Washington’s AI Playbook Shifts to Security Tests, Not Gatekeepers

A new U.S. executive order puts frontier AI under a voluntary security lens, while Europe keeps betting on formal model obligations, incident handling, and cybersecurity duties.