What Is Anthropic’s Mythos AI and Why Was It Blocked?
Anthropic’s AI tool Mythos, designed to detect software vulnerabilities, was restricted to 200 partners due to risks of misuse by attackers. The U.S. Trump administration ordered Anthropic on June 12 to block access to its advanced models, including Mythos 5 and Fable 5, from all foreign nationals after discovering potential 'jailbreak' vulnerabilities in Fable 5.
Anthropic PBC developed Mythos, an AI system capable of identifying security flaws in software and computer systems. The company limited access to 200 partner organizations, citing concerns that malicious actors could exploit its capabilities to steal data or disrupt critical infrastructure. On June 12, the Trump administration issued an order requiring Anthropic to prevent all foreign nationals—both inside and outside the U.S.—from accessing its most advanced AI models. The directive followed the discovery that Fable 5, a version of Mythos, could be bypassed through a process called 'jailbreaking,' allowing unauthorized use of its cybersecurity functions. In response, Anthropic immediately shut off access to Mythos 5 and Fable 5 for all customers. The company stated that the U.S. government’s decision was based on the risks posed by the potential circumvention of safeguards in Fable 5. The restrictions highlight growing concerns over AI security and the need for stringent controls to prevent misuse. Anthropic’s move reflects broader regulatory pressures on AI developers to ensure their technologies align with national security interests. The incident underscores the challenges of balancing innovation with the risks associated with powerful AI tools. As governments enforce stricter oversight, companies like Anthropic face increasing scrutiny over how they deploy and secure advanced technologies.
This content was automatically generated and/or translated by AI. It may contain inaccuracies. Please refer to the original sources for verification.