White House Pushes Back Against Anthropic’s Mythos Expansion

The White House is resisting Anthropic’s plan to expand access to its powerful Mythos AI model, citing risks that the model could be used to identify and exploit software vulnerabilities and concerns over the company’s ability to contain sensitive information. Government officials point to recent security issues—unauthorized access to Mythos via private channels and an accidental exposure of internal source code for Anthropic’s Claude Code—as evidence that broader distribution could imperil critical infrastructure and that compute capacity and trust must be addressed before rollout.