Trump's Anthropic Export Ban Sparks AI Sovereignty Debate at G-7

The Trump administration's abrupt decision to restrict access to Anthropic's most advanced AI models has created diplomatic tension among Western allies and accelerated calls for nations to develop independent AI capabilities.

Last Friday, the U.S. government imposed export controls blocking all non-U.S. citizens—including Anthropic's own employees—from accessing the company's Fable and Mythos models. The restriction, which cited the models' effectiveness at uncovering software vulnerabilities, forced Anthropic to immediately shut down access for all users. The move came with minimal public explanation, alarming democratic allies who now question their reliance on AI technology that can be withdrawn without warning.

Why it matters

The export controls represent an unprecedented use of national security authority to shut down a commercial AI product, signaling that the U.S. may treat advanced AI models as strategic assets requiring the same restrictions as weapons technology. This approach forces allied nations to choose between accepting potential disruption of critical technology or investing heavily in domestic AI development—a gap that currently favors U.S. companies by a significant margin.

Industry leaders advocate for democratic AI alliance

At a G-7 working lunch Wednesday in Evian-les-Bains, France, Anthropic CEO Dario Amodei urged world leaders to coordinate the sharing of powerful AI technologies among democratic nations while excluding China. According to a person familiar with the discussion, Amodei emphasized that AI breakthroughs, though led by U.S. companies, resulted from researchers across all G-7 nations and that benefits should be distributed accordingly.

OpenAI CEO Sam Altman called for establishing an international forum to manage AI safety, while Google's AI head Demis Hassabis proposed an international standard-setting body to enact development rules. Hassabis argued that AI's economic, philosophical, and technical challenges shouldn't fall solely to companies to resolve.

Allied leaders express concern over U.S. unilateralism

French President Emmanuel Macron acknowledged the legitimacy of concerns about dangerous AI models but criticized the Trump administration's approach as "strictly nationalist," according to the Associated Press. Canadian Prime Minister Mark Carney warned earlier this week that the sanctions highlighted risks of "overreliance on certain models" and urged allies to diversify their AI capabilities.

Chris Lehane, OpenAI's head of global affairs, told reporters that foreign leaders recognize they face essentially two options for accessing cutting-edge AI: China or the United States. This reality, he said, underlies discussions about forming a safety club for democratic nations with the U.S. in a leadership role.

Trump told reporters after the lunch that negotiations with Anthropic over Fable and Mythos were "going fine" but did not address the export ban directly during the meeting.

These details were first reported by The Washington Post.