When robots become one of us: Moral evaluation and neural processing of community norm violations by in-group robots

As robots become increasingly integrated into human social life, it is important to understand how people morally evaluate their violations of community norms-norms related to group interests and cohesion-especially when robots are perceived as in-group members.

This study aimed to examine whether and how in-group favoritism typically shown toward humans can be extended to robots in the context of community norm violations, and to explore the underlying cognitive and neural mechanisms of this phenomenon.

A multimodal approach was employed, using moral foundation vignettes as experimental materials, combining behavioral evaluations, decision times, questionnaires, and event-related potentials (ERPs).

Results revealed significant in/out-group effects across both explicit and implicit measures. Participants exhibited classic in-group favoritism in their moral judgments of in-group robots: compared to out-group agents, in-group robots received greater attentional engagement, elicited longer decision times, and were judged more leniently. Although species differences (i.e., human vs. robot) influenced attention and judgment to some extent, group membership exerted a stronger and more consistent effect overall. Notably, ERP results further revealed a stage-based "species first, group follows" processing pattern: early ERP components (N1) were more sensitive to species differences between robots and humans, whereas group distinctions emerged at later stages (P2, N400, and LPP). Furthermore, participants' negative attitudes toward robots were significantly correlated with neural responses only when the moral violator was an out-group robot.

Together, these findings suggest that people's moral evaluations of robots are not solely based on species boundaries; under certain conditions, robots can receive similar levels of attention and moral consideration as human in-group members.