Anthropic temporarily suspended access to its AI models Mythos 5 and Fable 5 after the U.S. government imposed export controls due to a reported “jailbreak” vulnerability, according to a company statement. The restrictions, which barred foreign nationals from using the models, forced the AI lab to shut them down entirely, according to a source familiar with the matter.
What triggered the U.S. government’s action against Anthropic?
The U.S. Department of Commerce notified Anthropic on Friday that it had identified a method to “jailbreak” Fable 5, a system designed to prevent misuse of the AI model. The agency classified the models as “military-grade” due to their ability to detect cybersecurity flaws, leading to export restrictions under the Export Administration Regulations (EAR), according to a government source. Anthropic confirmed the suspension in a statement, noting the restrictions applied to “any foreign national, whether inside or outside the U.S.”
How did Anthropic respond to the government’s decision?
Anthropic disputed the government’s approach, arguing that the jailbreak did not justify “sweeping action” that could stifle AI innovation. “This action does not adhere to principles of transparency, fairness, or technical grounding,” the company said in a statement. The lab has since sent staff to Washington to negotiate with the Trump administration, per a Wall Street Journal report, though no official response has been released from the Department of Commerce.
What are the broader implications for AI regulation?
This marks the second conflict between Anthropic and U.S. authorities. Earlier this year, the Department of Defense labeled the company a “supply chain risk” after it barred military use of its Claude AI for autonomous weapons or mass surveillance. The Pentagon sought broader permissions, while Anthropic emphasized ethical safeguards. The current dispute highlights tensions between AI developers and regulators over balancing innovation with national security, according to cybersecurity experts.
What happens next for Anthropic and its models?
Anthropic has not outlined a timeline for restoring access to Mythos 5 and Fable 5. The company has previously advocated for “statutory processes” that allow governments to block “unsafe deployments” while ensuring transparency. The outcome could set a precedent for how U.S. agencies regulate advanced AI systems, particularly those with dual-use capabilities in cybersecurity and defense. A Department of Commerce spokesperson declined to comment, citing ongoing discussions.