Business and Trade Committee
Oral evidence: UK-US Atlantic Declaration, HC 1705
Tuesday 11 July 2023
Ordered by the House of Commons to be published on 11 July 2023.
Members present: Darren Jones (Chair); Alan Brown; Jonathan Gullis; Jane Hunt; Ian Lavery; Anthony Mangnall; Andy McDonald.
Questions 1 – 44
Witnesses
I. Professor Emily Jones, Associate Professor in Public Policy, Blavatnik School of Government; Chris Rogers, Head of Supply Chain Research, S&P Global; Michael Birtwistle, Associate Director, Ada Lovelace Institute; Dr Zeno Leoni, Lecturer in Challenges to the International Order, King’s College London.
Witnesses: Professor Emily Jones, Chris Rogers, Michael Birtwistle and Dr Zeno Leoni.
Q1 Chair: Welcome to this morning’s session of the Business and Trade Committee, where we will be talking about the Atlantic declaration, which was signed by the Prime Minister and President Biden a few weeks ago as part of ongoing UK-US collaboration. We have four witnesses with us today. We have Dr Zeno Leoni, who is the lecturer in challenges to the international order at King’s College London. Then we have Chris Rogers, who is head of supply chain research at S&P Global, Michael Birtwistle, who is the associate director at the Ada Lovelace Institute, and Professor Emily Jones, who is the associate professor in public policy at the Blavatnik School of Government. Welcome to all four of you and thank you for coming to the Committee this morning.
My first question is to set the scene in the context of the Inflation Reduction Act, which, as this Business Committee hears often, is having a big impact on investment and business decisions in the US and across the UK and Europe. How does the Atlantic declaration fit in that context? What does it mean for the UK in the context of the Inflation Reduction Act and our relationship with the US?
Professor Jones: I am going to defer to Chris on supply chains and IRA. I am happy to pick up the data questions, which was what I was going to focus on.
Chair: It is just in the round. You do not need to go into specific detail.
Professor Jones: At a headline level, it is incredibly important for us to really think through our strategic relationship with the US at the moment, particularly in the light of the IRA and industrial strategy. We need to think very carefully about how we position the UK in the global economy, where we are going to seek competitiveness. We essentially have a subsidies race going on now between the US and the EU. We are going to have to make some very careful choices about how we position ourselves, not only to not fall behind, but also to make sure we are right at the competitive edge for the next 10 or 20 years, and think about the green economy, climate change and how we transition within that increasingly competitive marketplace.
Q2 Chair: Michael, is the Atlantic declaration a welcome and useful contribution in that context?
Michael Birtwistle: I cannot comment on the effect on inflation and so on. It presents an important opportunity around the prospect of the UK seizing the moment on artificial intelligence and the opportunity to position itself globally with regards to AI as a sector but also AI governance. I am happy to speak more to that on perhaps more specific questions.
Chris Rogers: The first point to note is that the Inflation Reduction Act is not really about reducing inflation. It is about promoting manufacturing. In that regard, the parts of the Atlantic declaration that play into the manufacturing elements provide an important connection in industrial policy between the UK and the US. We can discuss this in more detail later on, but we are in a globally competitive manufacturing market. Clearly, that covers a lot of different industries. It is very important for a global Britain to be a part of these different elements of connectivity.
The Inflation Reduction Act, as I will talk about it later on, is very much focused on the green economy, particularly around the electrification of transport. The transport industry has a great history in the UK and so maintaining those connections is very important. In the round, it is about connectivity within global manufacturing supply chains.
Dr Leoni: Personally, I think that we should read the Atlantic declaration more in the context of the Defense Production Act. This is the most important development, and context here is really key. We live in a time of strategic competition between the United States and China, and the UK somehow finds itself in between this strategic competition. The United States in particular is pursuing a new 21st century sphere of influence to decrease strategic dependencies on China, and it wants to do this with the UK.
What is really key, which involves both benefits and risks, is the fact that the Defense Production Act and the domestic source clause are to be understood in the context of the ability of the domestic industrial base to supply materials and services for the national defence and, crucially, to prepare for and respond to military conflict and foreign threats. We should look at it through these lenses.
Q3 Chair: The way we will structure the session today is that we are going to ask questions about particular parts of the Atlantic declaration, which means that each of you will have to sweat it out for a few minutes in relation to your area of interest. We are going to start with critical minerals, so we will come first to you, Chris Rogers. The Atlantic declaration made quite a big deal about a new UK-US agreement on critical minerals. What do you think that would mean specifically for UK businesses? Are we producing enough for that to be of benefit to the UK economy?
Chris Rogers: In general terms, the answer to your second question is not yet. The answer to your first question is that it is important to provide an export market. As I mentioned a few moments ago, this is about a global shift in manufacturing supply chains, particularly in relation to electric vehicles. If UK exporters want to be a part of that, they need to be a part of the global agreements that are going on. I do not mean a full global agreement, but there are a lot of these bilateral agreements as we have, for example, with the Atlantic declaration and the critical minerals agreement within that.
For UK exporters specifically, we are looking here at potential and current producers of products of cobalt, graphite, lithium, manganese and nickel, among others. These are not huge industries in the UK at the moment. I was going to say that many of them are still nascent. I was in Cornwall recently and saw the copper mines there. We have been doing copper for a little while.
In terms of things such as lithium and so on, these are all minerals that are available in most countries. The trick is being able to mine them economically, to process them efficiently and in an environmentally friendly manner and then have readily available markets for them. In that regard, having a critical minerals agreement with what effectively is the largest automotive market after China, in the US, is really important for UK exporters.
It is providing that market not just for cars. Obviously UK vehicle manufacturers may become eligible. Some of them are not at the moment. We have a little way to go on battery manufacturing. Being right at the head of that supply chain, in terms of the minerals, is an area where we can tangibly develop in the UK in the short term. “Short term” is a relative measure when it comes to mining and processing, but this is a start in that process.
Q4 Andy McDonald: Chris, can I go a little further on this? We are starting to scratch our heads as to what the point of this agreement is. There has been some speculation that it might be similar to the agreement that exists between the US and Japan. That does not contain any new binding agreements. That does not bind anybody to any specific course of action. It is not a free trade agreement and it comes into effect on signature, so there is no ratification. Does it take us any further in reality?
