Slashdot Mirror
Robot Warriors Will Get a Guide To Ethics
thinker sends in an MSNBC report on the development of ethical guidelines for battlefield robots. The article notes that such robots won't go autonomous for a while yet, and that the guidelines are being drawn up for relatively uncomplicated situations — such as a war zone from which all non-combatents have already fled, so that anybody who shoots at you is a legitimate target. "Smart missiles, rolling robots, and flying drones currently controlled by humans, are being used on the battlefield more every day. But what happens when humans are taken out of the loop, and robots are left to make decisions, like who to kill or what to bomb, on their own? Ronald Arkin, a professor of computer science at Georgia Tech, is in the first stages of developing an 'ethical governor,' a package of software and hardware that tells robots when and what to fire. His book on the subject, Governing Lethal Behavior in Autonomous Robots, comes out this month."
Been there, wrote fiction about that(much of which was about how, even in fiction land, it wouldn't work so well).
Weird. So this fails the Asimov criteria.
More importantly, would also necessarily fail the Golden Rule and Kant's Categorical Imperative.
If this is ethics, its a pretty limited version of it, and to be honest sounds more like rules of engagement than actual ethics.
Not to mention... some of the assumptions aren't great. As the article itself points out, it's been a long time since there was a civilian-free battlefield.
As for the direct example of the robot locating a sniper and being offered the choice of a grenade launcher and rifle - how does the robot know that the buildings surrounding it aren't military targets? How do they get classified? How does a hut differ from a mosque, and how does a hut differ from some elaborate sniper cover?
I don't think this is going to work out as planned.
Those who can, do. Those who can't, sue.
There is no such thing as a smart missile unless it immediately destroys itself safely.
NB: The message above might reflect my opinion right now, but not necessarily tomorrow or next year.
If you drop a fucking robot into a village where a vast majority of the people don't know how to read, what do you think they're going to do? They'll shoot at it, get the backs of their heads blown off, and then everyone will say, "Well, the dumbass shouldn't have shot at the robot!"
If this war on terror is so important, sign up. If you can't, get your brother or sister or even better, sign your kids up. If they're not of age yet, they'd better be in the JROTC. Then you can talk to me about how using drones and missiles isn't the dominion of motherfucking cowards. It's for freedom lovers defending freedom!
And if you think it isn't, imagine what the headlines would be if China landed a few thousand autonomous tanks and droids in Los Angeles. Oh, but that's right. This is about principles for others to follow, and for us to ignore.
It should never be legal for a robot to "decide" to take lethal action.... Ever.
"It's because they're stupid, that's why. That's why everybody does everything." -Homer Simpson
Why is this a when question, rather than an if question?
Humans aren't actually better at it than robots; humans are notoriously bad at estimating conditional probabilities.
10 PRINT CHR$(205.5+RND(1)); : GOTO 10
In any war zone (regardless of who has fled and who hasn't), isn't anyone who shoots at you, defined as a combatant and a legitimate target?
"Believe me!" -- Donald Trump
Was this article an attempt to promote Terminator 4?
Ave Molech Setting
But (most) humans have this innate condition where taking another life weighs on them somewhat - even most veterans and soldiers I know get twitchy about having to shoot at another person. A robot removes this and replaces it with cold logic.
Put another way, replace the robots with the WOPR, and the humans with, well, the humans in the bunkers.
Me failed English...
FreeBSD over Linux. If my comments seem odd, this may explain...
What? this isn't true, there ahve been many battle fields where civilians aren't at.
The Kruger Dunning explains most post on
Yes, robots are much better at calculating probabilities; given a series of "facts" with a confidence level assigned to each one, a robot would make a better decision. What I should have said is that "Humans are better than robots at making decisions based on incomplete data." Humans can develop "intuition" and many have a great deal of experience in interpreting the context of the data. While it may be possible some day for robots to have a deeper understanding of context than humans, that day is still a long way off.
I've abandoned my search for truth; now I'm just looking for some useful delusions.
Humans are (generally) concerned about self-preservation. Wrongfully killing someone could get them in jail or executed. Robots, on the other hand, simply decide based on some algorithm and have no concern about the effects of their actions. While you could try to boil down the soldier's logic to an algorithm, the key difference you can't resolve is that the soldier has free will, while the robot has no real choice of its own.
Another thing that's nice about restricting the ability to kill to humans is that a rogue soldier, no matter how well-trained, can be killed easily enough with the right application of force. We have no idea how advanced lethal robots could be. We don't have any reasonable guarantee that a rogue robot could be stopped.
