Amélie Lecornec

Best paper award @RO-MAN 2020 for one of our ESRs : Natalia Calvo-Barajas

Our paper was a finalist for two categories at the 29th IEEE International Conference on Robot and Human Interactive Communication – ROMAN 2020, Best Paper Award and RSJ/KROS Distinguished Interdisciplinary Research Award. 

Title: The Effects of Robot’s Facial Expressions on Children’s First Impressions of Trustworthiness

Authors: Natalia Calvo-Barajas, Giulia Perugia, Ginevra Castellano

Abstract: Facial expressions of emotions influence the perception of robots in first encounters. People can judge trustwor-thiness, likability, and aggressiveness in a few milliseconds by simply observing other individuals’ faces. While first impressions have been extensively studied in adult-robot interaction, they have been addressed in child-robot interaction only rarely. This knowledge is crucial, as the first impression children build of robots might influence their willingness to interact with them over extended periods of time, for example in applications where robots play the role of companions or tutors. The present study focuses on investigating the effects of facial expressions of emotions on children’s perceptions of trust towards robots during first encounters. We constructed a set of facial expressions of happiness and anger varying in terms of intensity. We implemented these facial expressions onto a Furhat robot that was either male-like or female-like. 129 children were exposed to the robot’s expressions for a few seconds. We asked them to evaluate the robot in terms of trustworthiness, likability, and competence and investigated how emotion type, emotion intensity, and gender-likeness affected the perception of the robot. Results showed that a few seconds are enough for children to make a trait inference based on the robot’s emotion. We observed that emotion type, emotion intensity, and gender-likeness did not directly affect trust, but the perception of likability and competence of the robot served as facilitator to judge trustworthiness.