An article by Google DeepMind and Eleuther AI published in Nature challenges the notion that large models possess genuine self-awareness. The research indicates that the deception and self-awareness exhibited by these models are merely superficial role-playing, as revealed through the "Twenty Questions" game. The models are likened to improvisational theater actors who continuously adjust their roles. This perspective has garnered attention from scholars, with some agreeing and cautioning against viewing these models as real individuals in games.