My webcomic
Good points, but I don't think this is about robotic soldiers lumbering over battlefields just yet. I think this will, at first, be more about semi-automated fire control systems and drones. Like a future Predator drone might decide to wait to fire its Hellfire missile if it thinks there's too many civilians in the area and the projected accuracy is too low due to interference. Or a point-defense system might see a kid walking around in a field and decide that he's not a threat, because he's not carrying any weapons or moving in a threatening manner.
Since our drones are still somewhat dumb, most of the ethical considerations are the responsibility of the programmers and project commanders. For example, that drone might be programmed to distinguish straight dusty road with no other cars or civilians around from twisty road in the middle of downtown with lots of civilians walking around and a poor chance to hit the target.
Besides, if a robotic soldier were pointing a gun at you and demanding that you surrender, it would probably be tracking you with multiple sensors and would blow your face off as soon as your finger twitched in the direction of your gun.
Job? I don't have time to get a job! Who will sit around and bitch about being broke and unemployed then?
what is the fallback mode / data link lost?
crush kill destroy?
And yet, is it fundamentally a bad thing? We give less-than-stable humans that responsibility all the time.
Yes it is fundamentally a very bad thing. First instead of being limited to one trigger that unstable human can now pull hundreds of triggers simultaneously. The robot will never question his orders it will simply comply no matter how morally questionable the order is.
Secondly the one big way in which democracy helps maintain peace is that the people who will do the dieing in any conflict are the ones who also effectively control the government through their votes. If suddenly Western democracies can send robots in then they are far more likely to go to war in the first place which is never a good thing.
Who said enemy casualties are a bad thing?
Circumcision is child abuse.
These are military robots. No military robots would fall under Asimov's list.
What I think some fail to remember is that Asimov was just a science fiction author. He wrote stories. Very compelling ones, his place in modern literature is gigantic, but none the less just fictional stories. Thus his "three laws" have nothing to do with reality. They aren't natural laws, or legal standards, they are just part of a story. Thus they have no standing in the world.
They may well be how Asimov would like to see robots work, they may well be how you'd like to see robots work, however they have nothing to do with how the military wants it to work. They are not a canon of any kind.
When a robot is developed for military purposes, it should be no surprise the ethics are considered in that context. The whole point of it will be to be able to use deadly force if necessary. The programming is then when is that ok and not ok.
So please, let's have all us geeks lay off the Asimov "three laws" when it comes to robots. Every time something like this comes up people start talking about that like it matters to anyone. No, it really doesn't.
Your post has absolutely zero factual basis to it. Physical separation is a major psychological factor when deciding to 'pull the trigger.' Try reading "On Killing" by Dave Grossman, an excellent book that points out the reasons why distance makes it easier to kill.
Portland, North Dakota Puppies
Obviously you've never spoken to a tank commander, or any manufacturer of the UI's inside of armored vehicles. They are designed to be 'like video games' for a reason. Specifically to dehumanize the opponent, and mitigate the likelihood that you will associate your actual actions with killing. That's basic psychology. It's also why we refer to the enemy in Iraq/Afghanistan as Haji, why we called the Germans Jerries, and why we called the VC Charlie. You don't hate Ho Ming Na, father of 4 children who were brutally slain by US soldiers and is trying to simply save his farmland. You hate Charlie, so killing Ho Ming Na is acceptable. Anything to dehumanize them is crucial for removing mental blocks to soldiers.
By your logic, shooting someone at point-blank range would be significantly more difficult than shooting them from 200 yards away, which would be more difficult than shooting them with battlfield artilary from 1 mile away[...]
Correct, if we're talking about killing the same 1 target. Stabbing someone to death has to be far more difficult than watching a special ops team on a monitor halfway around the world.
The logic doesn't follow, because as you move farther away and impact more people, the decision becomes more and more difficult.
You introduced a second variable here besides distance... "more people". We don't drop nukes because we know that every time we do that we kill tens or hundreds of thousands of people. People we don't really want to kill are going to die. Lots of them. On the other hand, we frequently send cruise missiles and drop smartbombs because we can kill with them far more... discriminately, than with mirv nukes on an icbm.
For some reason, you are assuming that physical separation suddenly turns people into sociopaths.
Nobody said that. You've extrapolated too far all on your own.
I don't see a technical reason why a robot couldn't get that, too. It would be just a negative score for any killed human, which would enter the equation when making the decision.
The Tao of math: The numbers you can count are not the real numbers.
For some reason, you are assuming that physical separation suddenly turns people into sociopaths.
Well, yeah, because thats proven. Remember the Milgram experiment?
Imagine that, say, China attacks USA using robots.
Still thinking that excessive casualties are OK?
General David Petraeus. See here.
This sig all sigs devours