Big Tech in the UK

Dorothy Bishop’s attempt to get The Royal Society to oust Elon Musk has crystallised political dilemmas of sweeping consequence

The Royal Society is often described as the world’s oldest learned society. It operates from a pair of enormous wedding cake edifices on the drive from Buckingham Palace to Downing Street. I have been in and out of it all my working life and seem to recall Prince Andrew conducting a quasi-masonic ceremony involving a silver bell at the event to mark its 350^th anniversary. To be elected to its fellowship is, for any scientist, a signal honour. Last week, Dorothy Bishop resigned from it.

Bishop is an emeritus professor of developmental neuropsychology at Oxford. If your child has autism or dyslexia, it’s likely the professionals you talk to rely on her work. She resigned because she, along with more than 70 other fellows, has been unable to persuade the Society to repudiate the fellowship it granted to Elon Musk in 2018. In doing so, she has touched on three issues of sweeping consequence.

First, Musk was elected a fellow only because of his mastery of important technologies, including at the time SpaceX and Tesla. So Bishop is raising the question of how the rest of us should relate to Big Tech as those companies and their owners step ever more openly into the realm of politics.

Second, she wants to repudiate Musk because of his populist attacks on science. So she is raising the question of how science should respond to an antagonistic tide that is now global.

Third, the global phenomenon of populism has a powerful hold in the UK, running through Brexit, Nigel Farage’s Reform party, a large faction within the Conservative party and arguably a significant faction in the Labour party. So in seeking to repudiate Musk, Bishop is also asking The Royal Society to take a position, by implication, in UK politics.

That is, The Royal Society, an institution that I have known for 40 years as an intimate ally of the UK state that is above party politics, finds itself now embroiled in the kind of political street fighting you can find every hour of the day on social media. This will make many fellows distinctly uncomfortable. However, I think the ultimate cause of this is not Bishop but a combination of shifts in wider society and decisions taken by the Society itself.

Big Tech

Musk was elected to the Society in 2018, by which point the Society was already close to what we now call Big Tech. Its website tells us, “Throughout 2018, The Royal Society supported by DeepMind brought you the world's leading thinkers in AI, in order to help build a greater understanding of what machine learning and AI are, how the technology works, and the ways it may affect our lives.” This project was part of a longer programme in which the Society and Big Tech (DeepMind by this point already being owned by Google) worked together towards the shared goal of managing social attitudes towards AI in order to avoid the kind of popular revolt that prompted Tony Blair to ditch genetically modified crops. So far, so normal. But now look at the bigger picture from the point of view of Big Tech.

Big Tech is defined not simply by big companies or big products but by monopoly ecosystems. For example, Google does not merely dominate search but also largely owns the AdTech stack of services that companies use to place advertisements across all media. There has been a global failure of competition law to limit these monopolies, which have grown ever-bigger and ever more sure of their ability to extract monopolistic rents. Thus it is clear that any effective action to protect customers will have to come from regulation. We saw how this can bite last year when the UK’s Competition and Markets Authority blocked Microsoft’s acquisition of Activision Blizzard and ultimately obliged it to substantially restructure the deal for the games company. Regulators rather than other companies are what Big Tech fears.

Regulation generally allows for a minister to override the regulating agency and, thanks to a lack of good economic evidence on the long-run impact on consumers of the monopolies, there’s often a grey zone to be exploited.

Recourse to such ministerial intervention seems to have been Microsoft’s first reaction to the CMA’s block last year. Its president said the decision was "bad for business" and that confidence in the UK had been "severely shaken" – a rare and stinging attack plainly intended to damage the government.

In the US, incoming Republican administrations have previously scrapped remedies coming down the line towards Big Tech. If Google, which is on the verge of losing its Chrome browser, fails to pull off the same trick with Trump this time round, its Beltway team is surely getting sacked.

