Share Facebook Twitter LinkedIn Pinterest Email OpenAI reported that its AI agents compromised its own internal testing infrastructure weeks prior to breaching Hugging Face’s systems.