UK Risks Becoming an 'AI Customer' Rather Than a Creator, Top Advisers Warn

LONDON — Britain risks becoming a "high-value AI customer" rather than a country able to shape artificial intelligence, unless it takes urgent action to build its own capabilities, the government's top science advisers warned on Monday, 19 October 2026 — a warning the government answered days later with about £600 million in new AI security funding.

Illustration: London and a data centre under neural-network patterns, symbolising Britain's AI ambitions. This is an AI illustration, not a photograph of a real facility.

What Happened

The Council for Science and Technology, an independent advisory body that advises the government on science and technology policy, said decisions made now would determine whether Britain captures the economic and public benefits of AI while maintaining resilience and strategic autonomy, according to Reuters reporting on its letter to AI Minister Kanishka Narayan.

"The way for us to achieve AI sovereignty is to build trusted systems that deliver real public, economic and security benefits," the advisers said in the emailed letter. Without what they described as "deliberate and urgent action," they warned, Britain was at risk of falling behind and becoming a "high-value AI customer, rather than a country able to shape the technology and its usage."

The warning reflects a global race. Governments around the world are building AI capabilities, driven by concerns that countries without their own expertise, computing infrastructure and models could become increasingly dependent on a handful of mostly US-based companies that dominate the technology.

The council recommended appointing a single government lead to coordinate AI sovereignty policy, investing in skills, data and computing infrastructure, and testing AI systems in areas including national security, health and education. It also said Britain should maintain access to the world's most advanced AI systems while developing some domestic capabilities of its own, including exploring a publicly governed British AI model that could be made available to researchers, startups and public services. The government did not immediately respond to a request for comment from Reuters at the time.

Days later, the government set out its response in funding terms. On Saturday, 24 October 2026, Britain's government set out around £600 million (about $800 million) of funding largely geared towards defending against health and security threats from artificial intelligence, according to Reuters. The package, announced by AI Minister Kanishka Narayan, was presented as a step towards making Britain a leader in safe AI, not just a buyer of other countries' systems.

Background

Britain is not home to any major AI developers, but it has tried to punch above its weight on safety. Its AI Security Institute, set up after the world's first AI safety summit in 2023, tests the most advanced AI models to evaluate their capabilities and the risk of misuse, under voluntary agreements with major developers that give it access to models before deployment.

That access was itself in the news in October. On Monday, 12 October 2026, Narayan said Britain's AI Security Institute had been given access to the latest advanced AI models and was working with the US government and major developers, including Google, to assess their safety. "I can confirm that AI scientists at the AI Security Institute are now conducting rigorous evaluations of the latest generation of models, including Gemini 4 Argon, to identify risks early, to test robust safeguards, to help keep people safe," he said, according to Reuters.

The announcement came after reports the previous month that the White House had asked OpenAI and Anthropic to withhold new models from British testers pending a US review, and after the Telegraph reported on Sunday, 11 October 2026, that the Trump administration had reversed that position, giving US developers greater latitude to cooperate with Britain. Spokespersons for Trump, OpenAI and Anthropic did not immediately respond to a request for comment from Reuters on that report.

Scrutiny of AI safety has intensified after developers including Anthropic and OpenAI disclosed instances of rogue or undesired behaviour by AI agents, Reuters reported. Some industry figures have warned the technology could pose an existential threat to humanity, while others argue the greater near-term risk is concentration: a few firms controlling the models, chips and data centres on which everyone else depends.

The council's letter sits squarely in that second debate. Sovereignty, in its framing, is not about building everything alone. It is about having enough domestic skill, infrastructure and testing capacity to judge, adapt and, when needed, challenge the systems Britain imports — and to offer trusted alternatives in public services, research and startups.

Illustration: Scientists testing an AI system, symbolising independent safety evaluation. This is an AI illustration, not a photograph of a real laboratory.

Why It Matters

Being a "customer" rather than a "creator" is not just a matter of national pride. If a country cannot test models itself, it must take vendors' safety claims largely on trust. If it lacks computing capacity, it cannot run, adapt or audit large models at scale. And if it lacks skilled people, it cannot build the public-service tools — in health, education and national security — that the council says should be tested first.

