The U.S. Commerce Department used national security export controls to bar Anthropic from distributing its newest AI models, Fable 5 and Mythos 5, to foreign nationals. Because the directive covers foreign employees inside the U.S., Anthropic disabled access to the models for all users, according to its statement. Anthropic said the government’s action followed a reported “bypass” technique that could unlock cybersecurity capabilities, which the company described as narrow. It also argued the export-control process should be transparent and grounded in technical facts. For universities and researchers, the restrictions add a new layer to AI governance—forcing schools to reassess vendor contracts, geographic access requirements, and how course or lab use of “frontier” models is documented and audited.