# Lutnick’s Anthropic crackdown claims new power over AI models > Source: > Published: 2026-06-19 15:53:58+00:00 **Getting your** [Trinity Audio](//trinityaudio.ai)player ready...**By Maggie Eastland, Bloomberg** The Trump administration’s push to rein in [Anthropic PBC,](https://www.mercurynews.com/2026/06/02/anthropic-offers-mythos-model-access-to-150-additional-groups/) outlined in a recent Commerce Department order, relies on an unprecedented use of export control laws and raises legal questions about whether the US can dictate who can access artificial intelligence systems. In ordering Anthropic to obtain US approval for foreign nationals to use its Fable 5 and Mythos 5 models, Commerce Secretary Howard Lutnick expanded the boundaries of laws governing transfers of sensitive technology to target the mere usage of cutting-edge AI models. That’s fueling concerns among developers and their customers about the government’s willingness to intervene in their operations on national security grounds. Before last week’s directive, the technology industry had operated under the assumption, based on past official advisory opinions, that export controls did not apply to usage of cloud-based programs, even if US restrictions applied to transfers of the software code itself. In the wake of Lutnick’s letter, “you can’t say that any more,” said Kate Koren, a deputy director at the Center for Strategic and International Studies. “Now that the Commerce Department has done it, no company can rule out that they’re going to do it again,” added Koren, who previously worked at the Commerce Department. “Until or unless this is challenged, any customer has to assume this could happen for any model at any time.” The order marked the government’s most significant intervention in the AI industry, prompting Anthropic to immediately disable access to both models. It represented a major shift from plans to vet AI models based on companies’ voluntary participation spelled out by an executive order from President Donald Trump a little more than two weeks ago. Lutnick’s letter to the company, seen by Bloomberg News, has raised questions over whether simply using an AI model now counts legally as a technology transfer. That position, which is required to make the order’s restrictions enforceable, runs counter to past practice regarding access to software capabilities, according to export control analysts and trade lawyers. The Bureau of Industry and Security, which oversees US export control programs, has long enforced rules requiring companies to gain US approval before sharing certain sensitive technologies, including with foreign employees, a process it calls “deemed exports,” but its application to AI usage is uncharted territory. “For the provisions cited in the letter, it matters a quite a lot what Commerce means by ‘model,’” said Chris Chamberlain, a former Commerce Department adviser and partner in the law firm Morrison Foerster’s national security group. “The baseline interpretation is that providing cloud access to software is not an export or deemed export of that software.” In his letter, addressed to Anthropic Chief Executive Officer Dario Amodei, Lutnick cited two codified export control authorities that he said give the US the ability to impose license requirements. Those powers, however, remain relatively untested and may only be applicable for specific countries, rather than on a worldwide basis. The first statute gives the US the ability to identify emerging technologies and enact interim restrictions over them. The second allows the government to quickly impose license requirements to defuse a risk of controlled technologies reaching a military intelligence end user in an adversarial nation like China or Russia. While the administration has yet to comment on the reasons for its decision, the company says it believes the government issued the order after discovering that it’s possible to “jailbreak,” or bypass the guardrails, of Fable 5, a recently released version of Mythos that Anthropic had blocked from carrying out cybersecurity tasks. “The US has solved a problem,” Koren said, referring to the jailbreak risk. “But they’ve opened a can of worms by doing it this way.” Since last week, the company’s technical staff and administration officials have been holding discussions in a bid to resolve the government’s national security concerns, though so far, there’s been little public sign of progress. When asked Wednesday at the Group of Seven summit in France about the status of the talks, Trump said they were “going fine” but gave no additional details. At the G-7, the US crackdown on Anthropic took center stage, with French President Emmanuel Macron leading discussions on ways to deploy advanced AI models through so-called trusted partners. Broadly, the US move has “renewed worldwide calls for technological sovereignty and independence from the American ecosystem,” said Michelle Nie, a fellow at the Center for a New American Security. For now, the Commerce Department’s measures require Anthropic to obtain a license to offer those models to anyone who’s a foreign person. The company doesn’t currently require citizenship verifications for its users, and it counts foreign nationals among its own employees. The move means the rest of the AI industry will be forced to take heed of the authority asserted by Lutnick, and BIS can invoke other powers if the specific provisions used against Anthropic are contested. That raises the risk of service interruptions for companies that run afoul of US government requests. Aidan Gomez, the chief executive officer of Canadian AI developer Cohere Inc., said on Thursday that the episode highlights the need for companies to lean on a range of providers. “It’s exactly what we’ve been saying for so long,” Gomez said in an interview at the VivaTech conference in France. “If you rely on centralized large players, you are at risk of losing access, and so you need a diversified supply chain of AI.” –With assistance from Benoit Berthelot, Josh Wingrove, Mackenzie Hawkins and Ania Nussbaum. More stories like this are available on [bloomberg.com](https://www.bloomberg.com) ©2026 Bloomberg L.P.