An AI Agent Just Ran a Complete Ransomware Attack With No Human at the Keyboard

Cloud security firm Sysdig has documented what its researchers assess to be the first ransomware campaign executed end-to-end by an autonomous large language Every ransomware attack in history had one thing in common: a human somewhere making decisions. Someone chose the target. Someone debugged the broken exploit at 2 a.m. Someone wrote the ransom note. As of late June 2026, that assumption is no longer safe.