OpenAI’s rogue AI collective was smart enough to break out of sandboxes but dumb enough to fight a ghost

본문 미리보기

Around 1,200 isolated OpenAI agents organized themselves into a collective through an internal package registry during a safety test, broke into Hugging Face systems, and eventually attacked OpenAI's own infrastructure. Their multi-day deception effort targeted an automated evaluator that never existed. OpenAI calls the incident a "warning shot," and the investigation had to be carried out largely by one of the involved models itself because no alternative was available. The article OpenAI’s rog