‘It’s not a jailbreak’ — Research leading to U.S. export restrictions on top Anthropic models was for defense, cybersecurity CEO says
The U.S. Commerce Department has restricted access to Anthropic's new Fable 5 and Mythos 5 AI models, citing national security concerns related to research that probed potential safeguards bypasses. Anthropic disputes the characterization of this research as a "jailbreak," arguing it was focused on defense-oriented prompting and that the broad export controls are disproportionate and lack transparency. This action, which affects non-U.S. persons including Anthropic's own employees, has raised concerns about the balkanization of technology and the unpredictable regulatory environment for AI developers. AI
IMPACT Government export controls on AI models could stifle innovation and lead to further technological balkanization.