UK seeks broad powers to regulate overseas AI developers

The British government has proposed a clause in the Cyber Security and Resilience Bill that would give ministers broad powers over AI companies worldwide. The amendment would allow inspections, platform testing and requirements for developers to test their models, with sanctions for noncompliance. Introduced at the final substantive amendment stage, it faces limited opportunities for parliamentary scrutiny.

AI models and their developers could fall within the bill’s scope as critical suppliers if their technology is used by an energy company or data center, including through open-weight models. The legislation leaves the details of regulation and sanctions to subsequent rules, potentially giving Parliament little influence over implementation.

The proposal creates a tension between cybersecurity oversight and access to technology Britain relies on foreign companies to supply. OpenAI tools helped discover a vulnerability in PHP, a coding language estimated to be used in 70% of all websites. Research from the Center for Governance on AI found that 7% of model releases were delayed or withheld in the UK compared with the US, while the figure for the EU was 11%. Britain currently depends on US and Chinese developers, making possible restrictions on frontier-model access a significant consideration for regulatory decisions.