Mythos AI: China's Request Alarms US Officials

A Chinese representative recently pushed Anthropic officials for access to their AI model called Mythos. This happened at a closed-door meeting in Singapore. Anthropic refused the request. Mythos is an AI model that can find software bugs and create working exploits without human guidance. Anthropic will not release it publicly. Access is limited to the U.S. government and about 40 partner organizations. U.S. officials were alarmed when they learned about the Chinese request. They believe Beijing likely approved it. Vice President JD Vance and Treasury Secretary Scott Bessent have expressed concerns about the model's potential for cyberattacks on critical infrastructure. The White House is now considering an executive order for a formal review process for powerful AI models. This situation highlights the growing importance of advanced AI in national security.