In a briefing published on June 14, The Economist criticized the Trump administration’s export controls on Anthropic as “capricious and chaotic,” leaving America’s closest allies caught off guard. The incident began at 5:21 p.m. on June 12, when Secretary of Commerce Howard Lutnick sent a letter to Anthropic CEO Dario Amodei, citing “national security” and ordering the company to immediately bar all foreign nationals (including foreign employees inside the U.S.) from using its two flagship models, Fable 5 and Mythos 5. Anthropic was given only about 90 minutes to comply and was forced to shut down access to those models for all users. The stated trigger was another company’s report to the government of a jailbreak vulnerability in Mythos — but Anthropic countered that the vulnerability was “relatively simple” and that other publicly available models had the same issue.
The diplomatic fallout spread quickly. At the G7 summit, British Prime Minister Keir Starmer directly asked President Trump for an exemption but was refused; the UK’s AI minister, Kaniska Narayan, subsequently stated the incident proved Britain needed stronger “sovereign AI capabilities.” Canadian Prime Minister Justin Trudeau described the episode as a warning against over-reliance on a single AI model, and French politicians called for AI to be elevated to the level of a sovereignty issue on par with nuclear energy. G7 diplomatic sources revealed that allies have presented Secretary Lutnick with a framework for a “trusted partner” arrangement, but no agreement has been reached so far. Earlier, after negotiations between the Trump administration and Anthropic over the Pentagon’s use of Claude broke down, the company was labeled a “supply chain national security risk.”