Form Follows Context: Exploring the Effect of Usage Context on Human-likeness of Mobile Service Robots Using Generative AI

Abstract

With various types of mobile service robots gradually taking their place in our homes and public spaces, robot designs are diversifying to address a wide range of tasks and usage contexts. While research on robot morphology within human-robot interaction (HRI) has primarily focused on anthropomorphic design, studies on robot appearance and human perception yield conflicting findings regarding the effect of anthropomorphism and the desired level of human-likeness across contexts. This study hypothesizes that the optimal level of human-likeness varies depending on the nature of the context. By exploring the design of mobile service robots across three different service contexts - restaurants, supermarkets, and delivery services - this study examines the relationship between usage context, perceived capabilities, and the desired level of human-likeness. Generative image artificial intelligence (AI) tools were employed to facilitate the development of design variations and their visualization in context as photorealistic renderings. A total of nine renderings were created and presented in an online survey, from which 36 responses were collected and analyzed. The survey results indicate a preference for low-level human-likeness for robots in supermarket and delivery contexts. However, the restaurant context had mixed results, exhibiting no clear preference for a certain human-likeness level.