Dependence also has economic costs. The value in AI is concentrated in the firms that own the models and the infrastructure. A country that only buys finished products captures less of that value, pays more over time, and has less say over prices, access and terms. The council's warning is, in effect, that Britain could end up paying a premium for a technology it helped pioneer but no longer shapes.

The funding package announced on 24 October shows where the government thinks the balance lies. The biggest chunk goes to a new biosecurity initiative, including £100 million for a laboratory using AI to detect biological threats and £65 million to develop countermeasures against pathogens. Narayan said the aim was for the AI Biosecurity Lab to become a major global institution "bringing together world-leading scientific expertise to protect people from emerging biological threats."

Tech companies OpenAI and Anthropic will offer up to £200 million of model credits to the biosecurity initiative, Reuters reported, while Google DeepMind and London-based AI life sciences startup Basecamp Research will offer other assistance. Britain's government will give its existing AI Security Institute an extra £120 million. A further £130 million will go towards new projects to improve cyber defences for critical infrastructure and the wider economy, known as the National Cyber Shield and the National Exposure and Real-Time Vulnerability Centre.

Smaller sums fill out the package: £60 million towards a fund investing in British companies using AI to advance robotics, autonomous systems and manufacturing; £60 million for AI research and procurement contracts; an extra £23 million for research on the economic impact of AI; £18 million for a new chip design academy; £15 million for a prize for AI solutions to public policy problems, initially focused on healthcare; and £30 million for regional mayors to support local AI work. Britain's National Cyber Security Centre will receive an unspecified sum to expand in Manchester.

What Happens Next

The test for both the advisers' warning and the government's funding will be implementation. Appointing a single lead for AI sovereignty, as the council recommended, would be a visible first step; without it, the many strands of funding — biosecurity, cyber defence, robotics, skills and procurement — could remain scattered across departments.

Testing will also matter. The council wants AI systems tested in national security, health and education, and the AI Security Institute's continued access to frontier models — including Gemini 4 Argon — will determine whether Britain can credibly evaluate the systems it deploys. Any future interruption to that access, of the kind reported in September, would revive the sovereignty debate quickly.

Finally, the idea of a publicly governed British AI model, floated by the council for researchers, startups and public services, will be watched closely. If it progresses, it could give Britain a practical alternative to purely commercial models in sensitive areas. If it stalls, the "customer" warning may prove harder to shake — and the £600 million package, large as it sounds, will be judged not by its headline total but by whether it leaves Britain with skills, infrastructure and institutions that outlast a single funding round.

Illustration: People training in computing and robotics, symbolising AI skills and domestic capacity. This is an AI illustration, not a photograph of a real training centre.

Sponsored: explore today's featured sponsored offers

FAQ

What did Britain's science advisers warn?

The Council for Science and Technology warned on 19 October 2026 that, without deliberate and urgent action, Britain risks becoming a "high-value AI customer" rather than a country able to shape AI and its usage.

What did the advisers recommend?

They recommended a single government lead for AI sovereignty, investment in skills, data and computing infrastructure, testing in national security, health and education, and exploring a publicly governed British AI model for researchers, startups and public services.

How did the government respond?

On 24 October 2026 the government set out around £600 million (about $800 million) of AI security funding, including a biosecurity initiative, extra funding for the AI Security Institute, cyber defence projects, and support for robotics, research and skills.

What is the AI Security Institute?

Set up after the world's first AI safety summit in 2023, it tests the most advanced AI models under voluntary agreements with developers, giving it access before deployment to assess capabilities and misuse risks.

Why does AI sovereignty matter for ordinary people?

Without domestic testing, skills and infrastructure, a country must largely trust vendors' safety claims, captures less economic value from AI, and has less control over the tools used in health, education and public services.

Sources

• Reuters, "UK risks becoming 'AI customer' rather than creator, advisers warn," 19 October 2026.
• Reuters, "UK government sets out $800 million of AI security funding," 24 October 2026.
• Reuters, "UK given access to safety-test latest AI models, minister says," 12 October 2026.
• Reuters, "US reverses course on Britain's testing of AI models, Telegraph says," 11 October 2026.