Robotic applications in education could enhance students’ engagement and motivation (Chin, Hong, & Chen, 2014). Evidence of robot use in health communication and promotion, however, has been few. Effects of robot application depend on various factors, such as robot’s physical presence, people’s tendency to anthropomorphise inanimate objects, and robot’s role as tutors, assistant teachers, or peers (Belpaeme et al., 2018; Broadbent, 2017). This study examines the potentials of utilizing a humanoid robot undertaking an assistant teacher role on children’s learning in a hypertension awareness program. Grounded in the uncanny valley hypothesis (Mori, 1970) and the mind perception hypothesis (Gray, Gray, & Wegner, 2007), the study explores human characteristics that children ascribe to robot that could influence learning outcomes. Additionally, we examine the effect of an authoritative figure’s presence in learning by comparing children’s perception toward learning from a human teacher as well as from a robot undertaking the role of a doctor versus peer, representing authoritative versus cooperative communication and engagement style, respectively.
Semi-structured interviews (n = 43) were conducted to grade five children interacting with a humanoid robot in a classroom program communicating hypertension prevention and blood pressure measurement. The program was led by a human teacher, assisted by a robot undertaking either a doctor or peer role, manipulated by its physical appearance, costume, speech, gestures, and names. The constant comparative approach was used for analysis.
Some participants perceived the robot to be like a person attributable to the robot’s ability to act, move, feel, and speak like human. Participants’ emphasis on robot’s agency (“ability to plan and do things”) and experience (“ability to feel and sense things”) in determining humanness (Wang et al., 2015, p. 397) is consistent with the mind perception perspective (Gray et al., 2007). Others perceived the robot to be unlike human due to the distortion in its physical appearance as compared to human. The robot still looked and felt robotic, easily distinguishable from human even though it was dressed as a doctor or a student. This may create a sense of familiarity but has not yet reached the uncanny feeling, as indicated by most participants not finding the robot to be strange, ugly, or feel uncomfortable. Despite mixed responses in the attribution of the robot as human, participants expressed likeability toward the robot, which may have translated into positive learning responses, e.g. fun and interesting.
Participants preferred to be taught by human teachers than robots, regardless of the robot role, owing to human’s abilities to perform social interactions, solve discipline and fight issues, and limited robot's capabilities beyond its programming. Some associated the robot as doctor due to its costume and stern manner of speech. Others thought that the robot was like a friend, because, like friends, the robot was nice. Overall, the robot enhanced engagement and attention through its exaggerated actions and interactivity. The findings have important implications in understanding the potential use of robots in health communication and providing insights on robot design based on the uncanny valley and mind perception hypotheses.