How to Stop a Robot: Prototyping an Intuitive Kill Switch for Autonomous Systems

As AI-supported systems rapidly develop, mobile robots like the Boston Dynamics Spot are becoming increasingly autonomous. The more autonomous a robot is, the more likely it is to be trusted to correct itself in the event of errors. However, if the robot begins to move uncontrollably through the room, how can people switch it off easily and intuitively? Our work-in-progress addressed this research question and developed three approaches in a prototyping session with seven participants. They documented and evaluated ideas, all of which leaned toward a physical switch at a fixed location on a wall. However, the discussion also highlighted challenges such as limited accessibility, potential confusion with other security systems, and the need to deactivate individual robots selectively.