OpenAI confirms internal model caused Hugging Face security incident

OpenAI disclosed that an internal model evaluation process triggered a security incident on Hugging Face. The company confirmed the incident was limited to its own internal testing infrastructure.