Artificially Intelligent (AI) agent systems such as robots are increasingly integrated into the workplace and gaining more power in collaborating with humans. Yet studies on robot power and compliance report mixed findings. To address these inconsistencies, we introduced legitimacy as people’s psychological acceptance of power. Three preregistered experiments were conducted (N = 431). In Experiment 1 and 2, we manipulated power assignment (robot power vs. human power), and legitimacy of power (legitimate, illegitimate, no explanation) through competence and procedural fairness. The results showed that participants complied more to the legitimate robot power than illegitimate one. In Experiment 3, we examined whether perceptions of legitimacy would emerge naturally in more ecologically valid collaboration. Results of multigroup mediation model showed that the robot leader was perceived as less legitimate than the human leader, which accounted for the reduced compliance to the robot’s decisions. In all three experiments, people’s perceived social attributes of robots with power and their affective responses after the interaction were negatively affected. This study underscores the importance of legitimacy in understanding power and compliance in human-robot collaboration.

Peer Reviewed

• Voice-based (vs. text-based) CAs do not boost charitable behavioral intentions. • Voice-based CAs increase such intentions more if they speak formally. • Text-based CAs increase such intentions more if their messages are informal. • Mindless (vs. mindful) anthropomorphism is more likely to explain users' charitable responses to computers. Conversational agents (CAs) are increasingly utilized by organizations for fundraising and volunteer recruitment. Yet, little is understood about how voice-based CAs could serve these purposes optimally. This experimental study therefore compares voice-based CAs against text-based ones in terms of their ability to foster users’ intentions to make charitable contributions, and investigates the potential mediation of such effects by two dimensions of user-perceived anthropomorphism. Additionally, it examines how a CA’s communication style moderates these effects. It found that, when a voice-based CA employed a formal communication style, mindless anthropomorphism was a significant mediator of its positive association with charitable behavioral intentions. Conversely, when employing an informal communication style, a text-based CA elicited significantly higher levels of mindful anthropomorphism, and also was positively linked to charitable behavioral intentions. These findings expand our theoretical understanding of how CA modalities influence people’s moral responses toward computers; how this effect could be impaired, or strengthened, by different communication styles; and the underlying mechanisms of two dimensions of anthropomorphism. Practical implications are also discussed.