UK must not be beholden to foreign AI, Alan Turing Institute head says

The UK must not become dependent on foreign AI systems that can be switched off at short notice, and must build its own versions of the technology, according to George Williamson, head of the Alan Turing Institute.

Williamson's argument rests on a simple dependency scenario: a country that runs critical functions on AI systems controlled elsewhere can lose access to those systems abruptly. "National resilience" is now a key focus of the Alan Turing Institute (ATI), he said, as the institute positions itself in a field dominated by two states.

The stakes he names are structural, not rhetorical. The US and China hold what he described as runaway leadership in AI. Britain, on this reading, is not competing for first place in the current race but for the ability to keep operating independently if access to leading foreign systems is cut off.

Who is making the argument?

George Williamson leads the Alan Turing Institute, the UK's national institute for data science and artificial intelligence. His statement carries institutional weight precisely because ATI sits at the intersection of UK research, government and industry. When its head frames foreign AI dependency as a national problem, that framing can shape research priorities and policy debates.

Williamson is not arguing that the UK should refuse to use foreign AI. His stated concern is dependency — the condition in which the country has no domestic alternative if outside systems become unavailable.

What does "switched off at short notice" mean in practice?

The phrase Williamson uses describes a scenario in which a foreign provider — whether a company or a state-linked entity — withdraws access to an AI system quickly. The UK's exposure, in his framing, comes from having no comparable systems of its own.

That concern follows from the market structure he points to. If the leading AI systems are built in the US and China, then countries that rely on them are exposed to decisions made in Washington, Beijing or the boardrooms of foreign companies.

The remedy he proposes is direct: the UK must develop its own versions of the technology. Not imports. Not licensed access. Domestic capability.

Why does this matter for UK policy?

Williamson's statement lands in an ongoing British debate over how much the country should invest in sovereign AI capability. His diagnosis — US and Chinese "runaway leadership" — sets the baseline for that debate. The UK is not starting from parity; it is starting from behind.

The ATI's own focus is shifting accordingly. Williamson said national resilience is now a key focus of the institute, which signals that questions of technological sovereignty are moving into the mainstream of the UK's AI research agenda rather than remaining a niche security topic.

What is at stake if the UK stays dependent?

The core risk, as Williamson states it, is abrupt loss of access. A country dependent on foreign AI systems that "can be switched off at short notice" faces two problems at once:

• Operational disruption. Systems embedded in UK workflows stop working when access ends.
• Strategic weakness. The UK has no domestic substitute to fall back on, because it has not developed its own versions of the technology.

Both problems trace back to the same root cause in his account: the absence of a stronger domestic position in AI, combined with the dominant positions of the US and China.

What comes next?

Williamson has framed the goal clearly: the UK must develop its own AI technology so it is not beholden to foreign systems. With national resilience established as a key ATI focus, the open question is whether that priority translates into concrete UK investment in sovereign AI capability — and whether it does so before dependency deepens.