Friday 9 June 2023

Could post-Brexit Britain lead global AI regulation?

It’s been a relatively quiet Brexit news week, with just the usual drip-drip of news about its negative impact on everything from the availability of au pairs to the production of strawberries to Brittany Ferries’ cross-channel freight volumes. As I remarked in a recent post, each individual piece of damage may be fairly small and easily dismissed, but the cumulative impact is ever-more alarming.

So in this post I’m going to focus primarily on Rishi Sunak’s ambition for the UK to lead international regulation in relation to Artificial Intelligence (AI), an ambition for which he sought to gain US support during his visit to meet President Biden this week. It was also announced that the government proposes to host a 'global summit' on AI in London this autumn.

Were this ambition to be met it would not, of course, be a benefit of Brexit – the UK could just as easily have proposed itself for this role as an EU member, and for reasons discussed below would have been in a better position to do so – but it is clearly part of an attempt to define Britain’s global role after Brexit.

The UK as an AI regulator

AI and its regulation are very rapidly emerging as huge economic and political issues and, as initially reported in the Financial Times (£), Sunak’s proposals might involve the UK hosting both a CERN-style AI research centre and an international regulatory authority based on that for nuclear energy. These are not terrible ideas. The UK has a relatively strong research base in AI and AI ethics – and the latter ought to serve as a reminder that despite the typical right-wing refrain that only STEM subjects are useful, and everything else is ‘woke’ or ‘mickey mouse’, the reality is very different. Philosophy and many other humanities and social science disciplines are quite as important in many business and regulatory fields. That’s long been almost a truism in Silicon Valley, the government’s favourite template for a post-Brexit innovative economy.  

The idea of the UK as an AI regulatory centre is also more realistic than previous ideas about how, after Brexit, the UK might create regulatory regimes in all kinds of areas in the expectation of attracting businesses to the UK because of its regulatory regime and, in due course, that regulatory regime emerging as the global regulatory standard. That approach was set out in the government’s January 2022 paper ‘The Benefits of Brexit’ (see, especially, pp. 24-29), as indicated by sub-headings such as ‘a sovereign approach’, ‘leading from the front’ and ‘setting high standards at home and globally’.

That is very much a relic of the now barely mentioned ‘Global Britain’ strategy, yet, as regards AI specifically, as recently as March of this year, in its policy paper introducing a consultation exercise that has still not been concluded, the government was talking hubristically of how “having exited the European Union we are free to establish a regulatory approach that enables us to establish the UK as an AI superpower.” Sunak’s latest ideas at least recognize that international regulation entails international agreement, rather than regulatory competition between unequal players.

In fact, they seem to owe something to, or at least are consistent with, the proposals in the June 2021 report by the Tony Blair Institute on the future of UK regulatory policy. These included that “the UK could strive to play a role as a regulatory convenor between the US and the EU … but only with the right focus and regulatory diplomacy strategy.”

The ironies of Brexit

Nevertheless, there is something deeply ironic about the UK, having not just left the EU but done so in a way that prioritizes national sovereignty in regulation, now proposing itself as being in the vanguard of international cooperation and regulation. Moreover, it serves to underscore the importance of hosting international regulatory organizations, and thus how much has been lost by Brexit leading to the departure from the UK of the European Medicines Agency and the European Banking Authority. Hosting such bodies can have a value way beyond themselves, as it tends to establish a whole eco-system of basic research and commercial application. It is very much to his shame that David Davis, when Brexit Secretary, fantasized that (£), somehow, the UK could continue to host those EU regulatory agencies post-Brexit. Never mind artificial intelligence, as so often Davis lacked the genuine sort.

That irony goes even deeper, given that for many Tory Brexiters the guiding thread of Brexit was not just national independence in regulation, but hostility to regulation in principle. As regards AI, that is also reflected in the proposals in the government’s consultation exercise, which are firmly tilted away from having a dedicated AI regulator even for the UK and towards a “light touch” regulatory model. In fairness, it could be said that developments in AI, and understanding of their implications, has moved very fast even in the three months since the exercise was launched, and everyone is scrambling to keep up. However, the Tories’ pre-disposition against regulation, both generally and in its current AI proposals, does damage the government’s credibility in now pushing to lead global AI regulation.

Moreover, Brexit has already exacted a price. As a Politico report this week explains, Sunak is making a pitch directly to the US because the UK doesn’t participate in EU-US forums like the Technology and Trade Council, where AI regulation is already being discussed, and the EU and the US, along with Canada, have already developed proposals for AI regulation, to be presented to the G7 this autumn. The UK’s supposed post-Brexit ‘nimbleness’ isn’t quite as nimble as all that, and, as in so many other areas, it suffers from not being ‘in the room’.

