A Mini-survey on Psychological Pillars of Empathy for Social Robots: Self-Awareness, Theory of Mind, and Perspective Taking

Recent studies in the field of Human-Robot Interaction (HRI) confirm the positive effects of robots' empathic behaviors in HRI. Most HRI studies investigating empathy, apply an empirical approach to implement empathy, i.e., the empathic model is derived directly from observations of empathic actions. This resulted in the emergence of numerous different empathic models that are only valid for a particular scenario that is highly tuned and, therefore, a slight modification in the scenario makes the corresponding empathy model infeasible. In fact, most of the proposed models suffer from a lack of generalizability. Since empathy is a complex concept that includes different dimensions, a coherent model of empathy that can be used in different scenarios or even be scenario independent, needs to consider several core concepts of empathy.Thus, the goal of this paper is to analyze and link different concepts of empathy and bring related existing models together, which can help researchers in the HRI community to have a better picture of an empathy model that might lead to the development of more general models of empathy for social robots.