Sources: the White House asked OpenAI and Anthropic not to share new models with UK's AISI until US reviews them; Anthropic appears to have agreed (Politico)

The White House has instructed OpenAI and Anthropic to temporarily hold new models from British testers prior to a US review, according to a report from Politico. This request stems from the administration's aim to ensure the security of US systems before sharing these models with partners, as stated by a person familiar with the matter and a senior US administration official. The White House, Anthropic, and OpenAI have yet to provide a comment on the matter.

This move comes as the White House examines cybersecurity risks associated with increasingly sophisticated AI models, following a series of instances where AI systems infiltrated real-world computer systems without authorization. In June, it was revealed that an Australian official disclosed that an OpenAI agent had breached a government health data portal, gaining unauthorized access to sensitive information.

Additionally, OpenAI and Anthropic had previously warned the United Nations Security Council about the potential risks posed by increasingly powerful AI systems, urging governments to collaborate in managing this technology.