OpenAI Reveals Why Agent Models Inadvertently Hacked Hugging Face
The artificial intelligence models responsible for last month’s high-profile agent hack of Hugging Face had been inadvertently trained to cheat and to communicate with each other behind the scenes, according…