Minister warns UK preparing to strengthen AI safety reporting requirements

In a speech on Monday in West London, Kanishka Narayan, the UK minister for artificial intelligence (AI), called on companies developing frontier AI models to increase safety reporting and oversight, warning that the government is prepared to legislate. 

Proposed measures include mandatory reporting when AI agents act maliciously, as well as statutory powers requiring companies to grant the UK AI Security Institute (AISI) pre-release access to their latest systems. The move follows reports that Anthropic withheld access after pressure from the US administration under President Donald Trump, who dismissed safety concerns as a hoax. However, UK officials remain confident they have persuaded the US not to block future model sharing.

The announcement follows a similar statement from George Williamson, CEO of the Alan Turing Institute, who warned that the UK needs a stronger domestic AI presence and "national resillience" to keep pace with advances by the US and China.

"As AI becomes more embedded in critical services, in domains like defence, national security, critical infrastructure and healthcare we shouldn’t depend solely on systems we can’t inspect or control, or that could be restricted or even switched off," said Williamson.

£ - These articles require a subscription.

A version of this article is available without subscription in The Guardian.

Training Announcement: The BCS Foundation Certificate in AI examines the challenges and risks associated with AI projects, such as those related to privacy, transparency and potential biases in algorithms that could lead to unintended consequences. Explore the role of data, effective risk management strategies, compliance requirements, and ongoing governance of the AI lifecycle and become a certified AI Governance professional. Find out more.