The Effect of Interdependence and Co-presence on Compliance with a Rule-breaking Robot

Abstract

As robots and virtual agents become integral workplace teammates, reliance on them increases. This increases the potential for conflicts between robot-issued guidance and established workplace rules. Therefore, it is vital to investigate which factors make individuals more likely to comply with rule-violating requests from autonomous agents in workspace settings. This thesis experiment investigated the effects of interdependence and co-presence on human compliance with rule-breaking requests from a virtual robot in a VR warehouse environment.The experiment involved two interdependence conditions: collaboration (high interdependence and co-presence) and coexistence (low interdependence and co-presence). Participants were involved in a box sorting task based on color. In the collaboration condition the robot would work with the participant by bringing them the boxes, while in the coexistence condition the robot would be doing unrelated work next to the participant. At certain points, boxes with ambiguous colors would show up, prompting the robot to request an action from the participant which violated the established rules set by the experiment. Results indicated no significant differences in compliance rates or hesitation times between conditions. The qualitative data pointed to task logic overshadowing the social influence of interdependence. Methodological limitations, such as short task duration, ambiguous instructions, and potential ceiling effects, were identified as possible overshadowing factors. We critically evaluated our design choices, and came with suggestions for future research involving extended tasks, emotionally charged scenarios, and explicit measures of perceived interdependence. This study contributes valuable insights into how contextual factors can dominate interdependence in human-robot interactions. This study establishes a theoretical foundation and provides practical guidelines for improved experimental design. We also published licence free material to use in future robotic VR experiments.