TY - GEN
T1 - Choosing the best robot for the job : affinity bias in human-robot interaction
AU - Trainer, Thomas
AU - Taylor, John R.
AU - Stanton, Christopher J.
PY - 2020
Y1 - 2020
N2 - ![CDATA[Humans subconsciously judge others as being either similar or dissimilar to themselves, manifesting as an unconscious preference, or affinity bias, for those who are perceived to be similar. In human-to-human interaction, affinity bias can significantly influence trust formation and lead to discrimination, for example, in decisions related to recruitment and team selection. We investigate whether affinity bias is observed in human-robot interaction during team formation with social agents that differ in gender and skin tone. In this study, we asked 61 participants to order the resumés of 24 different avatars that varied in gender, skin tone, and competency under the pretext of choosing the “best” avatars to be the participant’s teammate. Then, using a wizard-of-oz style experiment, participants performed a task with two avatar teammates (one most preferred and one least preferred) to measure trust. Results showed that while avatars were predominantly chosen based upon competency, avatar appearance generated an affinity bias in resumé sorting, and participants were more likely to trust their preferred teammate.]]
AB - ![CDATA[Humans subconsciously judge others as being either similar or dissimilar to themselves, manifesting as an unconscious preference, or affinity bias, for those who are perceived to be similar. In human-to-human interaction, affinity bias can significantly influence trust formation and lead to discrimination, for example, in decisions related to recruitment and team selection. We investigate whether affinity bias is observed in human-robot interaction during team formation with social agents that differ in gender and skin tone. In this study, we asked 61 participants to order the resumés of 24 different avatars that varied in gender, skin tone, and competency under the pretext of choosing the “best” avatars to be the participant’s teammate. Then, using a wizard-of-oz style experiment, participants performed a task with two avatar teammates (one most preferred and one least preferred) to measure trust. Results showed that while avatars were predominantly chosen based upon competency, avatar appearance generated an affinity bias in resumé sorting, and participants were more likely to trust their preferred teammate.]]
KW - human-robot interaction
KW - prejudices
KW - psychological aspects
KW - robots
UR - http://hdl.handle.net/1959.7/uws:57965
U2 - 10.1007/978-3-030-62056-1_41
DO - 10.1007/978-3-030-62056-1_41
M3 - Conference Paper
SN - 9783030620554
SP - 490
EP - 501
BT - Social Robotics: 12th International Conference, ICSR 2020 Golden, CO, USA, November 14-18, 2020, Proceedings
PB - Springer
T2 - International Conference on Social Robotics
Y2 - 14 November 2020
ER -