ON June 12, 2026, the Trump administration ordered Anthropic to restrict access to its AI models, Fable 5 and Mythos 5, exclusively to U.S. citizens after Mythos allegedly breached NSA and Cyber Command systems within hours. Senate testimony revealed that Mythos successfully infiltrated nearly all classified systems, leading to a U.S.-government ordered shutdown of the models without prior notice, impacting both domestic users and international allies, including Five Eyes partners.
Additionally, while Anthropic claims Mythos has identified numerous vulnerabilities in cybersecurity, concerns arise over the handling of AI capabilities and geopolitical implications, as Europe struggles with regulatory frameworks against reliance on external technologies for advanced AI.