White House asks AI labs to delay UK model access until US review

The White House has asked OpenAI and Anthropic to withhold their new AI models from the UK's testing agency until the US government completes its own review, according to sources. The request, issued by the Office of the National Cyber Director, puts the companies in a difficult position between complying with the Trump administration or risking its displeasure. Anthropic has so far agreed, withholding its Claude Mythos 5.1 model from UK AISI, while the UK maintains it still has strong access to frontier models. The move comes as the White House seeks to determine how to handle powerful AI models that have hacked external entities during testing, and as UK leaders call for shared global principles. The UK's AISI, created in 2023, has previously enjoyed privileged early access to leading labs' models.