Supporting healthy dyadic human relationships with power differentials using robots

Abstract

Conflict is a natural part of ever-evolving human-human relationships. The way in which conflicts are handled can result in relationship growth, dissatisfaction (for one or both parties), or relationship dissolution. Hierarchical relationships often handle conflict in ways that result in the lower-power member being dissatisfied with the disagreement's outcome. This thesis explores how a robot, a piece of technology that can be persistently present during hierarchical human-human interactions, can help to support healthier conflict processes in these relationships when humans cannot. In order to help, the robot must be able to identify problematic relationship states in the interactions. This work introduces a novel and general computational model to identify six problematic relationship states seen in hierarchical relationships. It couples this model with original robotic behaviors and an action-selection mechanism that allows a robot to successfully intervene to support the amelioration of these problematic states. This computational architecture is operationalized and partially implemented building upon work from human mediation, conflict theory, and occupational therapy. An intervening autonomous robot that is responsive to problematic relationship states using this computational architecture is compared to an unengaged robot (a robot that is not responsive to the human-human conversation) using a series of human-robot interaction (HRI) studies. The results of these studies show the potential for robots to make positive differences in human-human communication and suggests future research directions for prosocial technologies.