Slashdot Mirror
U.S. Army Robots Break Asimov's First Law
buanzo writes "The US Army is deploying armed robots in Iraq that are capable of breaking Asmov's first law that they should not harm a human.
SWORDS (Special Weapons Observation Reconnaissance Detection Systems) robots are equipped with either the M249, machine gun which fires 5.56-millimeter rounds at 750 rounds per minute or the M240, which fires 7.62-millimeter rounds at up to 1,000 per minute.
" update this story refers to this article from 2005. But com'on, robots with machine guns! I don't get to think about that most days!
From TFA:
They are still connected by radio to a human operator who verifies that a suitable target is within sight and orders it to fire.
While they are harming a human, it's ultimately a human that makes the decision to fire. And who cares about fictional "laws", anyway?
I guess it depends what you consider to be a robot? And under what conditions it could kill another human? The Phalanx defense system, currentlly employed on U.S. Warships, would allow itself to shoot down an enemy aircraft if it were attempting to crash into the ship. The Phalanx uses radar to detect incoming missiles and shoot them out of the sky by unleashing an insane amount of bullets in direction of the target. Pictures and info here. -C
These are actually robots, but they're not the fully-autonomous solutions that Asimov was suggesting that mankind needed protection from. Thus the "laws" of robotics don't apply here, because it's still a human who's doing the thinking for the machine.
In effect, this is a safe way for ground troops to line up a kill zone, then cause lots 'o bad guys to get torn to shreds. Prior to this, troops needed to use a vehicle-mounted machine gun to get this sort of rate of fire. This was extremely limited in close quarters, where a Humvee or Tank might not fit. While it was theoretically possible to carry a machine gun to the combat zone, such weapons are difficult to transport, setup, and use in close quarters.
Javascript + Nintendo DSi = DSiCade
...is no "law" at all.
If the submitter wants to troll about the military, the least he could do is spell Asimov's name correctly.
What makes a "robot"? Progressively more complex machinery has been able to inflict bodily harm, and kill, for quite some time.
THE US Army is deploying armed robots in Iraq that are capable of breaking Asmov's first law that they should not harm a human.
Sorry to break it to the folks over at the Inquirer, but Asimov's Laws do not actually exist....any more than his 'positronic brain' does. It's fiction.
Next week on the Inquirer: Computers Built That Break The Orange Catholic Bible's Commandment of 'Thou shalt not make a machine in the likenes of a human mind'.
Sheesh.
They are still connected by radio to a human operator who verifies that a suitable target is within sight and orders it to fire.
OK....so the're not even robots, then. They're telepresence devices.
Then the robot has the job of making sure lots of bullets are sent towards the target.
Statement from the Iraqi forces regarding the use of these 'robots':
Nice to know we can take what we've learned in FPSs and apply them to the real world.
Later the US plans to replace the control system of the bots with a "Gameboy" type of controller hooked up to virtual reality goggles.
Yes! Finally, all my training has paid off! I can be a soldier from the comfort of my basement! Where do I sign?
____
~ |rip/\/\aster /\/\onkey
These robots will have a pre-set kill limit.
The enemy must merely send wave after wave of men until that limit is reached and they will shut down.
Comment removed based on user account deletion
Oh no! Robots are breaking fictitious laws!!! Someone call the Fiction Police!
This guy's the limit!
. . . a place where Asimov's Laws, like the US Constitution or the Geneva conventions, don't really apply.
The very idea of a rule against hurting humans implies that a robot knows:
1. What hurting means
is it pain? death? financial impact? what about indirect effects? If I help human 1 build a better mousetrap, I am indirectly harming some other human's way of life.
2. What people are
3. Where they are
These are highly non trivial problems. In fact, they're unsolvable to any degree of certainty. They only make sense in a *science fiction* book in which a highly talented author is telling you a story. In the real world, they are meaningless because of their computational intractibility.
In the real world, we use codes of ethics and/or morality. Such codes recognize the fact that there are no absolutes and sometimes making a decision that will ultimately cause harm to someone is inevitable.
So can we please stop with these damned laws already?
So what this really shows us is that the winner of future wars will be determined by the country who has the most skilled gamers. I think I like the direction things are headed. Let's be sure to stay friends with the Japanese tho.
I firmly believe in the right to bear arms -- all arms.
Sorry, have to differ with you there. I don't want a tac nuke in private hands, because I don't believe you're capable of only hitting those who are actually posing a threat to you personally. I also wouldn't let you have land mines, pursuant to the common law principle of prohibiting reckless endangerment.
-jcr
The only title of honor that a tyrant can grant is "Enemy of the State."
Many researchers are spending lots of time researching AI, and the problems for which the Laws of Robotics are a an attempted solution; Namely how do you keep the robotis from taking over and/or indiscriminately killing mere humans, as seen in so many hollywood movies. So fictional laws are important as experiments in looking at potential solutions to a real problem.
As I see it, the main problem consists of two factors. One factor develops as a result of the first.
The first factor is consciousness, also known as self awareness. The second factor sounds like it is the first, but it includes other areas.
The second factor is Identity. Identity is not restricted to Self Awareness, but also includes group awareness, etc in expanding circles to include universes, subjective and otherwise. When someone else is considered part of a group identity, as "one of us", then you tend not to act against yourself. When the other person is seen as being "one of the Not Us but Them" then you tend to get an opposition, etc.
In wars, it is more a universe thing, the Hitler Universe vs the Churchill Universe, for example. Or Religious Figure One (tm) vs Religious Figure Two (tm). Or a religious universe vs a scientific universe.
Part of the problem of psychopaths, sociopaths, etc. is that they tend to group their victims into the "One of the Not Us/Not Me" category. No sense of being or identity is allowed or granted to the other person, and so, to one degree or another, this rationalizes pigeon-holing people into things that can be abused one way or another. Or else the identity given is some other alteration of reality that legitimizes criminal activity.
This is difficult enough to deal with in humans. Psychologists and psychiatrists have no cure for psychopaths, since it is seen as being in the genes. You can't make a pill for it, and no psychopath would take it as they do not have the luxury of seeing that anything is wrong with themselves.
Now we try to apply this to Robotics. Probably the only real solution for the problem is to redefine Human as self aware creatures from earth, and incorporate this awareness somehow into robots, to some slight degree, so that Robots see Humans as "One of Us".
It is a little touchy on how you would do this. It exposes some of the potential hypocrisy of humans in actions towards other potentially self aware creatures on earth, as well as each other. A self aware robot could see the hypocrisy without the emotional justification people exhibit. At this point, we could be in trouble.
"It is a greater offense to steal men's labor, than their clothes"
First: two observations:
1) SWORD is remote controlled it is not autonomous like I always thought a true robot in the Asimovian sense had to be.
2) Since we are now including remotely operated vehicles in the definition of a true robot, SWORD is not that different from a Paveway bomb or a Hellfire missile except SWORD doesn't self destruct when it destroys the target.
This begs the question wasn't Asimov's first law broken decades ago, perhaps even by the V1 which was strictly speaking a remote operated vehicle?
Personally I won't begin to worry about Asimovs laws as long as Humans are on the other end. apons.
Only to idiots, are orders laws.
-- Henning von Tresckow