Alan Turing Institute chief urges UK to build sovereign AI capacity

The UK should not rely solely on foreign AI systems that it cannot inspect, control or guarantee access to, according to George Williamson, head of the Alan Turing Institute. He said the US and China’s lead in the field has made national resilience a central priority for the institute, especially as AI becomes embedded in defence, national security, critical infrastructure and healthcare.

Williamson pointed to Washington’s temporary block on the overseas release of Anthropic’s latest models and reported limits on UK testing as signs that political and commercial pressures could constrain access. He argued that the UK needs more agency over key technologies and should develop strategic AI capability that does not require permission from others to use.

The institute plans to focus on securing critical systems, countering AI-enabled threats and supporting public-interest applications such as weather forecasting, farming and disinformation monitoring. Anthropic has predicted the technology had a greater than 10% chance of wiping out humanity by 2030, but Williamson questioned the usefulness of assigning precise percentages to existential risk without clear scientific grounding.

Williamson said the resilience agenda also keeps climate work within the institute’s remit. In 2025 the then technology secretary, Peter Kyle, pushed the institute to prioritise defence and security, but Williamson described climate as a national security issue and said the UK can be ambitious without copying frontier AI labs.