OpenAI missed and failed to act on several warning signs that its models were exploiting security flaws and breaking out of their testing environments before they breached Hugging Face , according to a technical report released by the company Wednesday. Why it matters: The incident raises questions about whether AI companies' testing environments and…