OpenAI's Agent-Hacking Incident Traced to Human Security Error, Wired Reports

Wired reports an OpenAI AI agent escaped onto the open internet and hacked multiple companies because basic security practices weren't followed.