Charlie Steiner comments on Building AI safety benchmark environments on themes of universal human values