Gender-Emotion Stereotypes in HRI: Exploring the Role of Gender and Speech Act on the Evaluation of Social Robots

Abstract

In the field of Human-Robot Interaction (HRI), there has been increased awareness that humanlike design of social robots could potentially reproduce societal biases. This study contributes to prior HRI research by conducting a behavioral experiment (n = 194), focused on the impact of stereotyping effects concerning a robot’s gender and speech act on people’s evaluation of warmth, competence and discomfort. The experiment manipulates the Pepper robot’s gender (male or female) and the robot’s speech act (assertive or affiliative speech) in an online video setting with a between-subjects design. The findings revealed that female robots are ascribed higher competence than male robots, independent of their speech act, and assertive robots are ascribed higher competence than affiliative robots, independent of their gender. Additionally, it was found that participants identifying as female tend to perceive the robot as more competent than those identifying as male. The evaluation of warmth or discomfort do not show significant effects, although there seems to be a tendency for assertive male robots to receive higher discomfort ratings. The results of this study suggest that robot gender as well as user gender and the associated norms and expectations may have complex effects on HRI, mainly in the competence dimension. This highlights the importance of designing social robotics without reinforcing gender bias. Moreover, it emphasizes the need for more theory-driven experiments to address gender issues in HRI, while considering the complexity and diversity of the concept of gender.