Pentagon deploys Anthropic’s Mythos AI cyber model despite plans to phase out company’s tools

The US Department of Defense is deploying Anthropic’s unreleased AI cybersecurity model, Mythos, to help identify and patch software vulnerabilities across government systems, even as the Pentagon moves ahead with plans to phase out the company’s products, according to a Reuters report.

Must read: Inside Claude Mythos, the AI is forcing a rethink of global cybersecurity

The deployment is part of Anthropic’s “Project Glasswing,” a controlled programme launched on April 7 that gives select organisations access to the Claude Mythos Preview model for defensive cybersecurity applications.

According to Anthropic, the model is designed to detect vulnerabilities across web browsers, infrastructure and enterprise software systems.

Speaking at a conference in Washington DC, Emil Michael, the Pentagon’s chief technology officer and under secretary for research and engineering, described the issue as a “national security moment,” Reuters reported.

Must read: Anthropic’s Claude Mythos triggers call for sovereign AI infrastructure in India: All details

Michael said the Defense Department continues to execute plans to remove Anthropic’s products from its systems over the coming months. However, he said Mythos possesses specialised capabilities that are particularly useful for finding and patching cyber vulnerabilities.

“The government has hardened up networks,” Michael said, adding that while such vulnerabilities have always existed, AI systems like Mythos now allow them to be discovered and fixed faster, though they could also be exploited faster by adversaries.

Must read: OpenAI takes on Anthropic’s Claude Mythos with new cybersecurity AI system ‘Daybreak’

Michael also suggested Anthropic’s lead may be temporary, saying competing models from rivals including OpenAI, xAI and Google are expected to develop similar capabilities soon.

The Pentagon had earlier designated Anthropic as a supply-chain risk after the two sides failed to reach an agreement over how the company’s AI models could be used within the agency. Anthropic later sued the Trump administration in March in an attempt to overturn the Pentagon’s blacklisting decision.