Human-Likeness and Moral Responsibility in AI: The Roles of Dyadic Morality and Emotional Unpleasantness

초록

This study examines how the perceived human-likeness of artificial intelligence (AI) influences judgments of AI’s moral responsibility. Specifically, it investigates (1) the direct effect of human-likeness on moral responsibility, (2) the sequential mediating roles of perceived moral patiency and perceived moral agency, and (3) the moderating effect of emotional unpleasantness on the relationship between human-likeness and moral patiency. Data were collected from 270 adult participants in South Korea using an online survey. Results revealed a significant negative direct effect of human-likeness on moral responsibility, contrary to initial expectations. However, a significant positive indirect effect emerged through a sequential mediation pathway involving patiency and agency, suggesting a suppression effect. Emotional unpleasantness further moderated the association between human-likeness and moral patiency, such that the relationship was stronger when unpleasantness was high. These findings provide new insights into how the physical appearance of AI, moral cognition, and affective responses interact to influence moral evaluations in human–AI interactions.

키워드