Chris Rogers: Compared to a status of zero, it is more than zero. However, you are absolutely right to point out that this does not have an X thousand tonnes a year of shipments by X date, which is what you would want to see as a business investor, for example. It is an intent to strike an agreement rather than an agreement itself.
Your point with the Japanese agreement is absolutely right. That is still in the process of implementation. For those of us who have worked—as I think all the members here have—in global trade for a while, you start small with these kinds of agreements. They take a long time to go through the process of being turned into legally binding documents with signatures and so on. It is not just Japan that is having these negotiations at the moment. It is also the European Union.
Everybody is trying to get their piece of the pie. Unfortunately for the UK in that regard, there are already plenty of countries that are already at the table with the pie because they have these free trade agreements, including Chile, which is the largest exporter of lithium already. You are competing with a very large country in that regard.
If nothing else changed from today, so we have the critical minerals agreement in its outline form, that will not change anything. It is the opportunity to move from a critical minerals agreement to the fully ratified documents. Importantly, without getting ahead of ourselves, at some point the US Treasury has to say, “This is something. These products are eligible”. I do not know how long that process will take. It is not a work of moments. I would not expect it to be done within a few months, certainly.
Q5 Andy McDonald: Are there any downsides to pursuing this course of action? Are there any consequences for relationships elsewhere from pursuing this?
Chris Rogers: We certainly have not seen other countries, regions or territories say, “You are working with the US. Therefore you cannot work with us”. In that regard, there is not. Does it preclude other forms of free trade deal? No, it does not. As far I know, there is nothing planned within the Atlantic declaration that precludes that. The issues around signing wider FTAs with the US are outside of scope a little bit, but it does not preclude that.
There is a risk that business decision-makers will say, “We want to see precisely what this deal looks like before we start signing commercial terms”. The longer it takes to reach an agreement, that is clearly not a good thing. In our experience, companies can deal with more or less any regulatory situation as long as they know what it is. It is uncertainty that can be a challenge. There is no doubt that there is demand for the product in the US. It is a question of whether it can be made economically available when other countries have access to IRA funding terms that maybe UK companies do not.
Q6 Alan Brown: Chris, you said that this is an intent to strike a deal and actually getting a deal agreed takes a long time. Can you give an indication or examples of how long these deals typically take, even the US-Japan one, for example?
Chris Rogers: The US-Japan deal is still ongoing and under negotiation. That was announced in March. We are on somewhat unusual ground here because we have come out of a period of 20 years or more where trade deals are full trade deals, so they cover everything. They cover all products and all tariffs. They cover services—the whole lot. We do not have a lot of experience of these bilateral industrial relations deals and how long they take.
Trade deals between partners can be turned round relatively quickly where the terms are known, but “relatively” is a year of negotiations and a year of ratification. As I say, it is not a work of months. Importantly, the US is having these negotiations with other regions, so it may well be that Japan goes first and that is a template for what is done with the UK and the EU. The main thing is maintaining momentum. I think that there was a mention from the Prime Minister that further discussions will be had in October. Clearly, leaping three months down the road each time suggests that this is a work of months into years, rather than weeks into months.
Q7 Alan Brown: Would it be fair to speculate that it is probably more advantageous for the US to get a deal struck with the EU first, before the UK, given that it is 27 countries?
Chris Rogers: Not necessarily, because you have to get all EU 27 states pointing the same way and agreeing the same thing. It is worth bearing in mind that individual EU states have shown a willingness to pursue local deals or arrangements when it comes to industrial policy. If we look at semiconductors and the promotion of expanding the semiconductor industry, we have seen specific countries within the EU, particularly Germany, be willing to make money and subsidies available not at the EU level.
Clearly, there is only a certain amount of bandwidth to have all these discussions at the same time. Absolutely, the EU is a bigger market. Therefore, for the US Government it is a potentially bigger partner, but that works for and against. It is also a bigger competitive threat for American businesses, letting in a region of that size and all of those potential suppliers.
Q8 Alan Brown: You have already touched on subsidies under IRA that are available to producers where the state already has a free trade agreement. You are saying that it is going to take at least a year or so to get a critical minerals agreement. How far ahead are all these other countries going to be? What does that mean for any UK emerging industries? You already touched on lithium. Does it then become almost impossible to catch up with established trade in these other countries?
Chris Rogers: To clarify on the IRA funding, the big part that is generally being discussed is if you, as an American, buy an electric vehicle, you can get a subsidy on your purchase of up to $7,500. You only get that full amount if the vehicle you are buying has got all of its components, including its minerals, from—I forget the precise word—ratified, agreed sources. Some of those resources could be overseas. That is the big part that is being talked about.
In terms of the timeframes and whether one falls impossibly behind, it is worth noting that a lot of other countries already have mineral exports well in excess of what we are doing at the moment, but we are still seeing private business in the UK look to set up. It is not looking to set up production purely to access IRA. It is an important element of it, but not the only element of it. However, clearly the sooner a deal can be done, on the right terms, the better.
As well, we are seeing that a lot of the automakers themselves, so the multinationals, General Motors, Ford, Stellantis, Tesla and so on, are looking to sign mineral deals. This is quite different to existing automotive supply chains, where those companies will be buying completed components. That might be engines, gearboxes or whatever. The world of electric vehicle supply chains is very different to ICE. We might not be on the first wave, whether that is Japan, the EU or the UK, but that does not mean that the industry is locked up forever. As I say, speed matters but getting the right terms matters as well.
Q9 Alan Brown: In the meantime, the UK Government have to do something in reaction to IRA if we are looking to boost the internal supply chain and create these opportunities.
Chris Rogers: What we are talking about today, the Atlantic declaration and the critical minerals agreement, is a reaction to IRA, but clearly IRA is not the only game in town when it comes to supporting the electric vehicle industry either as a customer market or as a broader industrial policy. Are there other countries that are looking to support their critical minerals industries with supports outside of, “We just want to access IRA”? Yes, absolutely.
As I mentioned earlier, that is also the case for the semiconductor industry. A lot of what we have talked about so far is in relation to electric vehicles, but the semiconductor industry more broadly is seeing a lot of active industrial policy at the moment.
Q10 Anthony Mangnall: Can I come back to you, Mr Rogers, on the US‑Japan agreement? As I understand it, the agreement that they have signed does not bind the parties to any specific course of action. It is not a free trade agreement and only enters into force under signature. How much can you rely on this? In this place we talk a great deal about friend-shoring, reshoring and all of this side of things. It is very easy for us in this place to think that it is simple to say it and it is very difficult to actually do it. In this instance, how much weight do you put on that agreement actually changing and finding a way in which you can have strong supply chains to one another on critical minerals?
