'll ask members of the committee before they answer, before they ask any question, to declare any interest that they might have. And I declare the fact that I am the senior counselor with the Cohen Group in in Washington, you will get a transcript of the account today.
And if you could expeditiously return that with any tiny technical changes that you might want to make, that would be would be useful. So let me start off the questioning by, by talking about the whole question of the European Union and Europe in general, in terms of the governance of of AI. It is said that the that Europe is, is, is lacking, is lagging behind both the United States and China in relation to artificial intelligence. And I wonder whether you accepted that characterization of the current debate. And if so, if that is. If that is the case, then can the European Union and Europe in general shape the international environment in order to capture whatever standards might eventually be be adopted by on a global scale?
So if we start off with Nick, back off the plane. If you wanted to start answering questions, then.
And if you could expeditiously return that with any tiny technical changes that you might want to make, that would be would be useful. So let me start off the questioning by, by talking about the whole question of the European Union and Europe in general, in terms of the governance of of AI. It is said that the that Europe is, is, is lacking, is lagging behind both the United States and China in relation to artificial intelligence. And I wonder whether you accepted that characterization of the current debate. And if so, if that is. If that is the case, then can the European Union and Europe in general shape the international environment in order to capture whatever standards might eventually be be adopted by on a global scale?
So if we start off with Nick, back off the plane. If you wanted to start answering questions, then.
10:37
Speaker 1
-
Speaker 1
Yeah, absolutely. So yes, I, I would, I would bring maybe a bit more nuance in terms of like the, whether it's lagging behind across the whole stack, right? The US and China are clearly leading in frontier AI development and in just the amount of investments and in compute, for example. But the EU leads in some aspects of the stack. And I think this is the overall nuance that we have to, to, to, to, to bring out also Europe. And that applies beyond the EU as well. And other governments that are neither the US nor China don't need to be a leader in frontier AI development to be able to shape the rules for governing it.
Right. So that's like a second aspect to, to this in, in terms of more nuance being like behind when it comes to the, the, the critical like frontier AI model development itself makes international cooperation a lot more urgent. Right? And for the ability to use other parts of the stack and other points of leverage is, is quite important. And that that will help to shape this, this international standards or norms or whatever regulatory regimes may slowly emerge. So the market size, the industrial capabilities, the talent pipeline, the the regulatory powers, as well as like the specific stack such as ASML or like the production of lithography machine and Zeiss or Siemens, which are other like technical layers of the stack where the EU is very well positioned, like the there could be actually like from this bargaining power, like a stronger influence than most would expect on how to produce and how to develop frontier AI.
And this is especially true when, like, it's in collaboration with other governments or so-called middle powers or like the UK itself. Right? I believe the EU, the UK and their partners can build very practical arrangements that protect them now and make a wider agreement easier to reach wider agreements on this international regime. I'm happy to dive a bit deeper into this, but it might be also useful to to get into the details a bit later. But like, there are a lot of precedents that have shown the bargaining power that the EU has as as maintained.
And I hope we will get to that at some point on this on this hearing.
Right. So that's like a second aspect to, to this in, in terms of more nuance being like behind when it comes to the, the, the critical like frontier AI model development itself makes international cooperation a lot more urgent. Right? And for the ability to use other parts of the stack and other points of leverage is, is quite important. And that that will help to shape this, this international standards or norms or whatever regulatory regimes may slowly emerge. So the market size, the industrial capabilities, the talent pipeline, the the regulatory powers, as well as like the specific stack such as ASML or like the production of lithography machine and Zeiss or Siemens, which are other like technical layers of the stack where the EU is very well positioned, like the there could be actually like from this bargaining power, like a stronger influence than most would expect on how to produce and how to develop frontier AI.
And this is especially true when, like, it's in collaboration with other governments or so-called middle powers or like the UK itself. Right? I believe the EU, the UK and their partners can build very practical arrangements that protect them now and make a wider agreement easier to reach wider agreements on this international regime. I'm happy to dive a bit deeper into this, but it might be also useful to to get into the details a bit later. But like, there are a lot of precedents that have shown the bargaining power that the EU has as as maintained.
