In a significant policy shift, the Trump administration has authorized Anthropic to distribute its advanced AI model, Mythos 5, to over 100 designated U.S. government agencies and private companies. This decision comes two weeks after a ban led Anthropic to withdraw both Mythos 5 and its variant, Fable 5, from the market due to security concerns.
The new directive permits access to Mythos 5 for specific organizations, including their non-American employees, as well as Anthropic’s own international staff. This marks a departure from the initial ban, which restricted non-U.S. personnel from engaging with these models. Commerce Secretary Howard Lutnick communicated this policy change to Anthropic’s Chief Compute Officer, Tom Brown, stating that appropriate safeguards are now in place to allow trusted partners to utilize the Claude Mythos 5 Model.
Notably, the administration’s directive does not address the status of Fable 5. This model, a version of Mythos 5 with enhanced protections, was released shortly before the ban but was also withdrawn after researchers reportedly bypassed its security measures. Anthropic has yet to comment on this aspect.
Anthropic acknowledged the development in a public statement, indicating that they have been collaborating with the U.S. government since June 12 to restore access to both models. The company is now working swiftly to reinstate Mythos 5 for the approved organizations and continues to engage with the government to broaden access and reintroduce Fable 5 for general use.
This policy reversal underscores the administration’s recognition of the critical role advanced AI models like Mythos 5 play in national cybersecurity. By allowing controlled access, the government aims to balance innovation with security, ensuring that key infrastructure entities can leverage cutting-edge technology while mitigating potential risks. Observers will be keen to see how this decision influences the broader AI policy landscape and the future deployment of similar technologies.
