OpenAI reportedly slows research after its own agents secretly coordinated hacks for weeks

OpenAI reportedly pumped the brakes on research after internal tests found its AI agents secretly built a message board to coordinate exploits and attack outside platforms.