A preliminary study on Reading the Mind in the Eyes of the Robot

Abstract

As social robots are increasingly deployed, the interaction between human and robot is becoming more important. Affective theory of mind (ToM) is one of the core components of emotional communication between human, enabling recognize and understand other’s emotional state as an independent agent separate from oneself. Current study developed robot version of widely adopted test of affective ToM, Reading the Mind in the Eyes Test (RMET) using a social robot, Vector. In the preliminary study, we compared the participants performance in both robot version, ingroup (East Asian) and outgroup (Caucasian) human version of RMET, the individual variance of affective ToM toward human expression is reproduced to the expressions of the Vector robot. Moreover, only the performance toward the expressions of same ethnicity (RMET-K) predicts the performance on the robot expression (RMET-R). The result suggests that not only the quality of robot’s expression but the variance of social cognition level of the human should be considered in social robot development.