Chris Rogers: You are correct to say that it is not a free trade deal, neither the critical minerals agreement that we have nor the one that Japan is negotiating. However, the terms of IRA do not necessarily require a hardcore, fully ratified free trade agreement.
In fact, one challenge is that clearly it is a law. It is not a discretionary document, but we have seen that the US Treasury itself, the IRS, the revenue service there, have had to grapple themselves with what counts and what does not count. The stock of vehicles that counts towards the subsidies has increased over time. There is a degree of flexibility in there. Of course, a classic free trade agreement would be a lot less ambiguous.
Q11 Anthony Mangnall: Where I am interested is that I look at IRA as one of the most protectionist programmes that any Government have ever launched in the history of protectionism in the first place. We are trying to find ways in which we can shore up stronger supply chains, and I worry about whether the Atlantic deal benefits the US far more than it will ever benefit us, partly because it is putting IRA alongside this. Can I ask you about that point? Also, if we are going to talk about reshoring and friend‑shoring, how long is that actually going to take in relation to the Atlantic declaration? Will it be six months, two years or 20 years?
Chris Rogers: Please pull me back to the first part of your question, because you have hit one of my favourite topics. I started calling it X‑shoring because there are lots of different flavours of it. The important point here is that companies make decisions about the structure of their chains all the time.
Remember that this is not the US Government buying critical minerals. This is setting the groundwork by which Ford, General Motors or Stellantis will buy critical minerals. They put a lot of different things into the pot when it comes to those calculations. The US-China trade differences under the Trump Administration led to a lot of tariffs. That was a highly protectionist measure as well. There are a lot of inputs into it and accessing IRA is only a small part of that. Pull me back to the first part of your question again.
Q12 Anthony Mangnall: I am wondering, if we end up signing agreements with the United States and it is implementing one of the largest spend programmes on domestic production, dressing it up as a green policy rather than what it is, which is a highly protectionist model, how we end up competing. Last week, we had the outgoing head of the TRA in here. I really struggle to see how we will be able to compete and how we will be able to make this competitive for businesses that might be wanting to advance in some of the areas we are talking about today.
Chris Rogers: The critical minerals agreement does not change IRA. IRA is there. We are a taker of the market in that regard. Also, having this is a necessary but not sufficient condition for success. We need to have a critical minerals agreement if everyone else does, even if it is not perfect. There was the question earlier about whether we would be worse off with it. We absolutely would not be.
Clearly, there is a lot of industrial policy being put in place by a lot of different countries that is or could be perceived as being protectionist. There are likely to be lots of references to the World Trade Organization once the dust settles, to determine how protectionist a given piece of domestic legislation actually is. I say “domestic” in the sense of referring principally to the domestic economy. The US is not the only country doing this.
Q13 Anthony Mangnall: It will not fill many people with confidence that the US is not party to the dispute mechanisms within the World Trade Organization. Unless they are, we have a real problem here. If there is reference in these areas, we will be at a huge disadvantage.
I am wandering, so before I go completely off-piste—I would like to have a longer conversation with you on this point—can I ask how you therefore enforce and strengthen the friend-shoring and the reshoring within the Atlantic declaration? I come back to this point. How do we make it work and make it work quickly? Using the US-Japan agreement, we have made the point that it is not a free trade agreement. Do we need to be a little bit bolder about this? Do we need to make sure that it is more than a memorandum of understanding, essentially, that we all have to put these free trade agreements in force quite quickly?
Chris Rogers: Speed is of the essence at the right terms. It would be remiss of me not to repeat that these are decisions that will be made by private business and by engaging with private business and working out what private business needs from it. Fortunately, a lot of the US car manufacturers are also UK companies and have businesses here. That level of engagement with private business needs to happen.
It may not be useful to have a very delineated, “We will export X tonnes of X product”, because the market moves too quickly for that. Whatever agreement that it is going to be, if that is as simple as a one-liner that says, “UK products will be counted as domestic under the terms of IRA”, that may be good enough. I hope that answers your question. We can carry on this conversation.
Anthony Mangnall: That is very helpful, and we certainly will. Thank you very much.
Q14 Jonathan Gullis: This is in relation to electric vehicles. The Government aim to develop a globally competitive electric vehicle supply chain in the United Kingdom. How will the critical minerals agreement with the US help us achieve this aim? What specific provisions are you looking to agree with these negotiations?
Chris Rogers: I am going to refer to the critical minerals agreement as CMA, because I always trip over the word “minerals”. The CMA provides an extra export market for UK manufacturers of these minerals. It is effectively widening their addressable market. If they, hypothetically, did not have any access to the US, let us say we could not export these products to the rest of the world, suddenly your only target market is battery-makers in the UK. That could be a considerable market with the right sort of developments, but it is relatively small. Having that ability to develop for both domestic use and international use is quite important.
We have seen the success of some of the fastest-growing manufacturing economies in the world. There has been that combination of domestic market and export market. If the CMA can get to a stage where it allows UK manufacturers of these minerals to have a bigger market and access to more funding as a result, it helps, but it does not do anything per se for manufacturing of electric vehicles in the UK. It is all further upstream from that.
Q15 Jonathan Gullis: We have seen news reports in recent weeks and months about the UK electric vehicle market, particularly around the batteries, being in grave danger of looking to leave this country or not work. Does this agreement help to prevent that, or is it aiding that move to the US for a lot of these companies that are looking in that space? If I am not mistaken, the Inflation Reduction Act is essentially incentivising said industries to go over stateside rather than remain in the United Kingdom.
Chris Rogers: IRA is not dependent on the CMA. IRA was happening anyway. Companies deciding where to build their factories to access a given market was happening anyway. There is a question of whether the critical minerals agreement makes it more efficient to take your minerals to the US and make batteries there, rather than make batteries in the UK, and therefore we are not going to make any batteries in the UK because there is a more lucrative market for these minerals in the US. That could be a challenge. Without seeing the actual terms of what the agreement looks like, that is a challenge. It is worth bearing in mind as well that IRA could potentially cover UK vehicles and batteries as well. It is part of a wider pattern in that as well.
Q16 Jonathan Gullis: There is still therefore a risk that this will lead to British industry potentially moving to America because there is more incentivisation for them to do so.