And I hope we will get to that at some point on this on this hearing.
10:37
10:37
Speaker 1
-
Speaker 1
I think that the lack of technical capabilities is a mischaracterization, because this is not across the stack. There is significant technical capabilities at the layer. For example, the the . The chip design, the lithography machine, some of the elements into into the chips that are key to making the, the, the data centers that eventually train these, these very large models, right? Same with like the gas turbines that are necessary to power these data centers like the EU as quite a bit of technical capabilities that is unique worldwide in, in these spaces and therefore can use that as leverage about this.
But more than this, more than technical capabilities, I would actually question the premise that technical capabilities are, are that influential, right? Because like market access, the 450 million fairly rich relative to the rest of the world consumers that the EU market provide actually is a significant bargaining chip into like setting up safety requirements that are basically making market entry, like conditional on certain satisfying some of the safety standards that the EU may want to see shaped. This is why we have seen companies, American companies in particular very interested in in shaping the international standards, actually listening and taking a cue from what the EU preferences are when it comes to to safety and security standards.
This is not done in a in a fairly kind of like happy way. Right? It is they feel perhaps a bit constrained by the by the EU preference for for safety and reliability, but it is definitely occurring. We have seen companies adapting their frameworks globally, their framework for safety globally in order to comply with EU rules. So there is a sort of a Brussels effect here that that seems to be at play.
But more than this, more than technical capabilities, I would actually question the premise that technical capabilities are, are that influential, right? Because like market access, the 450 million fairly rich relative to the rest of the world consumers that the EU market provide actually is a significant bargaining chip into like setting up safety requirements that are basically making market entry, like conditional on certain satisfying some of the safety standards that the EU may want to see shaped. This is why we have seen companies, American companies in particular very interested in in shaping the international standards, actually listening and taking a cue from what the EU preferences are when it comes to to safety and security standards.
This is not done in a in a fairly kind of like happy way. Right? It is they feel perhaps a bit constrained by the by the EU preference for for safety and reliability, but it is definitely occurring. We have seen companies adapting their frameworks globally, their framework for safety globally in order to comply with EU rules. So there is a sort of a Brussels effect here that that seems to be at play.
10:37
10:37
Speaker 2
-
Speaker 2
Yeah. Thank you very much. Quickly by way of background my name is Dan Yan. I'm director of research at Arc Foundation. We are a philanthropically funded think tank based in Brussels, and our mission is to help you prepare for transformative AI. Yeah, I'm very glad to be here today. It's a real privilege to be able to speak. I think I have a slightly different take on this question than Nick. I would agree with your characterization that Europe is lagging behind the US and China on most parts of the AI stack. And I also think this is an important disadvantage for the EU and the UK in terms of their ability to shape the international environment.
It is often said that in AI there is no number three, and I tend to agree with that. The US is clearly in first place, China second place. And then for a long time I would say no one comes really close. Europe in particular has relatively poor positions in multiple parts of the stack. So, for instance, when it comes to the frontier models itself, the best European model is currently roughly one and a half years behind the the frontier of AI, according to epoch AI. This independent AI research, monitoring and evaluation organization. Europe doesn't own any of the big players in the chip design space.
These are Nvidia, AMD, Broadcom, all American companies. Europe does not own any leading edge fabs where the AI chips are being fabricated. All the big cloud providers are American owned. And when it comes to the physical compute on European soil, Europe only hosts 5% of that,
It is often said that in AI there is no number three, and I tend to agree with that. The US is clearly in first place, China second place. And then for a long time I would say no one comes really close. Europe in particular has relatively poor positions in multiple parts of the stack. So, for instance, when it comes to the frontier models itself, the best European model is currently roughly one and a half years behind the the frontier of AI, according to epoch AI. This independent AI research, monitoring and evaluation organization. Europe doesn't own any of the big players in the chip design space.
These are Nvidia, AMD, Broadcom, all American companies. Europe does not own any leading edge fabs where the AI chips are being fabricated. All the big cloud providers are American owned. And when it comes to the physical compute on European soil, Europe only hosts 5% of that,