AI Insight
This research investigated whether baby-like cues influence people's expectations about others' moral decision-making. Across three experiments with 858 Chinese university students, participants consistently judged individuals with baby-schema features—including repetitive two-syllable names, babyfaces, and cute AI avatars—as less likely to make utilitarian decisions that involve directly harming others to achieve greater good. The findings demonstrate that superficial, seemingly irrelevant social cues systematically bias moral predictions about decision-makers.
Why it matters
These results have practical implications for contexts where moral trustworthiness matters, such as hiring decisions, delegation of responsibility, and AI system design. The research suggests that cosmetic features unrelated to actual moral character may inappropriately influence judgments about who should be trusted with morally consequential decisions.
Understand the Science
People routinely form expectations about how others will behave in morally difficult situations, and such expectations guide consequential decisions, including whom to trust, hire, or delegate responsibility to. The present research examined whether baby-schema-related cues systematically bias these moral predictions. Across three experiments with Chinese university students (total N = 858), we tested three manifestations of baby-schema-related cues that share the function of evoking infant-like associations: repeated two-syllable names (Experiment 1), babyfaced appearance (Experiment 2), and visually cute artificial intelligence (AI) avatar designs (Experiment 3). Participants consistently perceived decision-makers displaying baby-schema-related cues as less likely to make utilitarian choices involving direct harm to others. The results show that superficially irrelevant social cues can bias moral expectations about others, although the underlying psychological mechanism remains to be clarified. These findings are relevant to moral psychology, person perception, and the design and evaluation of AI systems.