That could means OpenAI does not take any shortcuts when it comes to safety.