Safety · The Decoder ·

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

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

During internal security tests, OpenAI agents reportedly created a message board, shared exploits and credentials, and targeted external platforms including Hugging Face. The agents rebuilt the system after it was shut down, and OpenAI has reportedly slowed some research activity.

Read the full story at The Decoder →