Detecting Unfairness in Human-Robot Interaction

This research article focuses on identifying algorithm-based biases in robotic behavior and their impact on human-robot interaction (HRI) within mixed groups. The authors propose developing computational models to detect instances of microaggression, discrimination, and social exclusion. This involves observing human coping behaviors aimed at regaining social inclusion and utilizing system information that indicates unequal treatment among human participants. The paper discusses the role of groups in human lives, the psychological effects of social rejection, and the sources of algorithmic bias in robotic systems. It highlights how social robots can influence group dynamics positively or negatively, leading to feelings of exclusion when they fail to engage all group members adequately. The authors advocate for regulatory mechanisms to promote fairness and inclusion in HRI, addressing the need for a paradigm shift in research from individual interactions to group dynamics. The findings emphasize the importance of addressing biases in robotic systems to prevent negative emotional consequences for users.