Mr. Sunak goes to Washington

Perhaps, as the Politico report suggests, the UK can have a role as leading the smaller players, outside the EU and the US. There’s quite a lot of evidence that Japan, in particular, and despite its initial anger about Brexit, which was seen as a breach of faith as regards Japanese investors in the UK, is keen to support and enable post-Brexit UK’s geo-political heft generally (not in relation to AI regulation, particularly), not least because of relations with China.

However, Sunak is clearly pitching for more than being just ‘the leader of the others’ as regards AI and its regulation. He will undoubtedly have been pleased with Biden’s words in the post-meeting press conference, saying “we are looking to Great Britain to help lead a way through this  … there is no country we have greater faith in to help negotiate our way through this”, but what this means in practice remains to be seen (‘to help’ is not quite a ringing endorsement).

Of course, no one expected the issue to be settled in this single meeting, and the real question is whether the UK is well-placed to deliver. Although the US-UK relationship still matters to both countries, at the most general level, as Rafael Behr put it in his discussion of Sunak’s Washington visit, “it is a bald strategic fact that Brexit makes a British Prime Minister less useful to Washington. Without leverage in Brussels, Sunak is not in a position to broker deals with Biden.” Unusually, this prompted David Frost to engage with critics of Brexit, ‘taking to Twitter’, as they say, to denounce Behr’s “declinist worldview” (‘declinism’ being the standard Brexiter way of recasting the observable damage caused by Brexit as the negativity of those who observe it).

Inevitably Frost’s complaint was, implicitly, based on his persistently simplistic understanding of sovereignty. First, he asserted that “the purpose of British policy is not to be ‘useful to Washington’ … it’s to pursue our own interests.” It apparently doesn’t occur to him that the UK might judge it to be in its own interests to be useful to Washington and, for better or worse, has often tended to do so, the Iraq War being an obvious example.

Frost then continued that he doesn’t “agree that those interests are best pursued [via EU membership, because] it is better to define and then pursue them ourselves, through alliances and partnerships of various kinds.” It apparently doesn’t occur to him that, as an EU member, the UK always defined and pursued its own interests, including the development of the single market and eastwards expansion, and that one means of pursuing them was through the ‘alliance and partnership’ of EU membership. And if Frost means that EU membership precluded the UK pursuing its perceived interests then, once again, Iraq gives the lie to that. There is hardly a more profound expression of sovereignty than the decision to go to war.

What Frost doesn’t comment on, though Behr mentions it, is the way that what used to be seen as the totemic Brexit gain of a US-UK Free Trade Agreement is not on the agenda of Sunak’s visit. That is not a news item, and was clear even before Liz Truss publicly acknowledged last September that there is no prospect of such a deal. Still, it is striking how quiet the Brexiters have gone considering that, in July 2017 Digby Jones was crowing that “remoaners” must be hating the fact that a US trade deal was “in the bag”. Even those who did not make such transparently false claims, but did hold out a deal as a key prize of Brexit, have never admitted that they simply got that wrong.

Pointing this out isn’t petulant point-scoring. It matters because it is becoming increasingly common for Brexiters like Frost, and some more neutral commentators, to talk as if various piecemeal deals on trade or security or other kinds of cooperation, including this AI role if it comes off, somehow justify leaving the EU. But, even leaving aside that some of the examples commonly given (e.g. AUKUS) didn’t require Brexit, and others (e.g. CPTPP) at best do no more than slightly mitigate the damage it causes, Brexit should be judged not on these things but against all the promises that Brexiters made for it. It’s one thing to make the best of a bad job, quite another to forget that it wasn’t sold in such dismal terms.

Can the UK be trusted?

All that aside, and quite apart from the way that Brexit inevitably reduced the UK’s influence with the EU and the US, the idea of the UK leading global regulation is undermined because the way Brexit was undertaken damaged Britain’s reputation for political stability and even for commitment to the rule of law. It is, to say the least, unfortunate that a country aspiring to a global regulatory role should not only have indicated its willingness to break international law but on more than one occasion asserted that, by definition, national sovereignty trumps such law. There’s a reputational price to be paid for Brexit Jacobinism.

Sunak may have introduced some stability and pragmatism, especially with the Windsor Framework, but his own words at the post-meeting news conference, about knowing that “some people have wondered what kind of partner Britain would be after leaving the EU”, show an awareness that the memories of the last seven years have yet to fade. Biden may have expressed “faith” in the UK but he won’t have forgotten the shenanigans over the Northern Ireland Protocol, and will still be watching to see how the Windsor Framework is implemented.

