ai: Trump Admin releases Anthropic Mythos to be used by more than

Trump Administration Reverses Ban on Anthropic's Mythos 5 for Critical US Entities

The Trump administration has reversed course on its ban of Anthropic's powerful cybersecurity AI model, Mythos 5, greenlighting its immediate deployment for over 100 specific U.S. government agencies and private companies. This significant policy shift, coming just two weeks after the model was pulled from the market, also extends access to non-American employees within these authorized organizations, as well as Anthropic's own foreign staff, according to reports from Semafor and Reuters.

Background on the Ban and Model Withdrawal

The decision marks a softening of a stringent stance taken earlier this month. On June 12, Anthropic was compelled to withdraw both its advanced cybersecurity models, Mythos 5 and Fable 5, from public and private access. This ban followed allegations that the models' protective guardrails, designed to prevent misuse, were easily bypassed by security researchers. Fable 5, a slightly more accessible version of Mythos 5, had been released just days before the ban, touted for its enhanced safeguards, only to be withdrawn alongside its more powerful counterpart.

Specifics of the New Directive

Commerce Secretary Howard Lutnick formally communicated the administration's revised position to Anthropic's chief compute officer, Tom Brown, in a missive seen by Semafor. Lutnick stated, "I have determined that appropriate safeguards are in place to permit certain trusted partners to access the Claude Mythos 5 Model." This authorization is strictly limited to the specified list of U.S. entities deemed crucial for national security and critical infrastructure. The directive notably does not address the re-release of Fable 5, leaving its status in limbo.

Anthropic's Public Acknowledgment and Future Outlook

Anthropic publicly acknowledged the development on Friday via a post on X (formerly Twitter). The AI lab stated, "Since June 12, we’ve been working closely with the US government to restore access to Claude Mythos 5 and Fable 5. Today, the government notified us that Mythos 5, our strongest cybersecurity model, can be redeployed to a set of US organizations that operate and defend critical infrastructure." The company emphasized its commitment to rapidly restoring access for these organizations and confirmed ongoing discussions with the government to broaden Mythos 5 access and make Fable 5 available for general use once again.

Implications for AI Regulation and National Security

This move highlights the delicate balance governments face in regulating rapidly evolving AI technologies, particularly those with significant national security implications. The swift re-authorization of Mythos 5, despite earlier security concerns, underscores the perceived strategic importance of such advanced models for protecting critical U.S. infrastructure against cyber threats. It suggests a pragmatic approach where the benefits of cutting-edge AI for defense outweigh the risks when deployed with carefully negotiated safeguards and restricted access. The continued restriction on Fable 5, however, indicates an ongoing caution regarding wider public deployment of powerful AI, especially after documented bypasses of safety protocols. The situation also reflects the intense collaboration and negotiation between leading AI developers and federal agencies as they navigate the nascent landscape of AI governance.

FAQ

Q: What is Mythos 5, and why was it initially banned?

A: Mythos 5 is Anthropic's powerful cybersecurity-oriented AI model. It was banned two weeks prior to this announcement, along with Fable 5, after security researchers reportedly managed to bypass its protective guardrails, raising concerns about its safety and potential misuse.

Q: Which organizations are now authorized to use Mythos 5?

A: Over 100 specific U.S. government agencies and private companies have been authorized to use Mythos 5. This access extends to their non-American employees, as well as non-American employees working directly for Anthropic, specifically for organizations that operate and defend critical infrastructure.

Q: What is the status of Fable 5, Anthropic's other AI model?

A: While Mythos 5 has been re-authorized for specific entities, the current directive from the Trump administration does not address Fable 5. This means Fable 5, which was initially released as a more accessible version of Mythos 5, remains unavailable for general use as Anthropic continues negotiations with the government for its broader release.