What happened
The White House has restored the UK AI Security Institute's early access to frontier AI models. The decision covers models from Anthropic, Google DeepMind and OpenAI.
What happened in September
In September, the Trump administration told the companies to stop sharing their models with the British agency. The move had limited the institute's ability to review next-generation models for safety.
What changed the decision
Diplomatic pressure from the UK's AI Minister, Kanishka Narayan, led to the reversal. The institute can now once again get early access to frontier models.
Why it matters
AI safety testing relies on independent bodies examining models before release. Cutting off access weakened that oversight chain, while the reversal signals that international cooperation is being restored.
