How a Robot's Social Credibility Affects Safety Performance

This paper connects the two domains of Human-Robot Interaction (HRI) and safety engineering to ensure that the design of interactive robots considers the effect of social behaviours on safety functionality. We conducted a preliminary user study with a social robot that alerts participants during a puzzle-solving task to a safety hazard. Our study findings show an indicative trend where users who were interrupted by a socially credible robot were more likely to act to mitigate the hazard than users interrupted by a robot lacking social credibility.