US Government Issues Export Controls on Anthropic's Fable 5 and Mythos 5 AI Models
The U.S. government has issued an export control directive impacting Anthropic's two most powerful AI models, Fable 5 and Mythos 5. Effective June 13, 2026, the directive suspended access for all foreign nationals, including Anthropic's own foreign-born employees, citing national security concerns related to a "jailbreaking" vulnerability. The decision has raised discussions in Europe and Canada regarding the control of advanced AI technologies.
The U.S. government has implemented an export control directive affecting Anthropic's Fable 5 and Mythos 5 AI models. The directive, effective June 13, 2026, mandated the suspension of all access to these models by any foreign national, both inside and outside the United States, including foreign national employees of Anthropic.
Anthropic had made Claude Fable 5 and Claude Mythos 5 generally available on June 9, 2026, following their development under a controlled access program named Project Glasswing. Fable 5 was described as a state-of-the-art model for general use, excelling in software engineering, scientific research, and autonomous work. Mythos 5, a more advanced version, was restricted to Glasswing partners and selected biology researchers.
Four days later, on June 12, 2026, Anthropic stated it received the export control directive at 5:21 pm ET. The order, issued by Commerce Secretary Howard Lutnick to CEO Dario Amodei, did not detail the specific security concern. Unable to filter users by nationality in real time, Anthropic reportedly disabled access for all customers to comply with the directive.
The U.S. government cited national security as the reason, specifically pointing to a method for "jailbreaking" Fable 5, which allows bypassing its safety guardrails. Anthropic, however, disputed the severity, suggesting the technique offered limited capability to review program code and identify errors, a function also present in rival models like OpenAI's GPT-5.5.
David Sacks, co-chair of the President's Council of Advisers on Science and Technology, stated on X that the administration had asked Amodei to either fix the vulnerability or withdraw the model, a request Sacks indicated Amodei refused. The Wall Street Journal also reported that Amazon CEO Andy Jassy had informed Treasury Secretary Scott Bessent and other officials that Amazon researchers had used Fable 5 prompts to obtain information potentially useful for cyberattacks. Amazon is a significant investor in Anthropic.
This dispute reportedly originated earlier in the year, stemming from Anthropic's stance that its technology should not be used for mass surveillance or fully autonomous weapons.
According to AI News, the swift implementation of these controls has highlighted global concerns over AI sovereignty.