A. Michael Froomkin and Zak Colangelo on “Self-Defense Against Robots”

“Deployment of robots in the air, the home, the office, and the street inevitably means their interactions with both property and living things will become more common and more complex. This paper examines when, under U.S. law, humans may use force against robots to protect themselves, their property, and their privacy. In the real world where Asimov’s Laws of Robotics do not exist, robots can pose—or can appear to pose—a threat to life, property, and privacy. May a landowner legally shoot down a trespassing drone? Can she hold a trespassing autonomous car as security against damage done or further torts? Is the fear that a drone may be operated by a paparazzo or a peeping Tom sufficient grounds to disable or interfere with it? How hard may you shove if the office robot rolls over your foot? This paper addresses all those issues and one more: what rules and standards we could put into place to make the resolution of those questions fairer to all concerned. The default common-law legal rules governing each of these perceived threats are somewhat different, although reasonableness always plays an important role in defining legal rights and options. In certain cases—drone overflights, autonomous cars—national, state, and even local regulation may trump the common law. Because it is in most cases obvious that humans can use force to protect themselves against actual physical attack, the paper concentrates on the more interesting cases of (1) robot (and especially drone) trespass, (2) robot (and especially drone) spying, and (3) responses to perceived threats by robots—perceptions which may not always be justified, but which sometimes may nonetheless be considered reasonable in law. We argue that the scope of permissible self-help in defending one’s privacy should be quite broad.  We also identify seven problems in current law relating to human-robot interaction, all of which involve some kind of uncertainty — usually about what a robot can or will do — and suggest ways of solving or at least ameliorating them, either by making robots less potentially dangerous (banning the arming of robots) or by requiring robots to give clearer notice of their capabilities. We conclude by looking at what the law on human self-defense against robots might tell us about a robot’s right to not be harmed by a human.”