The Trump administration this week abruptly ordered Anthropic to cut off access to its newest AI models for all foreign nationals — including users inside the United States and the company's own employees. According to The Verge, the order forced Anthropic to block access to its Fable 5 and Mythos 5 models, and the company has spent much of the week fighting to bring them back online under export rules that, the outlet notes, few people seem to understand.
The fallout has reached the highest levels of diplomacy. According to the Financial Times, Anthropic CEO Dario Amodei — backed by rival Sam Altman, along with Demis Hassabis — used the G7 summit to call for US-led international cooperation on AI rules, while French President Macron and India's Prime Minister Modi raised concerns over the US block on Mythos. Amodei attended a working lunch with G7 leaders and tech CEOs at the summit in Evian-les-Bains, France, on June 17.
Tom's Hardware reports that European and Canadian leaders are alarmed, with some warning that if they lose direct access to frontier American models, they may need to build their own national alternatives.
The dispute also has an internal dimension. The New York Times reports that Anthropic employees accuse the Trump administration of targeting them directly.
At the heart of the standoff is a technical demand. Trump administration officials told WIRED that for Anthropic to rerelease Fable 5, it must guarantee the model's guardrails cannot be circumvented — in other words, block all jailbreaks. Security experts told WIRED that simply can't be done.
Why it matters: a single government order has shown how quickly access to powerful AI can be switched off across borders, pushing US allies to rethink their dependence on American technology.