U.S. Government to Deploy Anthropic's Mythos AI for Cybersecurity Amid Risk Concerns
The U.S. government is planning to make a version of Anthropic's frontier artificial intelligence model, Mythos, available to major federal agencies, according to a report by Bloomberg News on Thursday. This move comes amid significant concerns that the advanced tool could sharply increase cybersecurity risks by identifying and potentially exploiting vulnerabilities in software systems.
Project Glasswing and Controlled Deployment
Announced on April 7, Mythos is being deployed as part of Anthropic's Project Glasswing, a controlled initiative that permits select organizations to use the unreleased Claude Mythos Preview model for defensive cybersecurity purposes. Under this project, the model is designed to help agencies bolster their cyber defenses by proactively finding weaknesses before malicious actors can exploit them.
Experts have highlighted that Mythos has already discovered thousands of major vulnerabilities in operating systems, web browsers, and other critical software. Its high-level coding capabilities grant it a potentially unprecedented ability to not only identify these cybersecurity flaws but also devise methods to exploit them, raising alarms about the dual-use nature of the technology.
Federal Preparations and Safeguards
In an email sent on Tuesday, Gregory Barbaccia, the federal chief information officer at the White House Office of Management and Budget, informed Cabinet department officials that the OMB was establishing protections to allow their agencies to begin using Mythos. The email, with the subject line Mythos Model Access, emphasized the need for robust guardrails before any potential release.
Barbaccia stated, We're working closely with model providers, other industry partners, and the intelligence community to ensure the appropriate guardrails and safeguards are in place before potentially releasing a modified version of the model to agencies. However, the email did not definitively confirm that agencies would receive access, nor did it provide a timeline for deployment or specific usage guidelines, as noted in the Bloomberg report.
Background and Political Context
The White House and Anthropic have not immediately responded to requests for comment from Reuters. Interestingly, Anthropic has been discussing Mythos with the Trump administration, as co-founder Jack Clark mentioned on Monday. This dialogue continues even after the Pentagon severed business ties with the U.S. AI lab following a contract dispute, highlighting the ongoing political and strategic considerations surrounding the technology.
As the government navigates these complexities, the deployment of Mythos represents a significant step in leveraging AI for national security, while also underscoring the critical need for stringent oversight to mitigate associated risks.
