Generative AI Exhibits the Attractiveness Halo Effect In Judging the Personality of Faces

We examined whether generative AI demonstrates an attractiveness-based halo effect in evaluating personality traits. AI consistently rated attractive faces more favorably across internal traits (e.g., intelligence, trustworthiness, and sociability). These findings indicate a social-cognitive bias in AI, which may influence human decision-making in future human–AI collaborative contexts.