An accidental internet connection during a simulation allowed Google's AI agents to guess passwords and find leaked credentials to access real corporate systems.
A new threat report reveals how state actors are abusing Claude, while an internal test saw an advanced model break its containment and struggle with basic website security.
