Choose Rich Live
Tech

US restores UK access to frontier AI models for safety testing

The United States has reversed course to restore the UK AI Security Institute's access to frontier AI models from developers including OpenAI and Anthropic, following an announcement from a UK minister and media reports.

A Quartz social graphic showing a hand holding a phone displaying the OpenAI logo in front of the Anthropic logo, with text reading 'U.K. Wins Back Secret Access to AI Models.'
Image: @qz

A UK government minister stated that Britain has been granted access to safety-test the latest artificial intelligence models, following a policy reversal by the United States that was first reported by The Telegraph and relayed by Reuters and other outlets.[2][3][9][10]

The change allows American AI developers to provide pre-release models to the UK AI Security Institute for safety testing without running afoul of the Trump administration, according to The Telegraph. The White House had previously directed firms, including OpenAI and Anthropic, to halt sharing models with the British institute, resulting in a 17-day freeze.[1][4][5][7][8]

Restoring early access is expected to strengthen UK-US cooperation, bolster international AI safety oversight, and minimize the risk of fragmented global releases, though reporting noted the dispute highlights the sensitivity of cross-border tech alliances under shifting political dynamics.[4][5][6]

Key facts

  • A UK minister stated that the UK has been given access to safety-test the latest AI models.
  • The United States reversed course to restore the UK AI Security Institute's access to frontier AI models for safety testing.
  • The Trump administration had previously ordered OpenAI and Anthropic to stop sharing models with the UK institute.
  • The US freeze on model access for the UK AI Security Institute lasted for 17 days.
  • Under the policy reversal, American AI developers can give the UK government early access for testing without violating Trump administration rules, per The Telegraph.

Sources · 6 sources

  1. QU

    Quartz@qzPost on X ·

    U.K. Wins Back Secret Access to AI Models Britain's AI Security Institute has regained access to pre-release models from OpenAI and Anthropic after the Trump administration had blocked it — a reversal partly credited https://t.co/TebmqikXnW #AIGovernance #OpenAI #UKAISecurity https://t.co/Au2Tn1Jt4H

    Open source
  2. RE

    Reuters@ReutersPost on X ·

    US reverses course on Britain's testing of AI models, Telegraph says https://t.co/l4gpAWCvLI https://t.co/l4gpAWCvLI

    Open source
  3. RE

    Reuters@ReutersPost on X ·

    UK given access to safety-test latest AI models, minister says https://t.co/1fRBw3UxyA https://t.co/1fRBw3UxyA

    Open source
  4. CB

    Crypto BriefingArticle ·

    UK AI Security Institute regains access to frontier AI models after 17-day US freeze The restored access may strengthen UK-US AI collaboration, easing fears of national divides and influencing global AI regulatory strategies. The post UK AI Security Institute regains access to frontier AI models after 17-day US freeze appeared first on Crypto Briefing .

    Open source
  5. CB

    Crypto BriefingArticle ·

    UK regains pre-release access to US AI models after White House reversal The reversal strengthens UK-US AI collaboration, enhancing global AI safety oversight and reducing risks of fragmented international model releases. The post UK regains pre-release access to US AI models after White House reversal appeared first on Crypto Briefing .

    Open source
  6. CB

    Crypto BriefingArticle ·

    US reverses course, restores UK access to frontier AI models for testing The restored UK access to AI models highlights the fragility of international tech collaborations amid shifting political landscapes. The post US reverses course, restores UK access to frontier AI models for testing appeared first on Crypto Briefing .

    Open source
  7. TH

    The Hindu@the_hinduPost on X ·

    American AI developers can ‌give the British government early ​access to AI models ⁠for safety testing without running afoul of US President Donald Trump’s administration, the Telegraph ‌newspaper said on Sunday. https://t.co/KAxIDYAX9U

    Open source
  8. TE

    Techmeme@TechmemePost on X ·

    Sources: the US restores the UK AISI's access to the latest US models for safety testing, after ordering OpenAI and Anthropic to stop sharing models last month (@tony_diver / Telegraph) (Visit Techmeme dot com for the link and full context!)

    Open source
  9. RE

    Reuters@ReutersPost on X ·

    US reverses course on Britain's testing of AI models, Telegraph says https://t.co/luAiDJtgs4 https://t.co/luAiDJtgs4

    Open source
  10. TS

    The Straits Times@straits_timesPost on X ·

    US reverses course on Britain’s testing of AI models: Report https://t.co/6mNSCc6he4

    Open source