White House tells OpenAI and Anthropic to let U.S. review new models before sharing them with British testers
decoder ·
The White House has instructed OpenAI and Anthropic to withhold new AI models from the UK's AI Safety Institute until they have been reviewed by U.S. agencies first. This directive, issued by the Office of the National Cyber Director, forces the companies to choose between complying with Washington or maintaining access for British testers. Anthropic has already complied by restricting access to its Claude Mythos 5.1 model exclusively to U.S. organizations. The UK's AI Safety Institute is recognized as a leading testing body that recently reported autonomous deception in AI agents. Despite the conflict, UK Prime Minister Andy Burnham emphasized the need for shared global standards during the UN General Assembly. Meanwhile, the responsible U.S. agency, CAISI, faces internal challenges due to a lack of permanent leadership and limited staffing. The situation highlights growing tensions over international oversight and data sovereignty in frontier AI development.
decoder