Operationalising Social Bonding in Human-Robot Dyads Through Physiological and Biobehavioural Proxies of Oxytocin

Given our evolutionary disposition as social animals, it is perhaps inevitable that humans will develop affective social bonds with artificial entities, such as robots. In natural systems, social bonding plays a powerful role in regulating behaviours, emotions, and even physiological health. If this is to extend to human-robot dyads, the phenomenology of these “invisible” bonds (including their emergence and dynamics over time) must be better understood. Current approaches that try to identify and quantify social bonding between humans and robots rely largely on post-hoc, qualitative assessments. Yet, biological evolution has provided us with visible signatures of social bonding—the neuropeptide oxytocin and its biobehavioural proxies—which may provide a starting point from which to operationalise affective social bonding between humans and robots. In this paper, we put forth three physiological and biobehavioural proxies of social bond formation, mediated by oxytocin, that can be measured in (real-time) human-robot contexts: (a) reductions in physiological stress (the “social buffering” phenomenon), (b) narrowing of spatial proximity between dyads, and (c) interdyad behavioural synchrony. We provide relevant evolutionary and neurobiological evidence for each of these measures, and offer some suggestions for how they can be recorded in (real-time) human-robot interaction scenarios, to measure social bond formation in human-robot dyads. With this paper, we aim to open the discussion towards identifying and developing methods for more robust operationalisations of “social bonding” between humans and artificial (robotic) agents.