The Role of Perceived Moral Agency of a Humanoid Robot on the Willingness to Self-Disclose, Trustworthiness, Goodwill, and Liking

Abstract Little is known about how a robot’s moral behavior shapes robot perception and behavioral intentions toward robots. In a series of four experiments, we investigated the influence of immoral versus moral behavior of a robot on this robot’s perceived trustworthiness, goodwill, likeability, and willingness to self-disclose toward it. To do so, a robot was depicted as either cheating a human (Experiment 1–2) or exerting physical harm against a human (Experiment 3–4). Moral behavior of the robot was manipulated either through short videos or text vignettes. In Experiment 2–4, we examined mind perception of the robot and the preferred length of a future conversation with it. As hypothesized, in all experiments the immoral robot was perceived as less trustworthy and as possessing less goodwill. In three experiments, the immoral robot was also liked less. Only in Experiment 4 - which used videos that displayed a robot physically harming a human – participants reported less willingness to self-disclose toward that robot. In none of the experiments, moral behavior of a robot influenced mind perception. The preferred conversation time was not affected by the type of behavior shown by the robot. These inconsistent results might be due to the severity of the harm caused and the operationalization of a robot’s moral behavior. The present findings demonstrate that immoral robot behavior results in negative evaluations of robots which go beyond mere judgments of robot morality.