We’re looking at what could become a major reset coming for Britain’s AI strategy. According to the Financial Times, Andy Burnham, who is tipped to be the next UK Prime Minster, is exploring a shift away from the current policy towards one built around UK AI sovereignty, regional growth and re-skilling, as opposed to the more AI-forward approach we’d become used to under Sir Kier Starmer, who had wooed US tech giants like OpenAI and Anthropic, while promoting of home-grown AI companies. Indeed, it even launched a £500 sovereign AI fund. But are we about to see the rise of a far more tech-sceptical UK administration? And what does the UK tech industry want to see next?
Featured on this podcast:
Host: Mike Butcher, Founder & Editor, Pathfounders.com and former Editor-at-large of TechCrunch. (X, LinkedIn)
Guest: Alys Key, an Editor, Writer and Journalist, who writes the Substack newsletter, UK2.0. (X, LinkedIn)
Guest: Meryem Arik, Co-founder and CEO at Doubleword, which builds high-performance open source model inference infrastructure for AI at scale. (X, LinkedIn)
Will Burnham’s AI reset slow Britain down?
Andy Burnham’s mooted reset of Britain’s AI strategy has raised alarm among some UK tech figures, who warn that the country cannot afford to lose momentum in the global race for AI infrastructure, talent and sovereign capability.
Speaking on the Pathfounders podcast, Alys Key, editor of the UK 2.0 newsletter, and Meryem Arik, co-founder and CEO of AI infrastructure startup Doubleword, said the ideas floated around a potential Burnham administration did not yet add up to a coherent technology policy.
The Financial Times reported that Burnham’s advisers are exploring a shift away from the current government’s approach to AI, with a stronger focus on British companies, regional growth, reskilling, public accountability and the impact of automation on workers.
But Key said her first reaction was “a lot of frustration,” particularly at suggestions that the UK had taken a reckless approach to technologies such as self-driving cars.
“It was frustrating to see things like self-driving cars described as having been a ‘gung ho’ approach to getting those introduced, when in fact the regulatory process has actually been extremely slow and careful on that,” she said.
She cautioned that the report may amount to political kite-flying rather than settled policy.
“I’m probably seeing this as a bit more similar to when we’ve had budgets in the past where several lines from them will be leaked in advance, and several of them will never actually make it into the real budget,” Key said. “It’s more of a let this out into the world, see what the reaction is.”
Arik said it was odd to criticise the Starmer government’s AI policy as too US-focused when it had also marked the first serious UK push into sovereign technology.
“It feels like a strange thing to try and diverge with him on,” Arik said. “It’s the first time we’ve really ever had conversations about sovereign tech and taking steps towards supporting homegrown champions.”
She added that some US-facing parts of the strategy had delivered practical benefits.
“The things that we have done that have been US-focused have actually been very smart growth decisions, for example Anthropic hiring a bunch of hugely high-paying jobs, which we desperately need,” she said.
Arik said the framing of issues such as self-driving cars made her question who was advising Burnham.
“It makes me concerned about who is advising Andy Burnham and whether they really know exactly what they’re talking about,” she said. “A lot of the things that were mentioned in this article just seem really far away from the necessity of what’s needed in terms of our position in the global AI market.”
Key said the economic impact of automation on drivers and gig workers was a legitimate issue, but warned that the language used around autonomous vehicles had irritated the sector.
“The worry about people who will lose their income because they’re Uber drivers or for a similar service — that’s an interesting one,” she said. “That’s a perfectly valid point to raise.”
But she added: “It was the quote specifically about this, like, ‘who needs this, who is this for?’ that really riled a lot of people up.”
Another big concern was infrastructure.
Arik warned that devolving more AI decision-making away from Whitehall could slow down the build-out of data centres and energy capacity, where the UK is already behind.
“My biggest concern is whether this will dramatically slow the development of data centre infrastructure and energy infrastructure,” she said. “The UK is already incredibly far behind when it comes to the amount of data centres that we have and the amount of capacity that we have.”
She said planning delays were already a major barrier.
“One of the biggest costs when building data centres is actually delays,” Arik said. “A year delay or year and a half delay is far more costly than energy or anything else.”
Key said there was a clear distinction between policies that could be devolved and those that should remain national.
Reskilling and local industrial strategy could sensibly be shaped by regions, she argued. But data centres, energy and AI infrastructure were matters of national security.
“There’s the aspect of AI that is of national security and geopolitical importance, and that has to do with our energy infrastructure and building up data centres and ensuring we’re competitive,” Key said. “That’s not a good idea to make devolved.”
On reskilling, Arik said local approaches could work if they helped regions identify where they can compete.
“Every area needs to figure out the industries that they can be competitive in and work to build them up,” she said. “Reskilling is a great way to figure out how you can be competitive in a brand new industry.”
But both warned that the UK’s biggest risk is losing time.
“What people are most worried about is losing time,” Key said. “In this AI race that is very fast, a reshuffle would just mean losing six months of time.”
She said the current AI minister, Kanishka Narayan, had won respect in the sector because he “knows what he’s talking about,” while a less experienced replacement could take months to get up to speed.
Arik agreed that another strategic review would be a mistake.
“The most important thing is that we don’t lose time,” she said. “This is moving incredibly quickly and is of real importance to our overall position as a country on the world stage.”
She said the UK should keep implementing Matt Clifford’s AI action plan rather than reopen the question.
“He spent a lot of time and effort getting that right,” Arik said. “Doing another review for the sake of it will not be productive and will waste time.”
Arik also said AI should be treated as a cabinet-level priority.
“I would actually like an AI minister to be a cabinet level position, because I think it is that important to our future,” she said.
For now, Key said the Burnham ideas remain hard to read as a programme.
“None of it is quite gelling together,” she said. “Are we investing more, or are we saying let’s own all our own data centres and not rely on anyone else? Or are we saying there’s been too much attention paid and we should stop having all these government rush-into-AI programmes?”
Arik put it more directly: “This just doesn’t add up into anything really coherent.”
