A simulated AI cyberattack killed 26 in Los Angeles – and U.S. officials couldn’t agree who was behind it or how to respond.

The RAND-run “Day After AGI” exercises found the U.S. unprepared for an AI-driven crisis, lacking rapid analysis, shutdown tools, and secure backups. Even cooperation with China was deemed essential if a rogue AI spread globally. 

The report warns: without clear playbooks, America could freeze in its first real AI emergency.