An Ethical Skynet?
Over at the Cato Institute, Nat Hentoff makes the case for, in so many words, an ethical skynet:
The Economist’s report on ‘remote-control warfare’ refers troublingly to an ongoing refinement in automated warfare aimed at answering those here and abroad who are questioning the ethics of this futuristic form of combat. Cited is Ronald Arkin of the Georgia Institute of Technology’s School of Interactive Computing.
He proposes involving the drone itself — or rather, the software that is used to operate it — in the decision to attack.
‘In effect,’ the article continues, ‘he plans to give the machine a conscience.’
Is this science-fiction? As I will demonstrate next week, Arkin is not alone among American high-tech explorers devising non-human target killings in attacks on terrorism. To elaborate on the inventive Arkin approach, “The software conscience that Dr. Arkin and his colleagues have developed is called Ethical Architecture.”
During attacks, the judgment of the automated and autonomous Predator or Reaper drone ‘may be better than a human’s because it operates so fast and knows so much. And — like a human but unlike most machines — it can learn.’ After a strike, this ever-alert machine can indeed learn from other sources whether the damage it caused — including dead civilians — exceeded its intentions.
With this information, a drone with a conscience can more precisely tailor future attacks and instruct other drones on how to more carefully direct their Hellfire missiles. Thereby, these ethical drones can provide support to future American officials defending the use of killer drones by showing how carefully the United States is working to be humane in its self-defense against international terrorism.
Full post here.
It’s a real interesting proposal, one that would be extremely compelling if it could work out. I’m not a programming expert, so I’m not sure just how feasible this is, or how long it would take to implement. I will say that I am incredibly reticent of putting the sole responsibility of appropriate force to use when killing people in a machines sole hands. There needs to be some sort of overall human control in the process, if nothing else because it removes the onus of responsibility when civilian deaths, or hell any deaths, occur. Seriously, who gets blamed when the drones kill an innocent bystander? It’s not like there’s a course of action that can legally be taken against a machine. If a machine decides to use excessive force because of a programming error, or it calculates that it’s needed to complete the mission then there would be nothing to stop this autonomous device from unneeded violence. Barring those *minor* problems, I think that at the very least these ideas are worth looking into. After all, what’s the worst that could happen when implementing killing machines?
-Phillip
