But that's a "safety risk", so for your own protection OpenAI will ensure that this technology remains safe in the hands of "trusted experts".