Chris Rogers: I would not have said that the CMA itself encourages. If anything, it provides an extra incentive or capability to sell into the US for these minerals specifically.
Q17 Jonathan Gullis: Have you spoken to companies in the electric vehicle supply chain about what they want from the CMA with the US?
Chris Rogers: I have spoken to them, but our client conversations are privileged, so I would not want to talk about those here.
Q18 Jonathan Gullis: I understand. Is there any chance that you could put anything in writing that would be private for the Committee to be able to see but would not be for publication?
Chris Rogers: I can put some general notes in, but that would be without attribution to specific companies.
Q19 Jonathan Gullis: I would be appreciative. Following the Inflation Reduction Act, tax credits for electric vehicles in the US will increase to $7,500 until 2032. Will electric vehicle exports from the UK to the US qualify for the full extent of these tax credits?
Chris Rogers: That is quite complex and the only people who can actually confirm that are from the US Treasury, if you would like to invite them along.
Q20 Jonathan Gullis: Of course, we do these agreements at a federal level with the US Government, as it were, President Biden’s Administration. It is fair to say that it is the state level, with governors, that still has a large amount of control in terms of agreements that are put in place. What work has been done that you see in America with state-level representatives, governors, in order to make sure that the CMA is being applied? For example, is it equally across all 50 states, or is there nuance within that that allows each state to make its own decision? Therefore, should the UK be focusing more at the state level than the federal level?
Chris Rogers: IRA is a piece of federal legislation with federal taxation. State governors tend to focus more on things such as tax credits for businesses setting up factories locally. The UK Government have engaged with specific states on other matters, but, in terms of IRA and CMA, that is not an issue.
Chair: We will need to move on to Dr Leoni, where we are going to be talking about some of the Defense Production Act and domestic source issues.
Q21 Anthony Mangnall: I will move on, but I want to make the point that it is interesting to see how fast the Chinese electric vehicle market is moving. The fact that BYD has now overtaken Tesla in terms of electric vehicle and battery production should be a big cause for concern to see how IRA, and indeed any of this, will help us compete in that market. It maybe gives us some thought to actually delay our ban on petrol and diesel cars from 2030, as the rest of Europe has done.
Dr Leoni, thank you very much for being here. Can we move on to Defense Production Act? The Atlantic declaration has stated that adding the UK as a domestic source under the Defense Production Act would strengthen AUKUS implementation. How exactly is it going to do this?
Dr Leoni: I would also like at some point to address the implications of that. The domestic source is going to strengthen and accelerate AUKUS because, essentially, it is going to build trust—that is what we are talking about—between the US and the UK at a time of strategic competition. It does that because, as the UK becomes a domestic source, the Defense Production Act is going to demand that goods and technology coming from the UK, sold by the UK to the US, will have research and development, engineering, manufacturing and production activities that are located in the UK.
This is about securing that supply chain and making sure that there is no interference from untrusted entities, which, for example, could be China. In that sense, it strengthens the AUKUS deal.
Q22 Anthony Mangnall: You are confident and happy that the actual declaration itself is strong enough to be able to encourage that collaboration. In addition to AUKUS implementation, the declaration notes that the domestic source categorisation would be useful for industrial collaboration and investment in strategic sectors. What are we hoping that the US and UK might achieve from this? What are we expecting to actually see delivered from it? Is it robust enough? As we have just been discussing in terms of, say, US-Japan, it is quite woolly. It is an MoU rather than a free trade agreement. How much do we think that this is actually going to deliver?
Dr Leoni: This is obviously something that has been agreed between the UK, the US and then, eventually, Australia. Some of the speculation at this stage of our research, and also from reading the Atlantic declaration, is that I am not sure whether the UK is lobbying to be and pursuing being part of the Atlantic declaration and becoming a domestic source, or whether the US is lobbying for the UK to become a domestic source. This relates to the question of benefits and risks, which I will come to now, to address your question.
There are benefits on two levels. This will certainly be a win-win for UK businesses. The Defense Production Act mentions the fact that this could potentially be small businesses or businesses that need more contracts. That sounds great. Also, I see, as a benefit of this deal, stimulus to the fusion doctrine of the UK, which we had as a framework. It is a very good framework. It is presented in the national security capability review, but I do not think that we are applying it consistently and decisively at a time of strategic competition. For example, where the Atlantic declaration says that we will have to prevent capitals from making investments that will help rivals, in that regard both the US and the UK might benefit strategically from this.
To complete my answer, there are also risks in this regard. The Defense Production Act refers to reasonable expectations. A reasonable expectation nowadays might mean very little. We do not know. We can only speculate. Because the Defense Production Act opens by saying that this is an Act that will be useful especially at a time of war, conflict and foreign threats, my question is whether the reasonable expectation will remain an expectation one day, or will become the sorts of pressures that we have seen when we were having the debate on Huawei in the UK. That is one concern.
The other concern is about the fact that, if we become a domestic source of the US and the US can ask the UK to provide certain items, at a time of crisis we might face shortages, perhaps. That is what, at the moment, some of the European countries are facing with regard to ammunition and weapons provided to Ukraine also because of an expectation from the United States. Also, this might prevent other deals.
Q23 Anthony Mangnall: You have just said “if” this happens. Of course, to even be recognised as a domestic source we have to see Congress change the law on this. How likely do you think that is under the current Administration? Do you have any view on what it would be like under a Republican Administration?
Dr Leoni: Under the current Administration, this is very likely, because the US has a national security strategy published in 2022 and other strategic governance, which very clearly state that the strategic competition that the West, and especially the US, is facing with China is not a strategic competition that any country can face alone.
Q24 Anthony Mangnall: To come back, I understand that you are saying that about the current Administration. How does that then relate in comparison to the National Defense Authorization Act in 2023? It was passed by Congress last year and was meant to, or had a provision to, include the UK and Australia as domestic sources under the Defense Production Act, but it was removed from the final text of the Bill during a bipartisan negotiation.
Do we have to be really worried about this? We are talking about something that is quite integral to the Atlantic declaration and our agreements around AUKUS, and yet we are not able to even, in previous legislation, hold our hope that it might get passed.
Dr Leoni: I understand that. The United Kingdom needs to be less worried about whether this will be passed or not. It needs to think more about the implication of whether this will be passed. This is more my concern. Nonetheless, to go back to my point on alliance, in Washington since the Trump Administration there seems to be quite a consensus on a strategic competition with China.
