The US government issued an export control directive on June 12, 2026, ordering AI company Anthropic to block all foreign nationals, including foreign employees inside the US, from accessing its latest AI models Fable 5 and Mythos 5 [1, 2, 3, 4, 5, 6, 7, 8]. Fable 5 and Mythos 5, released just days earlier on June 9, are advanced AI systems with built-in cybersecurity and biotech safeguards [3, 6, 9, 8].
Anthropic responded the same day by disabling global access to the two models to comply with the order. Access to the company’s other AI models remains unaffected [2, 3, 4, 5, 6, 7, 8]. The company said, "The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance" [4].
The US government cited national security concerns as the motivation but withheld specific details about the risks involved [2, 3, 4, 5, 6, 8]. Officials are worried that a method exists to "jailbreak" Fable 5 safeguards, potentially exposing cybersecurity vulnerabilities that could be exploited [2, 3, 4, 5, 6, 7, 8]. However, Anthropic conducted internal tests on the alleged jailbreak and found only minor, previously known vulnerabilities. The company disputes that this limited jailbreaking warrants blocking access to these widely used models, calling the order a misunderstanding [2, 4, 5, 6, 7, 8]. "We do not believe that finding a narrow potential jailbreak should trigger recalling a model deployed to hundreds of millions of people," Anthropic said [6].
The ban affects allied researchers as well, including foreign employees in Canada, the UK, and other countries. Experts outside the company have criticized the measure as poorly thought-out. Jimmy Goodrich of the UC Institute for Global Conflict and Cooperation said, "It even bans Canadians and Brits employed at Anthropic from doing research and development" [10].
Amazon CEO Andy Jassy and other tech leaders reportedly raised security concerns about Anthropic’s AI models to the previous Trump administration weeks before the ban [10]. Commerce Secretary Howard Lutnick sent the export control letter to Anthropic CEO Dario Amodei [1, 11, 7]. Amodei has previously supported government authority to restrict AI deployments considered risky, though Anthropic is now seeking clearer, transparent processes to manage such restrictions [12, 7, 8].
The government action comes amid ongoing tension between Anthropic and federal agencies over the use of AI for military and surveillance purposes [6]. Experts warn that similar AI capabilities are expected to emerge in other companies’ products soon, which could limit the long-term impact of controlling access to Anthropic’s latest models [9].
Anthropic is working to restore access and has requested a clearer legal and technical framework from regulators to avoid abrupt disruptions in the future [2, 12, 7, 8].
