OpenAI / Wikimedia Commons (Public domain)
The Trump administration is treating frontier AI like a weapons export, and America's closest ally is getting locked out.
The US government is actively pressuring OpenAI and Anthropic to restrict access to their most advanced AI models from the United Kingdom, marking a dramatic escalation in how Washington treats cutting-edge artificial intelligence: less like a software product, more like enriched uranium.
The restrictions, which have intensified over recent months, represent a sharp break from the historically open AI collaboration between the US and UK. They also signal that the Trump administration views frontier AI capabilities as a national security asset that shouldn’t leave American shores, even when the destination is Washington’s most reliable intelligence-sharing partner.
What happened, and what got restricted #
In mid-June 2026, the US Commerce Department issued an export-control directive targeting Anthropic’s Fable 5 and Mythos 5 models. The stated justification centered on cybersecurity risks, specifically the potential for “jailbreaks” that could compromise built-in safety protocols. The directive was sweeping enough that Anthropic suspended foreign access to both models entirely, a global shutdown that affected all users outside the US, including the company’s own employees abroad.
OpenAI followed a parallel path with its GPT-5.6 Sol model, restricting access exclusively to customers approved by the Trump administration. Notably, OpenAI itself does not appear to believe this gatekeeping approach should become permanent policy, but compliance with the directive left little room for dissent.
By early September 2026, the situation escalated further. Anthropic declined to grant pre-release testing access to its Claude Mythos 5.1 model to the UK’s AI Security Institute, known as AISI. This was a first. AISI had previously received pre-release access to frontier models as part of established collaborative safety-testing arrangements between the two countries. Instead, Anthropic offered access only to vetted US organizations.
Macro, rates, and crypto—what moved markets and what matters next.
Daily. Free. Join 34,000+ readers across crypto, finance, and policy.
The diplomatic fallout #
The UK government did not take this quietly. British officials launched a lobbying campaign seeking an exemption from the June restrictions, arguing that the Five Eyes intelligence-sharing relationship and the UK’s own robust AI safety infrastructure warranted special treatment.
The response from Washington was blunt. UK officials were told there was “zero chance” of an exemption being approved, with no indication that viable negotiations were even on the table.
The underlying rationale from the US side involves concerns that advanced AI capabilities could be exploited by foreign military intelligence services in adversarial nations.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our