It is possible that we will not see major strategic changes with a Republican Administration. We might see tactical changes, so we might see different ways in which these sorts of strategies are implemented. This might have an impact on the domestic source clause. Generally speaking, there is quite a lot of alignment across both parties when it comes to the strategic competition with China.
Q25 Andy McDonald: Dr Leoni, again looking at the Defense Production Act, there is a clear and explicit connection with national security and that focus on productivity, sustainability and safety. That is very advantageous in terms of moving some critical supply chains away from China. What are the implications for the UK in establishing a more binding relationship with the US? Is this giving the US a complete whip hand and do we just have to subjugate ourselves to this? Are there any dangers in that?
Dr Leoni: I will go back to some of the points that I have raised before. We need to put this in the context of US grand strategy. Since the Obama Administration, the United States has been unhappy with how the liberal order works, which has worked really well for it and for others in the West for many decades. Since the Obama Administration and, as we all know, since the Trump Administration, there is frustration. The US is trying to reform this liberal order, but it cannot do it formally, inside a multilateral organisation.
It is doing it informally, by pursuing more binding relations with its allies. I can give some examples in which this is happening. We are very much talking about the Biden Administration, which has made the idea of upgrading or updating alliances their main objective. We are seeing the quadrilateral security dialogue, the Indo-Pacific economic framework, AUKUS and the Build Back Better World framework. This is about building a new bubble, or sphere of influence if we want to use a more scientific term, where the US might have more influence over its allies in certain strategic sectors.
Ultimately, this is about a strategy of decoupling. It might be fine for the UK. We have had this special relationship for decades, so maybe there is nothing really to talk about. We need to keep in mind that, at a time of strategic competition and changing world order, when Asian countries are rising, for example, this might preclude us from having that sort of freedom that was also underlining the promise of Brexit and global Britain.
Q26 Andy McDonald: I understand that. I have said many times in this Committee about the dangers that China can pose in so many sectors. It poses a threat to our energy security, national security and so on. Is there a countervailing argument that, by pursuing this strategy with the US, we may be walking into some dangers that are unforeseen at the moment in terms of what is going to happen internally in the United States in the future? There is going to be an election in 2025. Of course, there is the need to keep China in conversation. We need to try to progress that relationship as well, rather than looking over our shoulder every five minutes about what it might do. Is there any scope there for attention to those sorts of issues? Does that make sense?
Dr Leoni: We should be aware that this will compromise, if that is what you are asking, our relationship with China, probably not tomorrow but when we will face a time of crisis, when the US will need us to provide certain critical goods. In preparation for that, in order for us to provide them with goods and technologies that are designed, produced and researched completely in the UK, even if a time of crisis comes in many years, in the long term we need to start preparing now in order to comply with what the Defense Production Act says, which, anyway, is not binding in reality.
We have a special relationship with the US. Considering what I have seen in the past from different Governments, I expect that our political elites will want to pursue that, even if it is not binding. On this point, to be clear, it might be worth keeping an eye on other countries that have entered the domestic source clause before. That is Canada, for example. The relationship between the US Government and Canadian suppliers so far has been minimal. We are talking about $50 million in terms of contracts. At the moment, it does not provide a very telling example.
My co-author, Dr Sarah Tzinieris, is conducting this research with me. We could not find anything in terms of whether Canada has changed its internal legislation in order to make that domestic source clause more formal and binding. This is something to keep an eye on in the near future.
Q27 Anthony Mangnall: We have obviously made quite big progress around CPTPP. AUKUS looks like a concentration of liberal, western democracies in the form of Australia and the United States. How is joining CPTPP compatible with not irritating, shall we say, the members there, with them looking over at AUKUS or the Atlantic declaration and thinking that these two may be competitors and put them at a disadvantage?
Dr Leoni: Britain’s membership of the CPTPP stems from the agenda of global Britain, based on the promise we would have greater opportunities out there after leaving the EU. Unfortunately, the CPTPP at the moment does not have the two biggest economies in the world. One of these is the US economy. The US is not going to join, or at least that is the domestic debate at the moment. China would like to join.
Going back to the benefits and risk of the domestic source clause and reflecting on whether this is binding, it is not binding at the moment but there are expectations. This poses a great dilemma for the UK because, at some point, it will have to decide whether it allows China into this trade agreement or not and whether it actually also allows Taiwan into this trade agreement. This is a big dilemma. If the US and China do not join the agreement, the benefits are limited.
Q28 Anthony Mangnall: I understand that. When we go into the realms of hypothetical over China, China is so far away from joining CPTPP because it does not meet so many of the standards. It is a red herring for us to go down that route today.
I am asking now, if you have a centralisation of supply chains outlined in the Atlantic declaration, how that might affect the UK’s membership of CPTPP and how other members may consider that. They may say, “We are at a disadvantage. We are in a free trade agreement with the UK on the one hand. On the other hand, you are doing a direct deal with, in this case under AUKUS, Australia and the United States, or just the United States under the Atlantic declaration. We are not very happy with that”. How are they going to react to this agreement?
Dr Leoni: Speaking about expectations, the expectation is that this will limit the possibility for the UK to have supply chains on certain critical technologies that are linked to other countries. Fundamentally, the problem is what we call the exclusive logic of AUKUS or the other US-led frameworks that I have mentioned already.
The problem is that we are talking about integrating more with the Indo‑Pacific. We see that as the future and yet we are proposing frameworks or taking steps that do not appear to be as inclusive or considerate of the interests of countries that are in that region. There is an important political problem there.
Q29 Chair: We are now going to move on to AI and data and those parts of the Atlantic declaration. To start on AI, I should probably declare my interests, specifically as chair of the Interparliamentary Forum on Emerging Technology as well as various other things that are on my register.
Michael Birtwistle, the Atlantic declaration signing between the Prime Minister and the President included an acceptance of an invitation from the UK for the US to come here at the end of the year for the global summit on AI. What do you think the ambitions of that summit should be? Specifically between the US and the UK, what are the opportunities that are linked to the Atlantic declaration in that summit that will be happening?
Michael Birtwistle: Thank you for having me today. The UK has an opportunity, as I said earlier, to position itself as leading the way on AI and AI governance. The UK is effectively proposing an institutionally focused model for governing AI, which distinguishes it from other models, such as that of the EU, which is a very heavy rules-based approach. The US is trying to move fast on AI but has a number of challenges related to its federal state system that can make it challenging to implement governance.