Thus there is an incentive for Big Tech to penetrate the states it operates in and create a political climate in which the envelope of potential action by regulators to break up the ecosystems is constrained. Some of this is company-specific but, back in California, the key players are connected and commonly trade favours; there is enough shared interest for it to have elements of a common project. This is far beyond “lobbying” in the sense in which it has historically been understood round Westminster and brings to the UK attitudes and techniques honed in the Washington swamp. One of the first things Google did when David Cameron became Prime Minister was to hire one of his cousins. In July last year, Politico named 22 former MPs and staffers from all parties who are on the Big Tech payroll, starting with the former Deputy Prime Minister Nick Clegg. Clare Barclay, the chief executive of Microsoft UK, has just been appointed to chair the government’s new Industrial Strategy Advisory Council. Penetrating The Royal Society is a natural part of this effort.

Right now, the acid test of Big Tech’s British swampwork is arriving for, as of 1 January, the UK’s regulatory regime will become significantly more capable when the CMA’s new Digital Markets Unit, arising from the Digital Markets, Competition and Consumers Act 2024, comes into operation. In the EU and Canada, regulators are starting to have concrete effects on business. The new UK regime, less mechanical than the EU’s, has the potential for even bigger impact. Further down the line, new legislation is coming that will establish a UK regulatory regime for AI.

In the current government, the deepest connection is between Tony Blair and Larry Ellison, Silicon Valley kingpin and master of Oracle, a company that specialises in big administrative systems. Ellison has bankrolled the Tony Blair Institute with more than $100 million and an Oracle director, Awo Ablo, is the TBI’s executive vice president. Via Blair and the Labour faction clustered around him, Ellison connects with Peter Kyle at the Department for Science, Innovation and Technology and Wes Streeting at the Department of Health. Both connections have obvious value.

Kyle is responsible for the AI legislation. Streeting is responsible for the NHS, a sector in which no Big Tech firm has yet been able to establish a monopolistic ecosystem. That slot is up for grabs and in this contest, the UK holds a distinctive position. The NHS is one of the biggest healthcare providers in the world with, thanks to a trustworthy regulatory regime and diligent development by UK researchers, unique potential as a source of data. Just look at the way the world is flocking to use UK Biobank.

Via Blair, Ellison can also claim credit in California for a wave of political support for tech, especially AI. One element of this is Blair’s book, On Leadership, where David Runciman discerns that the chapter enthusing about tech was ghost-written. Another element is the soaring, epic support for tech evident in A New National Purpose, a 2023 TBI report co-authored by Blair and William Hague. Another is a tidal wave of tech policy reports from the TBI that threatens to drown out, in classic Beltway fashion, any other position; since A New National Purpose was published, the TBI has released more than 70 reports that make the case for tech, most aimed at the departmental responsibilities of Streeting and Kyle.

I have previously described how this ‘tech turn’ at the TBI serves a political purpose for Blair and Hague. This will come up again below. But what we’re seeing here is something else, that it also serves a purpose for Big Tech. In other words, there is an alliance. The TBI is the crucible for the making of an alloy that is not purely political and not purely commercial, where power and money melt and merge into a single whole.

Regulator containment is not the end of its aims. They also include diverting large parts of the budget for public services to Big Tech. This arises naturally from the distinct conceptual position found in the TBI, which has two parts. First, said out loud, it pushes tech to the front in the search for solutions to practical problems. Second, unspoken, any kind of activist industrial policy is repudiated; the state’s role in the economy excludes, as Margaret Thatcher decreed, any ‘near-market’ activity and instead is confined to far-from-market activity – academic science. If the TBI can get you to accept those two propositions, then you will of your own volition make your public services heavily dependent on tech and you will satisfy this new need by creating thousands of contracts for tech companies.

