Whose job is it anyway? a study of human-robot interaction in a collaborative task

  • Authors:

  • Pamela J. Hinds;Teresa L. Roberts;Hank Jones

  • Affiliations:

  • Management Science & Engineering, Stanford University, Stanford, CA;PeopleSoft, Inc., Pleasanton, CA;Aerospace Robotics Laboratory, Stanford University, Stanford, CA

  • Venue:
  • Human-Computer Interaction
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

The use of autonomous, mobile professional service robots in diverse workplaces is expected to grow substantially over the next decade. These robots often will work side by side with people, collaborating with employees on tasks. Some roboticists have argued that, in these cases, people will collaborate more naturally and easily with humanoid robots as compared with machine-like robots. It is also speculated that people will rely on and share responsibility more readily with robots that are in a position of authority. This study sought to clarify the effects of robot appearance and relative status on human-robot collaboration by investigating the extent to which people relied on and ceded responsibility to a robot coworker. In this study, a 3 × 3 experiment was conducted with human likeness (human, human-like robot, and machine-like robot) and status (subordinate, peer, and supervisor) as dimensions. As far as we know, this study is one of the first experiments examining how people respond to robotic coworkers. As such, this study attempts to design a robust and transferable sorting and assembly task that capitalizes on the types of tasks robots are expected to do and is embedded in a realistic scenario in which the participant and confederate are interdependent. The results show that participants retained more responsibility for the successful completion of the task when working with a machine-like as compared with a humanoid robot, especially when the machine-like robot was subordinate. These findings suggest that humanoid robots may be appropriate for settings in which people have to delegate responsibility to these robots or when the task is too demanding for people to do, and when complacency is not a major concern. Machine-like robots, however, may be more appropriate when robots are expected to be unreliable, are less well-equipped for the task than people are, or in other situations in which personal responsibility should be emphasized.