U.S. Government Halts Anthropic’s AI Models Citing Security Risks

The San Francisco-based company Anthropic announced that the U.S. federal government has ordered the suspension of its Fable 5 and Mythos 5 artificial intelligence models. This decision stems from concerns over national security.

Anthropic disclosed the government’s directive in a social media update, stating it must immediately disable these AI systems for all customers to ensure compliance. The suspension affects foreign nationals working with the company as well.

In April, Anthropic introduced the Claude Mythos model but did not make it publicly available. They feared it could become a tool for hackers to infiltrate computer networks. The potential security risks posed by this AI model have alarmed experts in Washington and Silicon Valley.

Anthropic had shared the Claude AI system with approximately 40 organizations, which manage essential computer infrastructure. This measure aimed to help them address security vulnerabilities before hackers could exploit them.

The Commerce Department ordered the access restrictions on the new AI models, according to an anonymous source familiar with the situation. The duration of these limitations remains uncertain.