some have argued that the use of technology has or will make the world a safer place. yet, the logic of drones is the same utilitarian calculus that facilitated the dropping of the A-bomb. we would lose less lives through the use of a weapon that could end the war (and would come out ahead in terms of cost-versus an ongoing war against japan).
drones and their eventual progeny are perfect soldiers-they never tire, will not go AWOL, do not require medical attention and are not concerned about self-preservation. maybe that is why the skynet analogy is used so often.
as to your question, however, it is not the algorithm itself but rather the nebulous region from which the inputs for the former come from. the scientist acting under the au es of the military. how often has that pairing proved to be rather problematic ?