Data centres and quantum culture wars
One of the most edifying aspects of the US midterms has been watching a number of virulently anti-science, pro-health (sic) candidates get rejected by the American electorate.
More germanely, much of the coverage has framed the results as a referendum on Trump first and then AI second, with almost every anti-data centre ballot measure sweeping into state law wherever voters were given the chance to do so. For those readers who have been doing their best to ignore politics across the Atlantic for the last couple of years, it really can’t be understated how artificial intelligence has fractured across party lines over there, with Democrats against and Republicans in favour.
The next question – and one which is pretty important for UK science – is whether the same political divisions will emerge here. Treasury minister Dan Tomlinson has taken to political-division-manufacturing-machine and erstwhile social network X to suggest that no they will not:
“A common Westminster mistake is to look at big moments in US politics and assume the same battle lines will be drawn here. AI and the mid-terms are a good example of this. The easy assumption is that a big polarising debate on data centres is about to hit the UK, just because it has in the US.”
Tomlinson argues that Westminster has learned lessons from the data centres “free-for-all” in parts of the USA, and the solution to building more infrastructure for AI lies in the government’s new data centre covenant. This somewhat woolly set of principles for construction was in fact one of the more tangible announcements at last month’s Budget (unless you count the further raiding of departmental capital funds, which to be fair is among the most consequential kinds of policymaking we have in this country).
The covenant is a great example of the UK public sector’s “everything-ist” approach to dealing with problems, where it attaches many many additional riders to contentious processes – lower local electricity bills! create apprenticeships! pass your waste heat along to hospitals! – in a way that on paper might please everyone but in practice seems to please no-one.
The current Labour government might think that it can walk the tightrope on data centres without it becoming a left-right issue, but whatever happens it clearly is going to become ever more fractious. This in turn creates sensitivities around R&D spending, to the extent that a growing chunk of the pot goes to compute, AI-related research, and companies working in the space (and getting pretty rich as well).
On the left of parliamentary politics there is already pressure to stop building data centres – witness this early day motion calling for a moratorium – and there are protests taking place around the UK throughout the month. On the other hand, the prospects of the squabbles becoming strictly partisan are complicated by the fact that some of the most pro-AI voices in think tank land – in particular the Tony Blair Institute – are Labour-aligned.
Anyway, what this really got me thinking about was whether the coming shape of the debate was clear say five years ago. In retrospect was it not obvious that a technology tied up as it is with employment, local development, natural resources, profit, copyright, and the USA would create this particular schism?
Well, probably not. But the question is then what happens with the next big technology: where will the faultlines be for – to choose a topical example – quantum? This is a big area of investment for UKRI at present; at £426m over the spending review, it’s the third biggest discrete area, after AI and life sciences.
Yet if you were to ask me what are likely to be the political arguments over it, and which parties would be most likely to want more spending, or more regulation, or more public input, I would find it hard to sketch out an answer. Seems like the kind of question worth thinking about though.
Spotted a politician meddling in the research system? Or a blurry line between academia and government? Email the man himself: haldane [at] researchagenda [dot] news