Anthropic Halts Access to New AI Models After U.S. Government Order

Anthropic disabled access to its newest AI systems, Fable 5 and Mythos 5, after the Commerce Department ordered the company to block use by any foreign national under national-security authorities. The move—prompted by a reported technique to bypass certain safeguards—marks an unusual federal intervention in a publicly deployed frontier model. Anthropic disputed the scope and evidence behind the government’s concerns, calling the workaround narrow and likely present in rivals’ systems, and criticized the process as opaque. The suspension interrupts service for all customers, including enterprises and government partners, and follows months of tensions between Anthropic and Washington, including litigation over federal restrictions and reports that the NSA has been using Mythos for offensive cyber operations. The episode underscores how export controls and safety reviews are increasingly shaping commercial AI rollouts.

Read more