UK examines economic hit from loss of access to frontier AI models

The UK government is urgently assessing the economic and security impacts of consumers losing access to the latest frontier AI models, amid growing concerns about the impact of the threat on growth.

The Cabinet Office has asked the Department for Business, Innovation, Science and Trade to examine the hit to the UK economy if Britons were blocked from new versions of AI models launched by companies such as Anthropic and OpenAI, according to people familiar with the matter. The work is expected to be completed within weeks.

US President Donald Trump’s directive in June that demanded foreign national access to Anthropic’s Fable 5 model be rescinded had fuelled alarm within Britain’s National Security Secretariat about future access, the people said.

While Anthropic relaunched the model globally after temporarily rescinding it, the episode has heightened concerns in Whitehall about the prospect of Britons being denied access to a cutting-edge model available in other locations in future.

Economic security officials want to explore the potential hit to the economy if UK companies could not exploit the latest innovations, as well as the security implications if cyber vulnerabilities in British companies could be exploited by foreign users of cutting-edge AI models before being patched.

The assessment comes as the Home Office has also been reviewing the prevalence of AI in criminal methods, according to government figures.

One government figure said: “When a new AI model is launched, all businesses are immediately looking at how to best use and exploit it. For the UK to be left behind by not having access — even for a couple of weeks — would cause a hit to the economy.”

US Big Tech companies are betting hundreds of billions of dollars that AI will yield huge productivity gains, with the latest systems already able to automate many aspects of white-collar work from software development to routine legal and accounting tasks.

But even though the impact of AI on economic growth is unclear, the cyber security capabilities demonstrated by the latest models from OpenAI and Anthropic have triggered alarm among defence and security experts.

Dame Chi Onwurah, Labour chair of the House of Commons science, technology and innovation committee, said: “It is good that the new government is assessing the impact of being cut off but we also need to see a plan to address unacceptable levels of dependency, and indeed how an acceptable level of dependency is defined.”

Her committee had “repeatedly emphasised that the government needs a realistic plan to achieve sovereign capabilities in critical areas or risk having its access cut off at the whim of its partners”, Onwurah added.

However, Ciaran Martin, former chief executive of the National Cyber Security Centre, said he found the export control experience regarding Fable 5 “reassuring” in some respects.

“What was the consequence? Anthropic took the whole thing off the market, as it couldn’t meaningfully distinguish between nationals and foreigners,” he said.

Recommended

Martin, who is now a professor at Oxford university, added that he considered it “slightly implausible that an American AI model would be restricted from foreigners”.

But he warned: “The US’s ability to manipulate and leverage the whole AI stack – to try and force countries to bend to its will – could be significant.”

Officials said the UK government was investing to strengthen Britain’s AI capabilities via its £1.1bn AI hardware plan and £500mn sovereign AI fund.

The government said the state took “national and economic security seriously and considers and plans for a whole range of scenarios” and that AI was “at the heart” of its agenda.

“We now have an AI minister at the cabinet table, leading the new AI task force at the centre of government. We’re also building the computing power needed to develop and use advanced AI in Britain, and backing UK companies developing the next generation of AI technologies.”

Additional reporting by Tim Bradshaw