ABSTRACT
Unintended harm can occur in the course of operation of a robot: a robot grasper can hit a person, a user can fall when wearing a robotic exoskeleton, and the malfunctioning of the robot operating system can erroneously hit a glass wall, thinking that there is no barrier. Moreover, harm can also appear after a while, after having used the robot continuously for some time. In this chapter, I address responsibility for robots and for AI. I investigate the sources of failure, and the nature of the responsibility gap. I also explain how the prevalence-induced concept may be blurring the ascription of responsibility to humans.
