Stanford CIS

A Robot Really Committed A Crime: Now What?

By Ryan Calo on

It is always fun, and sometimes worrying, to see imagination come to life. I was on a panel last year at UC Berkeley around robotics and law. We talked about some of the conundrums robots and artificial intelligence might pose for law and policy–the subject of my forthcoming work Robotics and the Lessons of Cyberlaw. One hypothetical involved a shopping “bot” that randomly purchases items on the Internet. What if the bot purchases an item that is illegal in the jurisdiction where the item was shipped?

Given that criminal law commonly requires mens rea (Latin for an “intending mind”), it would seem the recipient of the package, even if she programmed the bot herself, might not be held criminally liable. After all, she intended to surprise herself, not to purchase contraband. This and other examples of emergent behavior create the interesting prospect of crimes without perpetrators, or so I argue.

Greg Miller did a wonderful write up of the exchange for WIRED, careful to note that the activity in question represents the speculation of a law professor (if based upon a live project by Darius Kazemi).

Well, now The Guardian is reporting that a version of this very thing has happened. Two London-based artists coded a bot that randomly purchased items from a hidden or “darknet” market using Bitcoin. The bot purchased, among other things, fake Diesel jeans and ten pills of ecstasy. But it also purchased perfectly lawful items such as a stash can and baseball cap with a camera in it. You can see the items here.

Read the full piece at Forbes.

Published in: Publication , Other Writing , Robotics