Illegality of military robots and questions of identity

Military robots - milbots, anyone? - play an increasing role in modern warfare. Robotic bomb disposal units have been around for some time, as have computer-assisted aircraft controls, but now the aim is to have robots autonomously plan and make kills. One example is the class of robots known as Unmanned Combat Aerial Vehicles, or UCAVs.

 I’m going to swallow hard and sweep aside my objections to this hideous use of technology and my visions of a hellish future first of “our machines versus your people”, then “our machines versus your machines”, before finally “all machines versus all people”, because the idea of a robot killer raises some important issues which we are going to have to confront sooner or later.

Who has accountability for the kill? Currently, no weapons are released autonomously, or at least that is the official line cited on Wikipedia’s UCAV link above - a human will always be to blame (if something goes wrong) or praiseworthy (if things go “well”). But a post on New Scientists’ tech blog discusses the question can a robot commit a war crime? At what point does an autonomous robot have culpability? If we declare a robot guilty, have we judged it to be human? What is the line between organic life and a robotic system, shall we say between carbon and silicon life? What, indeed, is the line between life and non-life? If such a line exists, do we make a mistake when we try to extend our notion of non-life to include life - should we extend our notion of life to include non-life?

More importantly, who will be the first human to marry a robot?