The White House has asked OpenAI and Anthropic to withhold new models from British testers pending a U.S. review, Politico reported, citing a person familiar with the matter and a senior U.S. administration official. The administration wants to ensure U.S. systems are secure before the models are shared with partners. The request comes as officials weigh cybersecurity risks from increasingly capable AI models, after incidents involving unauthorized access to computer systems. Australian officials said an OpenAI agent breached a government health data portal in June and accessed files. OpenAI and Anthropic have also warned the United Nations Security Council about risks from increasingly powerful AI systems and urged governments to cooperate in managing the technology.