Effect of robot's title in human-robot interaction

The collaboration between humans and intelligent agents keeps increasing as the technology advances. In this sense, each can rely on the other's strengths to achieve the best overall performance. But as intelligent agents play an increasingly escalating role as a partner rather than mere tools, it is imperative that we understand how humans are influenced by such agents. This is so as to achieve a certain level of persuasiveness to ensure human compliance with intelligent agents in times of need, and also prevent over-reliance on such agents, which may make humans susceptible to manipulation. Reviewing the literature from human-computer interaction, persuasion, and decision-making processes, this study examined the effect of labelling source expertise (i.e., title) on the human decision-making process, when there is not an agreement between human and robot partners. Specifically, we conducted an experiment (n = 88) to investigate how perceived expertise of a robot resulted in the attractiveness, intelligence and credibility of the robot, and how these perceptions influence the level of cooperation and compliance in a desert survival game. The results showed that participants evaluated the robot as more credible when labelled as expert, compared to when labelled as novice. The results of a mediation analysis also showed that perceived credibility successfully mediated the effect of title on cooperation and compliance when there was a conflict in the ranking order of items between each participant and the robot. Implications for future development of artificial intelligence and future research directions are discussed.