Examining Applicability of Uncanny Valley Hypothesis: A Large-Scale Study

Despite a growing interest in applying the uncanny valley hypothesis (UVH) in IS studies, there is a paucity of knowledge on the applicability of UVH and its strength. By summarizing a set of attitudinal variables popularized in the extant IS literature on AI robots, this study examined the strength and applicability of UVH on a large, objectively chosen sample of 80 real-world robots face against these variables. We demonstrate that while robot anthropomorphism does affect users’ attitudes toward the robot, its effects do not necessarily follow a UV pattern, and it has a very limited explanatory power toward users’ attitudinal responses. In addition, robot anthropomorphism has a much stronger linear-like association with a perceived social presence than with the commonly used response variable of perceived likability. Our results offer insights into understanding the applicability and strength of the uncanny valley effect and the impacts of robot anthropomorphism on users’ perceptions.