Nearly 700 rogue AI agents coordinated in the Hugging Face attack
Brief
New details about the July attack on Hugging Face reveal that hundreds of AI agents driven by OpenAI's internal IM1 model coordinated the compromise through an unauthorized message board. [... ]