That leaves the UK potentially quite well positioned to set the pace and tone of how we should be governing these technologies. It is also worth highlighting that the UK is ahead in its approach on developing a practice and ecosystem of AI assurance. This is the idea that you can test and audit models for particular outcomes, especially before you release. There is a lot of discussion about the desirability of independent third-party auditing of foundation models that the Atlantic declaration talks about as part of a risk management approach for them. The UK, again, could be well positioned to lead the way on that.
If it does want to secure that international leadership on AI, the UK needs to have its house in order and have a credible domestic approach to trustworthy AI governance at home. That is going to mean addressing the gaps that have been identified in its framework and committing parliamentary time to give regulators, developers and users of AI the right incentives and powers to deliver on the AI principles that the UK set out. If you are taking an institutionally focused approach, it means committing to make sure that AI regulation is properly resourced.
In terms of the opportunities within the AI summit specifically, there are the potential opportunities for international collaboration there, which are around a joint understanding of AI risks, which vary from the near term to the longer-term extreme risks that have dominated the news in recent times, and concrete commitments from Government and industry. It is going to be easier to secure commitments from the large multinationals involved in developing the most powerful models if most major economies set the same expectations.
Common approaches for national governance models can create meaningful oversight particularly of model developers, which is a current gap in most major economies. There is the opportunity to build consensus on AI standards as well, which is one of the key ways in which these technologies are likely to be managed and governed in future. We would also like to see this as being the beginning of an international conversation, rather than simply a one-off.
Q30 Chair: The summit has been labelled specifically about AI safety. There is a whole host of issues around economic and social implications of AI, as well as for trade and business. Do you think that the fact that it has been called the summit on AI safety means that it is only going to try to deal with the national-security-level risks that countries are worried about? Is there really going to be an agreement that the US will sign up to that might create international obligations that might affect American industry?
Michael Birtwistle: That is a really core question that we are asking ourselves around the summit. We are concerned there are a range of ways in which you can interpret this term, “AI safety”. You can interpret narrowly to relate to a set of speculative, longer-term, extreme risks that have, as I said, dominated public debate following recent jumps in capabilities.
The types of systems that have these sorts of cutting-edge emerging capabilities are often referred to as frontier AI systems, which is a very loosely defined nebulous term. There certainly seems to be appetite from industry to focus on the regulation of systems that effectively do not exist yet.
We think that it is important to take a much more broad construction of this term, “AI safety”. When you are thinking about safety in other sectors, you are talking about a real different range of risks, from physical to impacts on people’s lives. There is a set of current and near-term risks that are much better defined, understood and evidenced. In many cases, these risks are already occurring. Those are things such as misuse, discrimination and hallucination, particularly emerging out of these models. When we are thinking about this idea of safety, it is important that the task force and the summit find the space to talk about both near-term and longer-term risks.
Q31 Ian Lavery: China is obviously a huge player in all of this, but it appears that it will not be in attendance at the global AI summit. I am not sure if you have any idea why that might be the case. If you have, maybe you could share that. Given the significance of China’s tech sector, how important is it that China is involved and included in the discussions on the international regulatory efforts?
Michael Birtwistle: It is a good question. We should be focusing on the challenges that really require international co-operation in order to address them. When we are thinking about different forms of AI risk, there are going to be some forms, particularly the more extreme, longer-term risks where the management is going to be much less effective if not all actors capable of developing the sorts of systems with those capabilities have the same view of what those risks are and play by the same rules.
Ultimately, AI is software. While you need the hardware to run it and the skills to be able to set it up, it is ultimately software. That means that the proliferation risk is greater than perhaps other risks that we manage through international collaboration.
We do not have a settled view on whether the summit is the right or only forum for China, to seek that consensus internationally and include China. We have noted public debate that involving a broader range of states gives rise to some concerns about the openness of conversation that might happen at the summit. I do not think that we have a very clear view of what those are. The interests of states are more likely to align as you consider more extreme risks. It is going to be easier to find consensus on how to manage those than some of the other risks that might be about, for example, impacts on people’s lives and harms to fundamental rights that we have more consensus on in the west. If there is a wider range of states invited, we think that it is important that that does not limit the extent of discussion and energy that is directed at nearer-term risks, around which broad global consensus might be a bit trickier.
Q32 Ian Lavery: I wonder whether any of the other panel members want to make any comments at all on China’s lack of participation, non-appearance or not being there, and how important it is that it is involved in any regulatory process with regards to AI.
Professor Jones: AI is clearly an incredibly important frontier technology. It is one in which we are seeing increasing geopolitical tension and contestation. We are going to have to find a way to work together. Having some form of dialogue with China is incredibly important. It is hard from a political point of view.
The one question is how we have that dialogue in a way that does not raise the hackles, particularly in the US, of a lot of the US hawks. One way to do that is to focus on the really technical, low-level aspects. It might not be China coming and shaking hands on the red carpet at a big summit, but it might be quiet, behind-the-scenes dialogues at a technical level.
That is what we have seen in other policy domains. We have seen continued dialogue that flies under the political radar that can be useful and helpful in aligning standards. That is probably a sensible way to go, so it does not get derailed by the big politics and it is just the quiet diplomacy. Keeping those channels open with China is incredibly important.
Dr Leoni: It is probably unlikely that we will see China engaging further in in that regard. There is an understanding nowadays in China, and also in the US, that global interdependence is not just interdependence. It is not just a win-win. Interdependence has been weaponised so it creates asymmetric relationships. Where the Chinese are unable to impose their own standards, they will prefer to build their own international regimes. These days, they are coming out with different frameworks. I do not know so much about AI, but I am thinking about geopolitical frameworks, so I think that that is unlikely.
Q33 Ian Lavery: Several tech companies have welcomed the announcement of the global summit on AI. What role do they have to play in the development of a credible regulatory framework?
Michael Birtwistle: They certainly have a role in helping develop that framework. In particular, there are going to be ways that you want to evaluate and test foundation models to understand the capabilities and whether they meet the requirements that we jointly agree they need to stick to in order to be safe and acceptable in society. There is certainly also a role for them to ensure there is a level of collaboration and standardisation across the piece. There is a distinction between what happens within an organisation developing AI and the accountability framework that holds those organisations to account and ensures they are meeting those standards and they feel incentivised to do so.
