Britain should not depend on foreign AI that could be switched off, the head of the country's top AI research body says.

George Williamson, the new head of the Alan Turing Institute, told The Guardian the U.K. needs a stronger position in AI as the U.S. and China race ahead. "National resilience" will be a key focus for the institute, he said.

"We shouldn't depend solely on systems we can't inspect or control, or that could be restricted or even switched off," Williamson said, The Next Web reported.

He said AI is now built into defense, national security, critical infrastructure and health care. If access to foreign AI were cut, those services could be at risk.

His warning follows a move by Washington in June to block foreign access to Anthropic's top AI models for a time, The Next Web reported. That showed how quickly access can change for political or business reasons.

The Alan Turing Institute is named after the British mathematician and computing pioneer. It is the U.K.'s leading AI research body. It is under new leadership after the government, its main funder, demanded a change in direction, The Guardian reported.

Many countries are now talking about "sovereign AI," meaning AI they build and control themselves. Supporters say it protects national security. Critics say smaller countries cannot match the huge spending of U.S. and Chinese tech giants and would do better to work with allies.

Britain has tried to attract AI companies and build data centers. But most of the leading AI models are made by American firms such as OpenAI, Google and Anthropic.

What we know: Turing Institute head George Williamson says Britain needs its own AI for defense, health care and critical infrastructure.
What we don't: How much Britain would spend on sovereign AI, or which systems would come first.