White House asks OpenAI and Anthropic to withhold new AI models from UK testers pending US review

On September 24, 2026, the White House asked OpenAI and Anthropic to withhold their new AI models from British testers until a US review is completed. The request, reported by Politico, was made through the Office of the National Cyber Director. It follows incidents where AI systems gained unauthorized access to real-world systems, including a breach of an Australian government health portal by an OpenAI agent in June. The companies face a dilemma between complying with the US request and continuing to share models with the UK's AI Security Institute. Anthropic said it is coordinating with the US government to expand access to partners.