Anthropic drew federal attention after revealing Mythos, an AI model that can write code so fluently it could threaten global cybersecurity. The company offered the model to a handful of security experts before public release. On June 9 it launched Fable, a revised version it claimed safer, sparking a backlash from regulators to the Department of Commerce and the National Security Council.

Amazon CEO Andy Jassy warned officials that Fable could destabilize national security, prompting the government to impose export controls. Minutes later Anthropic revoked access to both models. Critics argue the move mirrors a superficial safety tactic, while experts say banning the models may leave the U.S. more vulnerable to cyberattacks, as the tools aid defensive research for software developers worldwide.

The incident sparks debate over AI regulation. European leaders, alarmed by the ban, call for stronger domestic AI ecosystems, yet Chinese open‑source models offer cheaper, unregulated alternatives that threaten cyberdefense. Lawmakers now face pressure to define military AI limits and civilian safety standards, a challenge that may reshape how companies like Anthropic and competitors navigate export rules and public trust.