Trump Orders AI Reviews After Anthropic Model Penetrates Classified Systems

President Donald Trump signed a June 2 executive order creating a voluntary framework for the federal government to review covered frontier AI models for up to 30 days before public release. The order aims to prevent cyberattacks and military misuse of advanced technology. Following the directive, major developers including OpenAI, Google DeepMind, Microsoft, xAI, and Anthropic agreed to provide early access for national security evaluations.

The urgency of the framework follows a testing exercise called Project Glasswing, during which Anthropic's Mythos AI model identified vulnerabilities in highly sensitive U.S. government computer systems within hours. Senator Mark Warner cited National Security Agency head General Joshua Rudd, noting the tool penetrated almost all classified systems. In response, the administration ordered Anthropic to suspend foreign national access to its Fable 5 and Mythos 5 models. Anthropic complied by disabling the models for all customers, though the company argued the government's steps were unwarranted.

Currently, Meta remains the only major U.S. AI developer that has not yet signed the voluntary review agreement. The company is negotiating terms and stated it hopes to sign the agreement soon. Meanwhile, over 100 cybersecurity experts from firms such as Nvidia and Adobe have urged the administration to lift the restrictions on Anthropic, claiming that limiting such tools could disadvantage the United States against foreign adversaries.