Should AI be allowed to make important decisions if it's proven to be more accurate than humans?