“Can we have pro-human AI?”