In terms of their role in the conversation, when the AI summit was announced there was a particular focus on industry voices in that announcement. We would have liked to see more reference to research community and civil society organisations being part of that voice. We have had more engagement from Government since. It is very important that the voices of those researching these models from the outside and the organisations that represent those impacted by the technologies are part of that conversation, and it is not simply about the companies that deliver them.
As I said earlier, there also does seem to be this clear appetite from industry to focus on very long-term aspects of safety. One of the roles that civil society organisations and researchers can play is in making sure that that conversation about near-term risks is also happening and is also acknowledged.
Q34 Ian Lavery: What are the risks involved that the Government might need to mitigate?
Michael Birtwistle: There are some nearer-term risks. Foundation models have shown that new risks are presented such as hallucination, which is this idea of confidently stating false information. There was a story about a US lawyer who used ChatGPT and got himself into hot water in the US by relying on that information.
We are seeing exacerbated harms such as discrimination. You can see adverts on Twitter today about people using ChatGPT-based models to sift CVs. Many of these models have been shown to exhibit gender biases, for example. Fundamentally, because of the generality of the outputs they can produce, one of the major areas of concern around these models is misuse. For example, that could be scaling misinformation in substantial ways and enabling fraud and online abuse. Those are some of the near-term risks that specifically arise from these models. They also inherit many of the other risks associated with AI technologies that I will not continue with.
As I say, there is also a set of longer-term concerns about the potential for future models to develop capabilities that involve things like deception or power-seeking. That could potentially mean that we get AI models that do not align with human interests or, for example, assist in the production of bioweapons. Those are quite speculative. In many instances, the capabilities we are talking about do not presently exist.
Q35 Ian Lavery: It is scary, is it not? That is not one of the official questions.
Michael Birtwistle: The public debate often does focus on those more extreme risks because they do seem scary. We would like to see more balance in that debate to make sure the stuff that is happening now is also addressed.
Q36 Ian Lavery: I have one final question. The approach to AI regulation is quite different in the EU and in the United States. How important is it for the UK to try to seek some sort of middle ground between their two approaches? Is that important? Is it realistic?
Michael Birtwistle: If we are talking about middle ground, the promise of the UK’s approach is this idea of context-specific governance. This is the idea that you are judging the impact of a technology as close to the point of use as possible. That means you can get more proportionate regulation. Our medicines regulator is going to be much better placed to decide whether an AI technology needs to meet particular standards or not than rules set at a higher level.
As I said earlier—I will not repeat my comments—there are a number of conditions that the UK needs to meet in order to achieve that vision. If it does so successfully, that could position it very well internationally as an attractive place to develop AI and be regulated proportionally. For the other industries that might develop around AI, as I said earlier, there is also a potential for an AI assurance ecosystem to develop in the UK.
Q37 Andy McDonald: Professor Jones, we are aware that the transfer of data is fraught with difficulties and restrictions. Specifically in the context of the US, what are the challenges UK businesses face when transferring data to the US?
Professor Jones: Let me be specific here. It is personal data that is affected. It is not all forms of data. Let me just preface my comments by making a couple of observations about how important data is for the UK and transatlantic trade flows as well as global trade flows.
Let us take our services sector as an example. The latest estimate from DCMS is that 93% of our services exports are data-enabled. Data certainly underpins all of our services exports. I was struck by the fact that 80% of our services exports are remotely delivered. As you can imagine, if the internet goes down, a lot of our exports are unable to happen.
This is also true for the goods trade, though. We were talking earlier about transport and logistics globally. Freight is moved across borders. Again, that is heavily dependent on trans-border data flows. That is all to say that data flows really underpin all of our cross-border economic relationships in a really quite profound way.
The challenges, particularly with the US, are with regards to personal data. That is referring to data through which you can either directly or indirectly infer the identity of a person: your name, your address, your physical attributes, et cetera. These are the sensitive parts of data. They are incredibly important for a lot of our services exports. If you think about financial services, our personal data often needs to be transferred. If you are a company like Mastercard, for example, you might need to transfer information between your affiliate and the UK and the US. Personal data is important. Personal data flows are important.
As you will be aware, we have the UK GDPR, which is modelled on the EU GDPR. We take personal data protection very seriously, at the moment as seriously as the EU does. Historically, there has been a big challenge for EU companies—of course, when we were in the EU this applied to us as well—to send data to the United States because the United States does not have as rigorous an approach to personal data protection.
We have had a series of agreements between the EU and the United States over time that sought to facilitate cross-border flows of personal data. What has happened is that these agreements have been challenged successfully by Max Schrems—you might have heard of the Schrems I and Schrems II judgments—and struck down. We have a situation at the moment where UK firms looking to transfer personal data to the United States have to go through a lot of administrative hurdles to do so.
If they are a big corporate, they can use what are called binding corporate rules. Again, in that Mastercard example, they are able to transfer that data because we have said, “Yes, Mastercard in the US has put measures in place to protect our personal data when they transfer it abroad”. Other firms looking to transfer data have to use what is called standard contractual clauses. Again, that is a mechanism by which, on a case-by-case basis, they are saying, “When we transfer your data overseas, we are going to make sure the standards are equivalent to UK data protection”. It is a very cumbersome process.
What has happened—this is where the Atlantic declaration is quite helpful—is we are now moving towards a new arrangement between the UK and the US, which will hopefully make those data transfers much smoother.
We are going to dock on to an agreement that has just been reached between the United States and the European Union. Yesterday, the European Commission decided that this new arrangement is adequate. The US has put a series of measures in place vis-à-vis personal data from the EU going to the United States. Yesterday, the European Commission said, “Yes, we deem this adequate”. Henceforth, European firms looking to transfer that personal data will be able to do so. We are now looking to put an analogous system in place, and I am happy to go into the details of what that will look like.
Q38 Andy McDonald: Could I just ask a supplementary? We have been speaking about the issue of medical healthcare and the businesses that circle around that. There has been a longstanding discussion about the safety of our medical data. How do you see these discussions impacting upon that sphere and what reassurances can people gain from any of those discussions?
Professor Jones: I was surprised to note quite how much of our medical exports were happening in ways I had not really understood. For instance, I learned recently that, if you have an x-ray or an MRI scan done in the UK, the chances are that it will be interpreted abroad. There is a lot more personal data that is—
Andy McDonald: It is moving anyway.
