Overview
- U.S. authorities issued an export-control directive that suspends access by any foreign national to Anthropic’s Fable 5 and Mythos 5 after researchers reported techniques that could bypass the models’ built-in safety controls.
- More than 50 cybersecurity leaders and tech executives have published an open letter asking the government to reverse the ban because they say restricting access will reduce defenders’ ability to find and fix software and security flaws.
- Anthropic rejects the ban’s justification, saying Fable was released with limits that block sensitive queries and that the reported jailbreak does not warrant cutting off foreign researchers and employees.
- Company executives are scheduled to meet with officials at the Department of Commerce to negotiate the directive as the dispute crystallizes into a test case over U.S. control of frontier AI tools.
- Observers warn the move could set a precedent for tighter U.S. controls on leading AI systems and may push allies to boost local AI investments and reshape how defenders access advanced tools for hunting vulnerabilities.