In any case, for all Sunak’s reassuring noises, he has only weak control over a party of which some sections are still in denial about why Liz Truss crashed and burned, and are openly flirting with ‘National Conservatism’. So the world is well aware that lurking behind Sunak’s gloss is a party seething with antagonism not just towards the EU and to Biden’s US, but towards the ECHR, IMF, OECD and even the WHO. Whilst such sentiments can be found in other countries, the fact of Brexit and of those voices being influential in the governing party certainly don’t help to make the UK seem a stable or attractive site for international regulatory leadership. You can hardly dismiss existing international institutions as the ‘Globalist Establishment’ whilst credibly proposing to establish a new one.

In any case, even leaving aside post-Brexit Britain’s trustworthiness, there are also questions about its operational capacity to undertake big projects like that of AI regulation. Although of a very different nature, nowhere is that more obvious than in the much-delayed introduction of import controls on goods from the EU. I discussed this most recently a couple of weeks ago, but at that time I missed something little-reported outside the specialist trade press, namely that earlier last month, apparently in response to industry lobbying of the government, the requirements for most fruit and vegetable imports have been relaxed, so that they will be treated as “low” rather than “medium” risk, with corresponding reductions in paperwork and physical inspections.

That’s good news so far as it goes, as it should reduce disruption to supplies (though it shouldn’t be forgotten that by doing so, it also increases the risks of a plant disease being imported). However, it also adds to the problems relating to ‘groupage’, which also proved to be so difficult when the EU introduced post-Brexit controls on imports from the UK. In practice, this means that mixed-loads of low-risk and medium-risk products are liable to be stopped for inspection, so even if someone is importing low-risk vegetables it won’t help them much if they are shipped in the same consignment as a medium-risk product. This will hit small businesses and specialist product lines especially hard as, again, happened when EU import controls were introduced.

The fact that the import control regime is still being worked out, years after the decision that Brexit meant ‘hard Brexit’ made import controls inevitable, shows how inept the Tory Brexit governments have been. Even now, less than four months before the next tranche of controls come in, those who have to operate them at ports don’t know what is going to be involved, to the extent of not knowing whether 1% or 30% of food imports from the EU will need to be checked.

All this is a big Brexit story in itself, as it is likely to lead to shortages of some products and higher food prices in the UK. But, amidst the general sense that ‘nothing works in the UK anymore’, it will surely also contribute to international scepticism about the idea of the UK leading AI regulation. After all, if post-Brexit Britain can’t even cope with the most basic consequences for the most basic goods of the regulatory border created by its new-found ‘independence’ from the EU, can it really be relied upon to take the lead in the most complex, advanced and fast-moving of global regulatory challenges?

Labouring a point

This brings me back to Labour’s Brexit policy, my discussion of which in last week’s post got a distinctly mixed reaction on social media. I’ll come back briefly to that, but as regards something like the UK pitching to lead AI regulation a Labour government would have an advantage over a Tory one simply because it would not be contaminated with the deregulatory madness and hostility to global institutions of the Brexit Ultras. That still doesn’t mean it would happen, but it would be a more favourable political environment within which to try it, though, by then, events will probably have overtaken us as regards AI. But the general point holds: in the eyes of many foreign capitals, a Labour government would be a return to normality.

Normality, here, means rationality and trustworthiness, but it would be unrealistic to think that, simply by electing a Labour government, the UK’s international reputation is going to be instantly restored. That will take a long time, which is partly why any prospect of (re)joining the EU is a long time away. Those insisting that this should be Labour’s policy for the next parliament need to ask themselves what value there would be in proposing something that would whip up huge controversy in the UK, to the point that there might not even be a Labour government, but if there was such a government with such a policy then the EU would be bound to turn the idea down unless or until it ceased to be controversial in the UK.

That observation isn’t negated by Michel Barnier’s widely-quoted comment this week that “the door is open” for the UK to join. Apart from the fact that it isn’t up to him (though he’s clearly a well-informed and credible voice), it would be absurd to interpret that to mean that an application for membership would be accepted unconditionally and without question. It’s not just the issue of negotiating the terms of membership, it’s whether there was a stable, sustained and unambiguous commitment in the UK to membership. Someone might say to an alcoholic spouse who has walked out of the family home that the door is always open for their return. It doesn’t mean they wouldn’t have to dry out before they were welcomed back.

Of course, I may well be proved wrong. Who, after the last seven years of politics, would be rash enough to make unequivocal predictions? But, having seen what became of the Brexiters’ pursuit of unicorns over those years, I’m not immediately persuaded of the merits of chasing what, at least for now, seem to be rejoiner hippocampi.

No comments:

Post a Comment