A former OpenAI researcher of nearly a decade suggested the AI industry has not addressed the potential problem of “rogue AIs” for “reasons that are largely psychological.”