The body images of social robots

The rapid development of social robots has sparked a challenge for both robotics and cognitive sciences to comprehend how humans perceive the appearance of robots. This understanding is a crucial prerequisite for achieving successful human-robot symbiosis. To uncover people's perceptions and attitudes towards robots, we analyzed image-associated words generated spontaneously by humans for 30 robots developed in the past decades. These words delineated a body image for each of the robots. We then used word affective scales and embedding vectors to provide evidence for links between human perception and the body images. Our findings revealed that the valence and dominance of the body images reflected human attitudes toward the general concept of robots. The study further demonstrated that the user-base and usage of robots significantly influenced people's impressions of individual robots. Moreover, we investigated the psychological and cultural implications of the body images by examining semantic distances between the robots and a human-related word, as well as gender- and age-distinguished words. This analysis revealed a relationship between the semantic distances to the word “person” and the robots' affects, as well as gender and age stereotypes towards the robots. Our study demonstrated that using words to build body images for robots is an effective approach to understanding which features are appreciated by people and what influences their feelings towards robots.