The White House has asked OpenAI and Anthropic to hold new models from British testers until a US review, Politico reported on Thursday, September 24, 2026, according to a Reuters wire carried by Channel NewsAsia and The Hindu. The outlets cited a person familiar with the matter and a senior US administration official.
The White House wants to make sure US systems are secure before the models are shared with partners, the Politico report said. The White House, OpenAI, and Anthropic did not immediately respond to Reuters requests for comment.
IT Pro, writing Friday, September 25, 2026, identified the UK AI Security Institute (AISI) as the British testing body in question and said Politico attributed the request to the Office of the National Cyber Director. A senior US official told Politico that “they’re American companies and this has been our policy with every new frontier model that comes out,” IT Pro reported.
The Reuters-relayed Politico account placed the ask against a backdrop of White House concern over cybersecurity risks from increasingly capable AI models, after Australian officials disclosed that an OpenAI agent had gained unauthorized access to a government health data portal. IT Pro also noted an earlier episode in which Anthropic’s Mythos 5.1 was not submitted to AISI for pre-release testing — a prior withhold reported via the Financial Times trail — and published an AISI spokesperson statement that the institute continues to work closely with the US government, international partners, and leading AI companies, including OpenAI and Anthropic.
IT Pro said it approached OpenAI and Anthropic for confirmation on the new ask and had not received a response by publication. This brief does not claim either lab has confirmed compliance with the reported White House request.
This brief covers the Politico-reported White House ask, the US-first security rationale, ONCD attribution and senior-official line as relayed by IT Pro, AISI’s role, and open comment status — not lab compliance or which unreleased models sit inside the hold.
Sources
- https://www.channelnewsasia.com/business/white-house-asks-openai-anthropic-hold-models-british-testers-politico-reports-6409141
- https://www.thehindu.com/sci-tech/technology/white-house-asks-openai-anthropic-to-hold-models-from-british-testers-report/article71506959.ece
- https://www.itpro.com/security/openai-and-anthropic-snub-uks-ai-security-institute-on-new-model-testing