Anthropic suspended access to its flagship AI models Fable 5 and Mythos 5 following a US government directive citing national security concerns. The abrupt move marks a sharp escalation in government oversight of advanced AI development.
The suspension affects Anthropic's most capable large language models. Fable 5 and Mythos 5 represent the company's cutting-edge research in generative AI. Users attempting to access either model now encounter disabled service pages.
US officials issued the directive under national security authority rather than through traditional regulatory channels. The exact rationale remains classified, though government statements reference risks tied to dual-use technology. AI models capable of processing complex information fall under scrutiny from both the Commerce Department and intelligence agencies.
Anthropic operates in a tightening regulatory environment. The Biden administration previously issued executive orders requiring AI companies to report safety testing results. The current suspension suggests enforcement mechanisms now extend to direct service restrictions.
The timing coincides with broader US-China AI competition concerns. American policymakers view advanced AI capabilities as strategically sensitive. Export controls on semiconductor technology and AI chips already restrict technology transfer to China and Russia. Suspending public access to frontier models aligns with containment strategy.
Anthropic's response shows limited resistance to government demands. The company previously cooperated with National Institute of Standards and Technology testing frameworks. This suspension pattern follows similar precedents in defense contracting where national security orders override commercial operations.
The broader AI industry watches closely. Other labs including OpenAI and Google DeepMind operate under similar scrutiny. Microsoft's partnerships with US government agencies already embed security reviews into product development cycles.
Users report unexpected service interruptions without advance notice. Anthropic directed customers to contact support channels for clarification, though official statements remain sparse. The company posted minimal public commentary about the suspension duration or conditions for reinstatement.
Revenue implications remain unclear. Fable 5 and Mythos 5 serve enterprise and research clients. Discontinuing access disrupts ongoing projects and contract obligations.
This action reflects government willingness to enforce AI governance through direct operational intervention. Rather than regulatory frameworks passed through Congress, the administration exercises executive authority over private AI development. The precedent suggests more suspensions may follow if officials determine security risks outweigh commercial benefits.