The US government’s Anthropic models ban was never about an AI jailbreak

Live updates

1. 1h ago

   US forces Anthropic to globally disable Fable 5 and Mythos 5 with no warning

   The US Commerce Department ordered Anthropic to immediately disable access to its Fable 5 and Mythos 5 models worldwide, affecting all users, not just foreign nationals. The directive came without explanation or prior notice, raising questions about the scope and justification of the security concerns. The move follows earlier restrictions targeting foreign users and ongoing disputes over AI capabilities. Experts question whether the broad suspension will hinder defensive tools while decentralized AI projects gain traction.

   What's confirmed:

   • The US government issued a directive on June 12, 2026, at 5:21 p.m. ET requiring Anthropic to disable Fable 5 and Mythos 5 for every customer worldwide.
   • The Commerce Department’s Bureau of Industry and Security delivered the order, citing unspecified national security concerns.
   • Anthropic received no advance warning or explanation for the suspension, which took effect immediately.
   • The ban was broader than previous restrictions, which had targeted foreign nationals only.
   • The suspension follows Amazon’s earlier warnings about capabilities like Defense Oriented Prompting (DOP) and ongoing disputes over AI vulnerabilities.

   Still unconfirmed:

   • The ban was triggered by a specific 'jailbreak' claim, though Anthropic disputes this as the sole reason.
   • There is speculation of Pentagon involvement or secret sabotage behind the decision, though no evidence has been confirmed.
   • The move may indirectly boost decentralized AI projects like Venice and Morpheus tokens.
   confidence 92%
2. 8h ago

   US government ban on Anthropic models tied to cybersecurity risks, not jailbreaks

   The Trump administration ordered Anthropic to suspend access to its Mythos 5 and Fable 5 models for all foreign nationals, citing concerns over bypassable safeguards and vulnerability identification. The move follows Amazon’s warnings and ongoing disputes over capabilities like Defense Oriented Prompting (DOP). Meetings between Anthropic and the White House suggest unresolved tensions, while cybersecurity experts argue the restrictions may weaken defensive tools.

   What's confirmed:

   • The US government banned access to Anthropic’s Mythos 5 and Fable 5 models for all foreign nationals, including those inside the US, citing concerns that safeguards can be bypassed and the models could identify software vulnerabilities.
   • Anthropic disabled access to these models for all customers following the US government’s export control directive.
   • The Trump administration’s decision appears reactionary and retaliatory, signaling the AI industry is not immune to US government interference.
   • Amazon’s warnings about Anthropic’s capabilities, including Defense Oriented Prompting (DOP), contributed to the White House’s decision to restrict access.
   • Cybersecurity experts warn that limiting access to models like Fable 5 may hinder defenders more than it restricts attackers.
   • Anthropic is still in dispute with the White House over the Fable 5 model, with no clear truce in place.
   • The US government believes the models’ capabilities could be exploited to identify critical software vulnerabilities, raising national security concerns.

   Still unconfirmed:

   • The Trump administration may be considering a broader crackdown on AI models with advanced cybersecurity capabilities, though no formal policy has been announced.
   • Anthropic’s internal research into Defense Oriented Prompting (DOP) was the primary trigger for the US government’s restrictions, though the company has not confirmed this as the sole reason.
   • The ban could accelerate a global AI arms race as other nations develop their own unrestricted models to counter US restrictions.
   confidence 95%