Anthropic disabled its Fable 5 and Mythos 5 AI models on June 12 after the US government issued an export control directive citing national security concerns, forcing immediate compliance for all users. The company publicly disagrees with the underlying reasoning, according to BeInCrypto. The powerful models had only been available for a short time before the takedown.
The episode highlights a reactive approach by both the administration and industry to rapidly advancing AI technology. It also raises questions about why Amazon—a major investor in Anthropic—would strike such a disruptive blow against the company. Amazon called administration officials Thursday night to share a report on how they jailbroke and accessed sensitive parts of the Mythos model, sources told Axios.
Anthropic had previously notified the government multiple times about the planned June 9 release of Fable, a general-use version of Mythos, and the government did not initially object. However, calls from Amazon and at least five other companies to senior administration officials Thursday evening and Friday morning led to the model being shut down by Friday night. The order forced Anthropic to suspend access for all users, including foreign nationals.
The move underscores growing tensions between AI developers and national security regulators. The administration's rapid response suggests a heightened sensitivity to potential risks from advanced AI models. Anthropic's public disagreement with the reasoning indicates potential pushback or legal challenges ahead.
Critics argue that such ad-hoc interventions could stifle innovation and create uncertainty for AI companies. The role of a major investor like Amazon in triggering the shutdown also raises concerns about competitive dynamics in the industry.