The Effect of Conceptual Embodiment on Human-Robot Trust During a Youth Emotion Classification Task.

Interactive robots are increasingly being used in social environments to provide humans with information, guidance and even recommendation. User trust plays an important role in the human decision-making process for whether to accept or reject a robot's recommendation. Prior work in HRI has investigated various robot-related, task-related, and user-related factors that influence human-robot trust. This work further examines task and user-related factors influencing user trust during a robot-assisted emotion classification task when the embodiment of the robot is left for the user to conceptualize. The chosen task-related factors manipulate the level of certainty in the images to be labeled and the gender alignment between the individual in the image stimuli and the participant. We further consider participant gender, parental status, level of interaction with children, level of comfortability with robots and conceptual robot embodiment. An online between-subjects experiment was conducted with 241 participants. Experimental results show that task certainty and gender alignment along with level of comfortability with robots, level of interaction with children and conceptual embodiment all had significant effects on human-robot trust. Notably, participants who described their conceptual embodiment of the robot with human-like characteristics reported higher levels of trust than those who did not. These findings provide insights that contribute to a greater understanding of factors influencing human-robot trust during a situated scenario involving robot recommendation.