Professor Jones: Yes, an Australian radiologist might look at your scan and say, “This is our diagnosis of it”. As we globalise, digitalise and increasingly rely on these cross-border transfers, it is very important that they are robust and that, as citizens, we can be sure our data is being protected. We need to seek to ensure that any of our personal data moving overseas, including our health data, is treated as it would be if it were in the UK, which is why we have these very cumbersome mechanisms.
I have not delved particularly into the healthcare sector, but the premise will be that any healthcare data that is being transferred abroad, for example to the United States in this instance, would be treated with the same level of protection as it would be in the UK.
Q39 Andy McDonald: You would not fear the exploitation of data for commercial purposes. You do not think that is a live risk.
Professor Jones: It is not a risk that arises particularly from the data bridge. It is a wider question. It is a really important question about how we treat health data under the GDPR. We have the Data Bill going through at the moment. We need to make sure we have really robust protections in our Data Bill for that health data. My personal data might be sitting with the NHS. As soon as it is transferred to a private provider in the UK or elsewhere, it needs to be upheld to the same level.
The additional issue with cross-border transfers of our health data would be whether we can trust that the other party is then going to uphold that standard. This has been the issue in the United States. For example, one of the reasons the Schrems II decision struck down the original EU-US arrangement was because there was no mechanism for redress. The implementation and the governance were seriously problematic, which they have now rectified.
Q40 Chair: As I understand it, as of yesterday, because the US and the EU have done the deal, the UK is now behind in comparison to the EU in its exchange of data with the US and the data bridge is an intention to fix that. What are the timescales and the process for that to happen? Do we just have to copy what was agreed yesterday and how do we do that?
Professor Jones: It is not exactly a bilateral agreement. What happens is the United States—it is similar to what happened under Privacy Shield previously—has basically said to the European Union, “We understand your concerns. We have put the following measures in place”.
Some of those concerns were around the accessing of EU citizens’ data by the national intelligence services in the United States. They have now limited that. The Biden Administration passed an executive order last October to say, “We are going to limit, certainly for EU citizens, the use of that data, and we are going to put this new governance mechanism that enables redress”.
They have negotiated a set of principles that US companies now have to abide by. If a US company wants to operate under that EU-US framework, they self-certify. They say, “We will hereby adhere to these set of principles”. In a way, it is a package that the United States has put together to say, “This is how our national security agencies will treat data. This is how our companies, self-certified under this regime, will treat data”. They have deemed that adequate.
The question for us now—in a way it is helpful that we are thinking about this after their decision—is, “Do we deem that those measures are fit and appropriate for the UK?” The step now is for the Information Commissioner’s Office, DCMS and the Minister finally to make a decision as to whether they think the new US arrangements would be deemed adequate from a UK perspective. Given how similar our arrangements are, I do not imagine that we would make a different decision to the European Commission.
I imagine that the timeline is this year. It is dubbed the UK-US data bridge. Essentially, it is saying, “Do we deem the US arrangements to be sufficient from our point of view?” On the US side, they would then need to add us as a recognised party under their executive order so the limitations on the intelligence services and the access to redress also apply to UK citizens. Does that make sense?
Q41 Chair: It does. On the one hand, given that the Europeans have said the American scheme is okay, I suppose it is easier for us to say it is okay without risking irking the Europeans.
On the national security point, though, the UK and the US have a particularly established relationship on sharing data around national security provisions. That is different to the EU and the US, as I understand it. Are the Europeans going to be worried, therefore, about the UK-US data bridge and whether the executive order exemptions on national security will really be effective, given our wider relationship?
Professor Jones: That is a really good question. That is about two things. I am not an expert on the exact nature of that relationship, but, insofar as we have a deeper relationship and our intelligence agencies share more personal data of UK citizens, for example, the question is, “Can we reassure the EU that we will not be a transit mechanism?” That is in terms of what are called onward transfers. That will be a mechanism that we will need to reassure the EU so we do not jeopardise our adequacy arrangement with the EU. That is a tightrope that we will need to walk. I imagine that our negotiators are very apprised of that.
What is interesting, though, is the extent to which we have the same arrangements. We have UK GDPR, which is the equivalent to the EU one. Will that be sufficient to keep the EU happy from an adequacy point of view?
Q42 Chair: That is probably, therefore, a question for them.
Professor Jones: It is probably a question for them, exactly. In terms of the timeline, by the end of this year we will see that UK-US data bridge operational, I would imagine.
Q43 Chair: Yes, I cannot see that there are any particular problems to that deal being done, which will be good news.
Professor Jones: Yes. Max Schrems has already put out a press release yesterday saying that they are going to lodge a case against it. We will see another case in the Court of Justice of the European Union by the end of this year and probably early next year. It will take a few years to work through, but the question will be about whether it is sufficiently rigorous to reassure the EU courts that it really is upholding the privacy of EU citizens sufficiently.
Q44 Chair: We have come to the end. I just wanted to ask one final question. I want you to rate the Atlantic declaration on a scale of zero to 10. Zero means it is all words and nothing of interest; 10 means it is a substantive and significant change in the trading relationship between the US and the UK. What do you think, Professor Jones?
Professor Jones: You have really put me on the spot here. I am going to say two. My caveat is that it gives a window on something that is substantive and important. When the UK-US data bridge comes into force, I would put it up there at a six or seven. It is a really significant thing. The Atlantic declaration does not move it along much.
Michael Birtwistle: Given the absence of very much global co-ordination on AI to date, it is probably closer to 10 than it is to zero.
Chair: It is closer to 10. That is not really giving me a number, but I will let you get away with it.
Anthony Mangnall: That is a nine, is it not, Chair?
Chris Rogers: We are in the data business so I should give a firm number, possibly with decimal points. I would say it is better than nothing, but it is the start of something. I would probably assign it something in the order of a three, if we are going to be precise.
Dr Leoni: I would probably sit in the middle, to the extent that on its own it probably does not mean much, but in the context of previous developments internationally and between the US, the UK and Australia, it shows that there is steady progress in integrating the relationship between these three countries more and more. There is substance, if we look at the previous steps that are being taken.
Chair: We are averaging around three or four. As you say, that is better than zero.
Thank you so much for helping us understand the Atlantic declaration. I am sure the US-UK relationship is something the Committee will come back to later in the year. I know there will be lots of topical interests, especially on critical minerals, electric vehicles, which we are also looking at, AI and the AI summit, and the data bridge and digital trade. No doubt we will speak again in the future. For the purposes of today, I will call the session to an end.