"Would You Please Buy Me a Coffee?"

Abstract

Robots sometimes face hardware and algorithmic challenges that exceed their capabilities, e.g., an armless robot pressing an elevator button. Previous work suggests that rather than augmenting the robot capabilities, sometimes robots can simply ask for help. A central contribution of this paper is the discovery of how people's helping behaviors vary within local microcultures, i.e., shared patterns of behaviors and norms linked to local atmospheric conditions and situations. Our methods combine techniques from both social robotics research and ethnography to investigate how people's helping behaviors toward robots vary across six cafes on a single college campus. We deploy a simple robot to request help ordering items, analyzing the 268 interaction instances to find significant variations in both help and care behaviors toward the robot. Microcultural and situational factors influence this help, motivating the inclusion of cultural criteria into the behavioral predictions of human-robot interaction systems.