The US Blocked Britain's AI Access. Now the UK Wants Its Own.

Key points

• The Trump administration issued an export control order in early 2025 banning foreign nationals from accessing Anthropic's Fable 5 and Mythos 5 AI models, then lifted the ban with no public explanation.
• The UK's Science, Innovation and Technology Committee warned last week that "the whim of a foreign government" could cut off Britain's access to vital technologies.
• The UK's National Cyber Security Centre and Department for Science, Innovation and Technology jointly published a "Cyber Shield" strategy, which plans to use advanced AI to defend against national cyber threats.
• A Proton risk report found that more than two-thirds of businesses in the UK, Spain, and France run primarily on US technology companies.
• Analysts warn that US providers control roughly 70 percent of the European cloud market, making rapid tech independence a steep climb.

Earlier this year, the Trump administration quietly issued an export control order, a legal instrument that restricts who can access certain technology for national security reasons, blocking foreign nationals from using two Anthropic AI models called Fable 5 and Mythos 5. Anthropic suspended access. No public explanation followed. The ban was eventually lifted, again in silence.

For the cybersecurity community, it was annoying. For the UK government, it was a wake-up call.

Last week, the UK's Science, Innovation and Technology Committee published a paper warning that Britain "may not be able to count on even its allies for access to vital technology." The paper noted that the same administration had also restricted OpenAI's latest models to a handful of hand-picked companies. The committee's word for the situation: blunt.

Should British organisations be worried about losing access to AI tools?

Yes, at least in theory, and that is precisely the point the UK government is now making publicly. Raphael Auphan, chief operating officer of Swiss technology firm Proton, told Dark Reading that if access to advanced AI can be switched off by a foreign government, whether for security, commercial, or political reasons, it creates a genuine strategic weak spot for any country that depends on those tools.

That dependence is not small. Proton's own research found that more than two-thirds of businesses across the UK, Spain, and France run primarily on US tech. Rik Turner, chief analyst at research firm Omdia, puts the cloud picture in starker numbers: Amazon Web Services, Microsoft Azure, and Google Cloud Platform together hold around 70 percent of the European market. The continent's largest home-grown cloud provider, OVH, sits somewhere between one and six percent.

AI is even more lopsided. Turner describes Europe's realistic choices as "almost entirely between US closed models, from the likes of OpenAI and Anthropic, or Chinese open ones, from DeepSeek and Alibaba." French startup Mistral AI exists, but analysts are quietly sceptical about whether its models qualify as truly frontier-level.

The UK's answer, at least on the defence side, is a programme called Cyber Shield. The National Cyber Security Centre and the Department for Science, Innovation and Technology published the strategy last week, proposing to build an AI-powered national defence system capable of identifying and neutralising cyber threats at machine speed. The NCSC is candid that this will require partnerships with private AI firms, cyber defence organisations, and universities; it is not a go-it-alone plan.

Louise Horton, head of UK government affairs at security firm NCC Group, says that is the right instinct. Sovereign control over decision-making and data matters, she argues, but the actual technology can still come from trusted outside partners provided the governance is genuinely British.

For ordinary people, the immediate practical risk is low. Nothing has broken yet. But if your employer, hospital, school, or local council uses US-built AI tools, it is worth knowing that access to those tools is, technically, subject to decisions made in Washington. Asking your IT team which critical services have a backup plan is a reasonable question, not a paranoid one.