Federal agencies could soon receive access to a cutting-edge artificial intelligence system designed to strengthen cybersecurity defenses, according to a Bloomberg News report published Thursday.
The Biden administration is exploring plans to distribute a specialized version of Anthropic’s advanced AI model, known as Mythos, to key government departments. However, officials are proceeding cautiously due to concerns that the powerful technology could potentially create new security risks.
Mythos was unveiled on April 7 as part of Anthropic’s controlled program called “Project Glasswing.” Under this limited rollout, selected organizations can access the unreleased Claude Mythos Preview model specifically for protecting against cyber threats.
The AI system has demonstrated remarkable capabilities, discovering “thousands” of significant security flaws in operating systems, internet browsers, and various software applications. Technology experts note that Mythos possesses advanced programming skills that give it extraordinary abilities to detect cybersecurity weaknesses and understand how they might be exploited.
Gregory Barbaccia, who serves as federal chief information officer at the White House Office of Management and Budget, sent an email to Cabinet department leaders on Tuesday outlining plans to establish security measures for agency use of Mythos, Bloomberg reported.
“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,” Barbaccia wrote in the message titled “Mythos Model Access,” according to the report.
Bloomberg noted that Barbaccia’s communication did not guarantee that agencies would ultimately receive Mythos access, nor did it specify when such access might be granted or how the technology would be implemented.
Neither the White House nor Anthropic provided immediate responses to requests for comment from Reuters.
Anthropic co-founder Jack Clark revealed on Monday that the company has been in discussions about Mythos with the Trump administration. These talks continue despite the Pentagon recently ending its business relationship with the AI company following a contractual disagreement.
