On October 12, according to the UK's Daily Telegraph, US AI developers can now provide early access to AI models to the UK government for security testing without facing opposition from the Trump administration. Previously, the Trump administration had required companies like OpenAI and Anthropic not to share new models with the UK's AI Safety Institute prior to release, but this decision has now been withdrawn. This change indicates that the Trump administration has adopted a more lenient stance on international AI safety cooperation. As regulators consider how to manage this increasingly powerful technology, US developers will have more opportunities to collaborate with the UK.
