UK AI Minister Says Working with US to Keep Pre-Release Access to Frontier Models

On October 12, 2026, UK AI minister Kanishka Narayan said in a speech in London that the UK has been working closely with the US government to ensure the UK AI Safety Institute continues to receive pre-release testing access to the most powerful AI models. “I want the UK to be one of the hardest countries in the world to be attacked by AI,” he said.

In September 2026, media reports said the Trump administration had asked OpenAI and Anthropic to hold back AI models from the UK until US authorities had first deemed them safe. The US has been tightening global access to Silicon Valley AI models, and earlier in 2026 banned Anthropic from releasing its latest models to foreign citizens.

Separately, Anthropic has not provided the UK AI Safety Institute with the latest version of its powerful Mythos model, without explanation, according to earlier reports.