The U.S. government restricted foreign access to Anthropic’s Claude Fable 5 and Mythos 5 models due to national security concerns. Anthropic responded by taking the models offline globally, representing a significant federal intervention in AI deployment.
The directive stems from a Trump administration executive order that created a security vetting framework for advanced AI systems. A White House adviser claimed the mandate followed the Anthropic CEO’s alleged refusal to fix a specific model vulnerability.
This enforcement highlights escalating friction between rapid AI innovation and government oversight. The decision creates new uncertainty for the international distribution and future deployment of powerful AI technologies.