MPs warn UK cannot rely on US for AI access

The UK may not be able to count on even its closest allies for access to vital AI technology, the House of Commons Science, Innovation and Technology Committee warned in a report published on Tuesday (7 July).

The committee pointed to the US government’s decision in June 2026 to place export controls on Anthropic’s latest AI models, banning non-US citizens from accessing them, followed by restrictions limiting access to OpenAI’s newest models to a small group of US companies and organisations approved by the Trump administration.

“The US’s decision to restrict some advanced AI models should be a powerful reminder to the UK government that it may not be able to count even on its allies for access to technologies,” the committee said. “Leverage may not be enough to guarantee sovereignty.”

MPs said it was essential to ensure the UK could not be cut off from key technologies “at the whim of a foreign government”, and that this could require diversifying partnerships to secure multiple points of access and developing sources of leverage outside a specific tech stack, or outside tech altogether.

The report found that the the government had repeatedly described building sovereign capability as a “critical priority” but had failed to explain what this meant or how success would be measured.

Prime Minister Keir Starmer told the Liaison Committee that publishing a list of sovereign capabilities risked “inadvertently telegraphing our specific vulnerabilities to hostile actors”.

The committee rejected this reasoning. “Limiting transparency around the UK’s ambitions for sovereign capabilities is unlikely to be hiding anything from competitors such as Russia and China, but it is leaving UK tech companies in the dark,” it said.

Minister for AI and Online Safety Kanishka Narayan said the “main lesson” of the US restrictions was that “access to AI capabilities is crucial”, pointing to the Sovereign AI Unit and a £1.1 billion plan to invest in AI hardware announced on 8 June 2026.

The committee recommended the government work with industry and academia to define sovereign capability for the five critical technologies in its 2025 Science and Technology Framework, supported by detailed supply chain analysis, and publish the results or share them with the Intelligence and Security Committee.

It also said the DeepMind sale illustrated the UK’s failure to retain innovation at scale. The London-founded company was acquired by Google in 2014 for between £400 million and £650 million and is now central to development of Google’s Gemini AI models.

“Had the UK been able to provide the scale-up investment required, the company might not have been sold to Google, and the UK would not need to rely on overseas firms for its generative AI capabilities,” the committee said.

Now read: HMRC built an AI to argue with before you argue with a human