Anthropic Disables Fable 5 and Mythos 5 AI Models Following US Export Control Order

Anthropic has announced it will disable access to its latest top AI models, Fable 5 and Mythos 5, which were released earlier this month. The decision follows a US government export control order that cites national security concerns.

The company stated that the order, received via letter on Friday evening, bars any foreign entities or individuals, including foreign national Anthropic employees, from using the models. Anthropic indicated that the "net effect" of this order is the abrupt disabling of these models for all customers globally to ensure full compliance.

Anthropic noted that the government's letter did not provide specific details regarding its national security concerns. However, the company believes the concern may relate to a potential method to "jailbreak" Fable 5. Anthropic has disputed the severity of this issue, describing the technique as narrow, not universal, and involving known vulnerabilities identifiable by other publicly available models.

This development marks a further escalation in Anthropic's ongoing interactions with the Trump administration concerning AI safety, national security, and government oversight of frontier AI models. Previously, in February, the Pentagon designated Anthropic as a supply chain risk after the startup sought to limit its AI models for certain defense applications. Anthropic subsequently sued the Defense Department over this designation, with two related lawsuits still pending.

Anthropic confirmed its compliance with the order but stated its disagreement with the government's findings. An exact timeline for when access would be disabled was not provided by an Anthropic spokesperson. The company also clarified that access to its other AI models will not be affected by this order.

(Source: Business Insider)