Using User-Generated YouTube Videos to Understand Unguided Interactions with Robots in Public Places

Professional service robots are increasingly being deployed in public places, which thus increases user exposure. However, we lack an empirical understanding of complex encounters taking place in dynamic and often crowded environments as well as how people overcome breakdowns during unguided interaction with a robot in a real-world scenario. In this paper, we conducted a covert, digital ethnographic study analyzing 104 user-generated YouTube videos focusing on people’s unguided interactions with robots in several public places. We identified several types of interaction breakdowns pertaining to someone (person-initiated interaction breakdown, IB) or something (environmental disturbances, ED) having a direct, negative effect on an ongoing unguided interaction. Our findings have implications for the design and development of service robots facing multi-user scenarios entertaining active (primary and secondary) users, inactive (commentators and observers) ‘users’, and Incidentally Co-present Persons (InCoPs) . Furthermore, we contribute to and built on the limited prior use of YouTube videos and digital ethnography in HRI research, thereby demonstrating its effectiveness in studying unguided interactions in public places, while supplementing and adding to the existing knowledge base of service robots in public places.