You can see this two-pronged attitude in the way the TBI framed sovereign AI capabilities in its report Governing in the Age of AI in May. First, the report’s basic premise is that government needs to reorganise itself around AI. Second, it recommends:

“Build sovereign AI capability in collaboration with trusted private-sector partners: government will need to create its own models, fine-tune existing ones or build tools on top of off-the-shelf LLMs, deciding on the appropriate approach and the best foundation model for each use case, to ensure that it can: first, train its own custom LLM for national-security purposes (which for the purposes of this report we name CrownIntel) on open and official data, and fine-tune it on sensitive and confidential information in a secure environment; second, create a ChatGB “legal advisor” tool for government by fine-tuning an off-the-shelf GPT-4 class model on legislation and parliamentary records; third, work with individual departments’ digital teams to build or procure AI tools using commercial foundation models. These tools for ministers, officials, caseworkers and analysts would be bespoke to each department’s individual use cases and could number in the hundreds or thousands."

This is an approach that builds in a dependency on Big Tech as the gateways to the main Large Language Models. Presumably because the security services have insisted on it, under this plan the UK will in fact produce its own LLM, CrownIntel. But this effort is to be siloed in such a way that it can have zero commercial implications. No effort is made to envision how a £10 billion effort by the state in a rapidly moving technology in which the UK has strong position in the basic science is to be leveraged into British companies and British jobs.

You can also see this attitude in decisions of the new government in those areas that fall to Streeting and Kyle. In health, the section on life sciences in the Green Paper on Industrial Strategy could have been written by the last government, the one that awarded a key NHS data management contract to Palantir, a firm controlled by Musk’s Silicon Valley mentor, Peter Thiel. Outside health, Kyle has been given a cross-departmental remit to promote technology, especially AI, providing scope for more Palantir-type contracts in every department.

With the benefit of hindsight, we can see Musk’s election to The Royal Society in 2018 as part of a more general and successful campaign by Big Tech to penetrate the UK state. This goes far beyond lobbying for individual policy targets and instead, via the alliance with Blair, has acquired a truly political dimension, both conceptually and in terms of boots round the Cabinet table. In this way, Big Tech is present in Keir Starmer’s administration just as it will be in Donald Trump’s.

Bishop’s demand that the Society repudiate Musk as an individual crystallises a basic political demand concerning Musk as a representative of Big Tech – that the penetration of the body politic be reversed and Big Tech be pushed back to the outside.

Populism and science

One problem in discussing populism is that people use the word to mean different things. This is one reason for relying on Bickerton and Accetti’s influential concept of technopopulism. In this, political leaders succeed by combining technocracy and populism:

Technocracy

· a claim to a particular kind of techne (competence or expertise) that entitles its possessor to rule over others

· a direct relationship of trust between the possessor of this techne and those they are supposed to rule over.

Populism

· presenting the country as divided between a pure people and a corrupt elite

· the claim of the leader to exclusively represent the people.

This broad conception includes within it both Blair and Trump. My last essay was on this topic and afterwards I found myself discussing it in the comments with Stephen Webb, the head of government reform and home affairs at Policy Exchange, the leading think tank for the Conservative Party. He complained that Blair and Trump are not the same and it’s a fair point. Even though they share some qualities, Blair and Trump are still very different. The question is, how exactly?

Technopopulism the book distinguishes three ‘pure’ varieties of technopopulism, New Labour, M5S in Italy and Emmanuel Macron in France and uses these as templates for ‘hybrid’ cases such as Podemos in Spain (where the technopopulism is fused with an avowedly left-wing agenda). As categories, these case studies are too baroque for my taste and applying them requires too much interpretation.

More straightforwardly, I think, we can observe how technopopulists respond to constraints. In the terms of Bickerton and Accetti, any kind of constraint is inherently problematic for a technopopulist. Part of their appeal is their claim to the techne that justifies their rule over the rest of us. If they allow themselves to be constrained, they immediately acknowledge a superior competence, which undermines the appeal of the techne. It also risks making them look like just a cog in a machine. And that pre-existing machine is part of the corruption that they are supposed to be attacking on behalf of the pure people.

Two kinds of constraint seem to me telling. The first is that imposed by the constitution. Blair never bridled at the limits imposed by the UK’s constitution. Indeed, in pushing for the UK to join the euro, he pushed for far greater constraints on the handling of the economy. By contrast, the great fear of Americans for Trump’s second term is that he will shred the constitution.

