Testing Empathy with Robots: A Model in Four Dimensions and Sixteen Items

The four-dimensional model of empathy presented in this paper addresses human-human, human-avatar and human-robot interaction, and aims at better understanding the specificities of the empathy that humans might develop towards robots. Its first dimension is auto-empathy and refers to an empathetic relationship with oneself: how can a human directing a robot expand the various components of empathy he feels for himself to this robot? The second is direct empathy: what does a human attribute to a robot in terms of thoughts, emotions, action potentials or even altruism, on the model of what he imagines and attributes to himself? The third dimension is reciprocal empathy that consists of thinking that a robot is able to identify with me, feel or guess my emotions and thoughts, anticipate my actions and wear me assistance if necessary. Finally, the fourth dimension, intersubjective empathy, is about thinking and imagining that a robot can inform me of things - emotions, thoughts that I am likely to experience- that I do not know about myself. Each of these four dimensions includes four different components: (1) Action (empathy of action), (2) Emotion (emotional empathy), (3) Cognition (cognitive empathy) and (4) Assistance (empathy of assistance). This theoretical model of empathy in four dimensions and four components defines sixteen items whose relevance will be tested in the near future through comparative experimental research involving human-human and human-robot interaction.