The question of what to do when you only have imperfect choices is a useful illustration for how we should think about AI and ethics. We don’t yet know how smart artificial intelligence will be or how ...