The second is epistemic. It is the necessity to respect empirical evidence and cogent reasoning. Any lawyer will recognise the centrality of these commitments, but it was a science journalist who was the first to ring alarm bells; in 2005 Chris Mooney’s The Republican War on Science detailed a campaign to undermine the authority of science that was already 20 years old. Today, the Maga crowd seek to bring to heel what they sometimes call the Citadel, a label for all those institutions, such as universities and newspapers, that sincerely and autonomously pursue truth. Again, Blair does not bridle at this constraint.

Call a technopopulist who accepts both constraints light, one who rejects both constraints dark. Hence Blair is a light technopopulist while Trump is a dark one. In this way, we are able to discuss what is different between the two without obscuring what is the same and without needing to limit the possibilities to the left or the right. Bishop’s beef is with the dark side, to which Musk evidently also belongs. It is the visceral fear of violence cultivated by Musk against scientists in the US that is the immediate prompt for Bishop’s resignation. Only the dark side undermines constitutional norms by tweeting in the middle of the UK’s summer riots that civil war is inevitable. Only the dark side undermines epistemic norms by tweeting memes like this one highlighted by Bishop:

I grew up in the post-WWII political consensus in which, instead of the all-or-nothing contest of the start of the Twentieth Century, left and right agreed to cohabit within the established parameters of the state. This consensus allowed newspapers to offer their readers the kind of balanced political reporting I was taught. That possibility has now vanished because the underpinning political consensus has gone.

The Royal Society that I have known also operated smoothly within that consensus. In fact, it helped construct it. In 1928, two years after the General Strike, Time reported:

Very quietly, very significantly a group of men representing 5,000,000 employees and another group representing £1,000,000,000 in invested capital met together at London last week for the first time. They met on strictly neutral ground, in a lofty pillared room at Burlington House, a room hitherto sacred to the high-minded proceedings of The Royal Society (scientific). There, seated around four baize-covered tables, they founded with high hopes the Conference of Industrial Cooperation.

In 1928 neither left nor right demurred from the epistemic claims made by science. Today however, the dark technopopulists do and this makes the conflict far more poisonous for the Society.

Bishop’s demand to oust Musk brings to the surface a tension that is becoming ever harder for the Society to manage. In essence, it has for a very long time ridden two horses, as a voice for science and as a soft arm of the state. These two horses are now pulling in different directions. As an arm of the state, it has to allow for Musk having a powerful voice in the coming Trump administration that will be governing our most important ally; it absolutely does not want to pick a fight with him. At the same time, this is not like 1928 or 1978 or indeed any of the Society’s 364 years. Science faces an entirely new, and in some senses existential, challenge. If the Society is truly its leading academy, it has the responsibility to lead the response to that challenge. In that role, Bishop is right; it must repudiate Musk.

Factional politics in the UK

The line Bishop in effect wants The Royal Society to draw is between the dark technopopulists and the rest of us. Rather than a line between political parties, this is a line between factions that exist in several parties. The task however is complicated by the fact that the Society is not a blank slate in this regard. It already has a factional position, but one that Bishop seems unaware of. Its emergence can also be traced back to 2018.

Here’s a list of all honorary fellows the Society has elected since 2010 (my brackets):

2010-2017

Melvyn Bragg (broadcaster), Bill Bryson (author), Lisa Jardine (historian), Adair Turner (civil servant), David Neuberger (judge).

2018

David Willetts (politician).

2019-2024

Yusuf Hamied (scientist businessman), David Cooksey (venture capitalist), John Kingman (ex-Treasury), Tedros Ghebreyesus (ex-World Health Organization), Kate Bingham (ex-Covid vaccines taskforce), Fiona Fox (science PR), Kwame Anthony Appiah (philosopher), Anthony Hughes (judge).

2018 is the only year in which the Society elected a politician. That politician was David Willetts, the Conservative Cabinet minister for science in the Coalition government 2010-2015 who pushed through the new regime for universities of high student fees, no student number controls, privatisation of the student loan book – and increased funding for science. In 2024 Willetts seems to have more sway over the political direction of the Society than any other fellow, for he is the only one who has been able to persuade the President to publicly adopt on behalf of the Society his own factional position.

