Tech and cybersecurity leaders call on U.S. officials to lift export controls on Anthropic's AI models, arguing the restrictions limit defenders' access to security tools without genuine risk justification.

What happened: A group of technology executives and security professionals, including leaders from Corridor, Socket, Ivanti, and academics from Johns Hopkins and University of Pennsylvania, sent an open letter to U.S. officials requesting removal of export control directives on Anthropic's Fable and Mythos large language models.

Why it matters: The signatories argue that Anthropic's models are good at finding software flaws and weaponizing exploits, but no more so than many other foundation and open-source models already in regular use for security audits. They contend that restricting access to these tools removes capabilities from American defenders while Chinese open-weight models are only months behind the best American models, potentially giving adversaries an advantage.

What to watch: The letter calls for any future AI regulations to be grounded in scientific evaluations developed with industry and academia input, created through democratic rule-making, enforced transparently with time to remediate, and used only to the minimal extent necessary to protect the American public.