Keir Starmer should pledge £11bn towards building “BritGPT” and a national artificial intelligence (AI) cloud in the next Labour manifesto or risk the UK falling ever further into dependence on American tech companies, an affiliated thinktank has said.
Labour for the Long Term, which campaigns within the party for it to adopt “long-termist” policies that mitigate dangers such as pandemics, climate breakdown, and AI extinction, argues in a report that the £1bn pledged by the government in the 2023 budget is not enough to protect Britain’s future independence.
The report calls for the creation of BritGPT, a homemade system with a remit to focus on market failures rather than simply trying to compete with Silicon Valley to build the biggest models.
“Private profit-seeking companies aren’t going to invest enough in ‘AI for good’ or AI safety, so the UK government should step in to correct this market failure and provide more public goods – such as medical research, clean energy research, and AI safety research,” it said. They suggested some of the budget could even come out of Labour’s £28bn annual “climate investment pledge” as a result.
“This is a hugely important technology, arguably the most transformative in the next few decades, and the UK risks being left behind,” said Haydn Belfield, associate fellow at the University of Cambridge’s Leverhulme Centre for the Future of Intelligence, said.
The government has pledged £100m to train new “foundation models”, similar to the GPT-4 system that underpins ChatGPT, and a further £900m on a new “exascale” supercomputer for similar work. But, Belfield warns, those numbers are an order of magnitude too small.
“Building up datacentres to make a new cloud region,” the sort of investment a company such as Amazon or Google makes to launch their services, “costs in the region of £10bn. And GPT-4 alone probably costs about $100m, and if you look at the cost trends, these are increasing rapidly: we should expect GPT-5 or GPT-6 to cost in the hundreds of millions of pounds, even before you account for salary costs. That’s what it takes to compete at this level, to support British companies and the British ecosystem.”
At the physical infrastructure level, a £10bn “Great British cloud” would mirror Labour’s pledges to establish Great British energy and to bring private rail franchise back into public ownership, and be comparable to the creation of the BBC and Channel 4.
Labour for the Long Term is not alone in calling for more state-backed investment in AI. In an interview with the Guardian earlier this month, Geoffrey Hinton, the co-inventor of “deep learning”, warned that AI development could doom humanity if it was pursued for purely commercial motivations.
“Google is the leader in this research, the core technical breakthroughs that underlie this wave came from Google, and it decided not to release them directly to the public,” Hinton said. Google was worried about all the things we worry about, it has a good reputation and doesn’t want to mess it up. And I think that was a fair, responsible decision.
“The problem is, in a capitalist system, if your competitor then does do that, there’s nothing you can do but do the same.”