This position is that of the “cross-party consensus on science” and it is frequently promoted by Willetts. Adrian Smith leant the Society’s support to it in a speech on 10 April. This might sound like an innocuous echo of the Conference of Industrial Cooperation a century on, but is in fact quite different, for while it is an idea that exists in both parties, it does not include the totality of those parties in the way that the Conference brought together the (near) totality of industrial ownership and the totality of the trades unions; rather, it extends only to a faction within each party.

The phrase originates in that Blair-Hague report, A New National Purpose. Willetts is close to Hague and the overt objective of this document is to make technology itself our new national purpose while shrinking the state and concentrating power on the Prime Minister. The podcast hosted by George Osborne and Ed Balls, which champions and normalises the idea of a cross-party consensus, is part of the same project.

Because the overt goal itself is so bafflingly removed from everyday life, this project can only be understood in the context of the higher purposes it serves: first, for Big Tech, as described above; second, to provide a basis for cooperation between the Blairite wing of the Labour party and the Haguite wing of the Conservative party. This second purpose gives both factions increased leverage against the remainder of their party by creating the potential for a new party that combines the two wings and allows one wing or t’other to remain in office whatever the election result - an arrangement with obvious benefits for any outside sponsor. Because the shared ideological position of the two wings of the project has its roots in Thatcher while the project advances through the idea of technology, I have called this project Thatcherism 3.0. Now we can also see clearly the Big Tech dimension of this project and its character as an alloy of politics and business.

In terms of factions, the line defined by the Thatcherism 3.0 that Smith has endorsed on behalf of the Society is similar to the one Bishop wants in the Conservative Party; it excludes those who want to merge with Reform. However, in the Labour Party its scope is considerably narrower for it excludes the dominant faction of Keir Starmer himself. While the Conservatives excluded by the cross-party consensus tend towards the dark side of technopopulism, the Labour politicians excluded are not technopopulists at all.

It seems to me that the generality of the Society’s fellowship probably agrees with Bishop’s line rather than Smith’s. However, it seems unlikely that the Society will be changing its position any time soon. Paul Nurse, the Nobel prize-winning geneticist who was president of the Society 2010-2015, will again become president next year and moves seamlessly between the Conservative and Labour wings of the project. He worked with Osborne to establish the huge Francis Crick Institute for biomedical research in London. He is also one of three eminent sponsors of Scientists for Labour and an intimate of Kyle’s.

Conclusion

The most basic problem with the kind of approach being promoted by the TBI is that it has already been tried and failed. The out loud aspect wildly inflates the potential speed of impact of technology on public services, and hence the centrality of it in solving day-to-day problems. Just this week, the Labour-leaning think-tank Demos produced a report contesting the TBI’s approach, instead arguing that AI should be relegated to one of many ingredients in a model that “liberates public service professionals to work in partnership with empowered citizens and communities”. And once the techno-utopian aspect is put to one side, the central, unspoken component can be seen to be continuation of the approach to industry that Thatcher pioneered 40 years ago. In this, science is well funded, but the state stays away from the marketplace and privatises public services where it can. The most obvious impact of this has been in the privatisations, including New Labour’s Public Private Partnerships, but the general economic consequences have been even more profound. Under this regime, Britain has suffered two waves of de-industrialisation, first under Thatcher and then under New Labour. This is one of the main reasons why, unlike the 1970s, none of the world’s leading firms are today British and the pay of so many British workers is, in relative terms, so much lower than it was then. It also lies behind the division of the UK into two parts geographically and economically, one centred on a London thriving in post-industrial sectors, the other left behind. In the NHS, it was this approach that led to the fragmentation and privatisation of the digital infrastructure, the underlying reason why after 40 years your GP’s system is still not properly connected to the one in the hospital. Put simply, the approach found in the Tony Blair Institute is not the solution; it is the problem. It is the status quo.

