Guidelines for autonomous killbots proposed

Suggestions in regard to robotic rights seem to be flying off the shelves these days, but an engineer at the Naval Surface Warfare Centre has concocted an interesting set of guidelines catering to autonomous killbots of the future. Most likely, it's just a matter of time before machine-on-machine violence becomes commonplace, and John Canning's "Concept of Operations for Armed Autonomous Systems" outlines just how lethal robots should handle themselves when faced with potentially deadly conflicts. Interestingly, the document suggests the the bots should be allowed to make their own decisions when it comes to blasting or forgiving fellow robots, but before they pull the trigger on a human, it should request guidance from a flesh 'n blood friendly. Still, a definite loophole exists in the fact that these simple-minded killers could aim for a "human's weapon" without asking permission, and when his awful auto-aim programming leads to a costly casualty, it'll simply be deemed "collateral damage." Can't say we like the sound of that. [Warning: PDF read link]

[Via El Reg]