“Even if you have good mechanisms that solve the AI control problem, you still have the political problem of not letting it be used for bad intentions. So we still have the humanities’ or subset of humanities’ control problem.”

— Nick Bostrom, Effective Altruism Global 2015: Oxford University