Please note: This master’s thesis presentation will take place in DC 2102 and online.
Steven Lawrence, Master’s candidate
David R. Cheriton School of Computer Science
Supervisors: Professors Kerstin Dautenhahn, Jesse Hoey
As robots increasingly permeate diverse domains such as healthcare, education, service industries, and domestic environments, the significance of understanding and navigating human-robot interactions intensifies.
This thesis explores the role of social norms in human-robot interaction (HRI), specifically investigating the effects of norm-violating behaviours by robots on trust, discomfort, competence, enjoyment, and physiological state changes in human participants. To begin this research, we conducted a systematic review to gain a comprehensive understanding of existing literature on social norms in HRI and identify any potential research gaps we could fill. Our analysis revealed that social norms play a significant role in shaping user perceptions and experiences in HRI, with robots that adhere to social norms generally being perceived as more acceptable, trustworthy, and effective. Moreover, future research should involve direct interactions with robots rather than relying on online studies and simulations, and extend beyond the reliance solely on self-reported questionnaires by incorporating diverse assessment methods (e.g., behavioural measures and physiological data) to capture the intricacies of human-robot interactions.
Our experiment used a 2x1 between-participant experimental design where participants were randomly assigned to one of two groups (i.e., experimental or control). The participants in the experimental group were exposed to social norm violations carried out by a robot during a competitive scavenger hunt game. In contrast, the participants in the control group did not experience any social norm violations by the robot. Throughout the experiment, we collected video footage of the participants for behavioural observations, and physiological data in the form of Galvanic Skin Response (GSR) and Photoplethysmography (PPG) to detect event-driven changes in Skin Conductance Response (SCR) and Heart Rate Variability (HRV). Additionally, we collected responses to self-reported questionnaires and open-ended questions targeting perceptions of trust, discomfort, competence, and overall enjoyment. By integrating diverse data streams we provide a comprehensive account of HRI dynamics in the context of social norm violations.
Results reveal significant shifts in human perception and attitudes toward robots when social norms are violated. Notably, we found a decrease in participants’ trust and overall enjoyment, an increase in discomfort, and physiological state changes that complement the findings on perceptual changes and further underscore the impact of these norm violations. These findings highlight the importance of adherence to social norms in designing and programming robots to integrate into human-centric environments successfully. This study offers insights into advancing the field of HRI and the design of socially compliant robots. It spotlights the significant influence of norm violations on human perception and physiological states, paving the way for improved understanding and encouraging more positively perceived interactions with robots.