Nobody could accuse the British Government of lacking enthusiasm for artificial intelligence (AI). According to its newly published AI Opportunities Action Plan, the technology’s blossoming will remake the British economy, boost productivity, smooth bureaucracy, and transform the quality of state services such as healthcare.
Through this, the government wants the UK to become an AI superpower, consolidating its position as the world’s third largest AI market, behind the US and China, the report said.
Part of the initiative is a £14 billion investment ($18 billion) by Vantage Data Centres, Nscale, and Kyndryl to build new data center infrastructure, on top of the £25 billion in AI investment announced at the International Investment Summit last October.
Its deeper theme is that the UK should be able to produce as well as consume AI technology, because receiving technology from others is a recipe for dependence.
“There is barely an aspect of our society that will remain untouched by this force of change,” said Prime Minister Keir Starmer in his report foreword.
“This government will not sit back passively and wait for change to come. It is our responsibility to harness it and make it work for working people,” he said.
AI Growth Zones (AIGZs)
Many of the 13,250 jobs the reports says will be created in the near term will be in “AI growth zones”, the first of which will be in Culham, Oxfordshire, also the HQ for the UK Atomic Energy Authority (UKAEA).
That’s significant because all of the data centers housing AI will consume huge amounts of power, at least 100MW to start with and eventually 500MW. Unfortunately, while Culham is the home of research into nuclear energy, there is no power station onsite, which explains why the report proposes a new AI Energy Council to fill the gap.
Cooling the new data centers could also be an issue in a potential drought zone, despite Thames Water promising to build a new reservoir nearby.
Additional requirements
National Data Library: AI, of course, needs data – lots of it. That’s becoming harder to source. The Report’s answer is a National Data Library, “underpinned by strong privacy-preserving safeguards.” The deeper detail of this has yet to be announced, but issues of privacy and copyright lurk.
Public compute: The UK needs another supercomputer, somewhere. This is despite the new government ditching plans as recently as October to build precisely this type of computer at Edinburgh University. The UK’s other supercomputer centers include Bristol (Isambard AI), and Cambridge (Dawn).
People: The plan proposed working out how many people the UK needs and then developing plans to fill gaps. That includes importing skills by exploring “how the existing immigration system can be used to attract graduates from universities producing some of the world’s top AI talent.”
Regulatory oversight: AI has yet to grapple with its daunting ethical concerns. The report’s answer is to turn the UK’s AI Safety Institute (AISI) into a statutory body with the power to intervene where it thinks fit.
White hot
The idea of a bold national plan for AI echoes that of another Labour Prime Minister, Harold Wilson, who in a famous 1963 speech proposed overhauling the UK through the “white heat of technology”.
This parallel isn’t necessarily happy; although the UK embraced Wilson’s white heat in pockets, it failed to capitalize on the wider opportunities offered by that era’s early development of computing and software. That was despite having plenty of clever people in the field, a decent education system, and big companies willing to invest.
The new plan is also uncomfortably similar to the recent Conservative administration’s championing of everything AI under Prime Minister Rishi Sunak, right down to his rhetoric about matching AI superpowers AI and China.
One difference is that the report’s vision was authored by industry figure Matt Clifford, an entrepreneur and investor appointed as the administration’s AI adviser after Labour’s election win in July.
He authored the plan’s 50 recommendations, all of which the government plans to implement. That suggests that the UK’s AI overhaul is not being driven as a pet project by politicians alone. Another backer is Demis Hassabis, co-winner of the 2024 Nobel Prize in Chemistry for his AI-aided work on protein folding.
Making not taking
In many ways, the challenges faced by the UK in enabling AI – indeed, all future computing projects – mirror what every country faces. Energy and cooling are problems, and tooling up to fill infrastructure and skills gaps is competitive and resource intensive.
One reason Wilson’s 1960’s plan struggled was that the state and public services ignored the computing innovation going on around them. However, given that AI is already being experimented with in the NHS, perhaps attitudes have changed.
But what will really count is that private companies can make money out of AI in the UK in the long run. Governments plans and infrastructure are helpful and set the mood, but only go so far.
“While the focus on investing in infrastructure such as computing power and a national data library is welcome,” said Dr Pia Huesch, Royal United Services Institute (RUSI) research fellow, “the UK Government must not forget the risks posed by AI technologies or the international partnerships that are needed to secure long-term benefit from AI technologies.”
