Embodied Active Listening: How Non-Verbal Backchannel Behaviour Influences Trust and Perception of a Social Robot

With the trend to make conversational user interfaces (CUIs) more fluid and humanlike we must consider the impact of anthropomorphism, especially when the CUI is embodied in a device, robot or avatar. Non-verbal backchannel behaviour and active listening is a key research area in modern CUIs and in this paper we explore the impact of non-verbal backchannel behaviour on active listening using a zoomorphic robot, Miro. We carried out a mixed-methods experiment, involving 20 participants interacting with the robot under three behavioural conditions: static, responsive backchannel, and random backchannel. Findings reveal that while participants occasionally misinterpreted the robot’s non-verbal cues, there were also notable moments of engagement and perceived connection. In addition, a responsive backchannel increased the perceived trust towards the robot. This suggests that trust in this particular embodied CUI was tied to their perceived social competence, highlighting the role of non-verbal backchannel behaviour in shaping human perceptions of CUI design.