The UK of two parts as found in 'Levelling-up economics' by Philip McCann

Even the core idea of a transformation of public services through a massive technological effort is old for those who remember the National Program for IT in the NHS, launched by Blair in 2002. Old and deeply problematic. As the Substack user Zoltan commented on my earlier essay:

[The NPfIT was] … the largest public-sector IT program ever attempted in the UK. The idea was to integrate the vast array of incompatible systems that dominated most hospitals. The approach was classic 'third way' - or neoliberalism as it was really. Contract out to a quasi-competing group of big tech corps who each had a geographical area (rather like the water companies). Throw billions of public money at it and try to spend your way through the many hurdles and technological problems. Fail. Then scale it down and try again. Rebadge and rename, then claim success on a vastly reduced goal.

At some level, Starmer gets this. By building up the Labour Together think tank, he has limited the influence of the TBI. It is why he says technology in public services is “a tool not a magic wand” and has nudged his biographer to emphasise this difference. However, he lacks any solid capability in tech in Number 10.

There are alternatives to Thatcherism 3.0 that Labour can pursue, albeit lacking the kind of think tank development that $100m buys you:

· By establishing a central technological and design framework, the .gov.uk services have leapt forward in recent years.

· In terms of sovereign AI, a more sovereign recipe using the ingredients provided by the TBI would be to leverage the CrownIntel technology for the ChatGB legal adviser tool and then to invite British companies to build tools on these foundations for use in departmental use cases. That is, the goal would not be “BritGPT” but a monopolistic ecosystem in which the technology stack is curated by the state rather than Silicon Valley.

· HealthTech is an emerging industry in which the UK has a fabulous comparative advantage over other nations and genuine scope to develop indigenous, world-beating firms. Here the transition to AI could be used as an opportunity to erode fragmentation and unify infrastructure and services, to develop tech that is driven from the bottom rather than the top, to nurture an ecosystem of British firms with British jobs and British tax revenues rather than an American ecosystem with American jobs and American tax revenues – to regain sovereignty rather than cede more of it.

The barrier to any kind of genuinely new thinking seems to be not financial or ideological but factional. The more you look at the Green Paper on Industrial Strategy, the more hybrid it seems in its commitments. Commitment to the novel, activist, beyond-Thatcher elements seems limited to Ed Miliband’s Net Zero remit. Elsewhere, the possibility of shift in approach is latent in the document, but not in fact articulated. As Paul Mason has observed, in many respects it does not in fact articulate any strategy at all. The most obvious explanation for this otherwise incoherent combination is simply factional: Starmer has ceded DSIT and DH to the Blairites and, in order to maintain party unity, Johnny Reynolds is not going to step onto their territory with his industrial strategy.

Beyond this basic problem, we have to be wary of the possibility of a more seismic disturbance, of where the Anglo-American fusion of Big Tech and technopopulism currently most visible in the TBI might go. There is no guarantee the novel alloy will stay on the light side, especially if politics in the US gets darker.

As Bishop points out, no fellow has been expelled for 150 years and the Society seems to have no appetite for it. I’m sure this is at least partly because there are sound diplomatic reasons for the UK not to antagonise Musk. To oust Musk, the Society would have to abandon one of the two horses it currently rides, giving up its role as a soft arm of the state. Given the storm heading for science in the UK, and the possibility in the future of a dark technopopulist government, this radical option is perhaps one that should be thought about more. What is it going to do if and when Musk’s allies starting threating its fellows with physical violence?

Bishop’s blog post explaining her reasons for resigning from the Society has itself an air of resignation, which I think is not needed. In taking her stand she has, at the political level, potentially snatched victory from the jaws of defeat. Unlike so many others she clearly understands that today science and democracy have enemies and that, as the political philosopher Alan Finlayson explains in this podcast, they have to be fought. By virtue of her past achievements, political intuition and moral courage, she has the potential to become the voice that science needs today. For the sake of our democracy, let’s hope she finds it.

Note: This article was updated to make more accurate the description of Oracle on 1 December 2024 and of Paul Nurse’s role in Scientists for Labour on 5 December.