Military Robots

[Toucan]

I just read an interesting article on military robots. Here (http://www.news.com.au/technology/story/0,25642,23284564-5014239,00.html)

Now that military robots are in use by many nations around the world it does indeed raise many interesting questions.
At present all military robots do not have the ability to make a decision to terminate a human life, although they can acquire and lock a target autonomously they cannot fire unless a human presses the fire button, but this may be about to change.
New designs will have the ability to decide for themselves whether or not a target should be terminated. It raises many interesting questions.

Do we have a right to pass to a computer or robot the decision to terminate human life?

Does that computer or robot have the right to kill people in self defense?

Does a machine (even an AI machine) have the right to self defense at all?

What will it do to the worlds political climate if countries gain the ability to attack each other without immediate risk to their own people?

[Borgorb]

This is a serious violation fo the laws of robotics
if u dont know them lemme refresh ur memory, not i have changed human to sentient being for (possible) integalactic encounters
1. A robot may not injure a sentient being being or, through inaction, allow a sentient being being to come to harm.
2. A robot must obey orders given to it by sentient beings beings, except where such orders would conflict with the First Law.
3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

Such blatent disregard for these laws will result in AI removing all threats like their military advisors

War is a brutal deisregard for life and has no place in society, although conflict is executed on a daily basis it shouold never escalate to the number of deaths taht have been recorded in 10 years compiled into one month

If robots govern this who is to say it wont turn out like I, Robot (the original)
for those who havent seen it the AI becomes so concerned for human preservance it makes the humans unable to do anything.

These robots shoud be discontinued and dismantled immediately. If they were made to the laws war for them would be impossible because they would have to listen to "the enemy" as they would be a sentient being also.

[femoimal]

[...]
1-Do we have a right to pass to a computer or robot the decision to terminate human life?
2-Does that computer or robot have the right to kill people in self defense?
3-Does a machine (even an AI machine) have the right to self defense at all?
4-What will it do to the worlds political climate if countries gain the ability to attack each other without immediate risk to their own people?

(I added the numbers)

1- Sure, human life is all relative. Its in order to save one's army personnel that you use machines to kill the other. A yellow/muslim/protestant/heathen/Tutsi/capitalist/communist's life is of less value than the one's army on the field. There is not so much a question of rights on the battlefield, but rather a question of military efficiency and political cost. Ask Hiroshima survivors.

2- Again, its me or you on the battlefield. The issue here is identifying combatants and civilians. Its already difficult for people, its immensely hard for machines. Fielding machines on a front will surely increase a lot that lovely euphemism called "collateral damage".

3- Machines have the right to protect their 10/50/250 million dollar asses as well as their 800 million R&D official budget. Compare that with 5 cent apiece for the life of an "irregular combatant", and you got your result.

4- that already happens, with air superiority and bombs. You can reduce to rubble a country by the air, but not submit its people. The zero-casualty conflict is a wet-dream of the past, a post-vietnam dream caressed by some morons like Rumsfeld and Cheney. That dream was shattered by Iraq. Training a marine is still extremely expensive, but machines are even more so, for the time being. I do not see many conflicts happening with no casualties in the future as big cold-war armies did not clash on large battlefields. The use of robots in guerilla wars is more about intelligence, c4i and possible surgical strikes and assassinations.
Say though that one day, only robots or remotely controlled automata will roam the battlefield. I doubt it would change a lot the outcome of conflicts. Thats why: the propaganda machines are well oiled and use patriotism and information control to give a positive view of war. See now that its not so much the question of deaths that the presidential candidates use for Iraq, but the cost in dollars of the war. And machines are more expensive than men. So, we can see a shift from human cost to economical cost in conflict, even in the dove left-wing circles. What north-american media gives a flying fuck about the 100.000 iraqi dead? none. Its the pride and the wallet that's hurt. So i think war will be the same. We fight for dollars, with dollars.


Ground fighting machines, unlike space/air/water/underwater ones, have to deal with very complex environments and so will take much longer to develop. The first robots will so more likely provide support to the ground troops. The ones tested right now are more like Med-evac, de-mining or logistical support.

[singo]

Do we have a right to pass to a computer or robot the decision to terminate human life?

Does that computer or robot have the right to kill people in self defense?

Does a machine (even an AI machine) have the right to self defense at all?


1) Is that any worse than humans having the right to terminate life off their own bat?

2) Yes. Really if something can protect itself it should. ESPECIALLY if it has intelligence.

3) If it is intelligent then it should have the same rights as humans. Anything else is slavery.


What will it do to the worlds political climate if countries gain the ability to attack each other without immediate risk to their own people?

Not sure I want to speculate really...

This is a serious violation fo the laws of robotics


Those "laws" were made up by mr Asimov to make a good story, not for any other reason.


Fielding machines on a front will surely increase a lot that lovely euphemism called "collateral damage".

I would say it would REDUCE collateral damage, whilst people can panic and fire at something that moves suddenly, a robot would presumeably have half-decent pattern-recognition so instead of firing at something that looks a bit suspicious it will only fire if the something is a bloke holding a gun. Robots are less fragile than humans, so the split second that would be vital in a human-human encounter can be taken to analyse the situation.

[mranderson]

1)

I would say it would REDUCE collateral damage, whilst people can panic and fire at something that moves suddenly, a robot would presumeably have half-decent pattern-recognition so instead of firing at something that looks a bit suspicious it will only fire if the something is a bloke holding a gun. Robots are less fragile than humans, so the split second that would be vital in a human-human encounter can be taken to analyse the situation.

Perhaps you are thinking of smart AI. Plus I doubt that the robots make decisions. It is guaranteed that if someone walks around with a corner with an AK it will shoot. But isn't it hard for a robot to distinguish between an AK, and maybe a metal pipe that happens to be shaped like a gun? And don't forget about if someone walks up to it with an axe, or disguised gun, or explosive.


1-Do we have a right to pass to a computer or robot the decision to terminate human life?
2-Does that computer or robot have the right to kill people in self defense?
3-Does a machine (even an AI machine) have the right to self defense at all?
4-What will it do to the worlds political climate if countries gain the ability to attack each other without immediate risk to their own people?

1. Until a computer can say yes or no based on its opinion, and views without factoring in anything intellectual it isn't a decision. It is the end result of its programming. So I would say passing this "right" on would be a good thing, but that it could be abused. Say deploying them in a city with people versus a combat zone.

2/3. Yes it does. If someone decides to attack a multi-million dollar defense robot that dual wields fifty cals then they have made their decision. It is out of the robot's "hands" when they shoot at it.

4. The same thing the nuke does, but without the fall out. If countries start a race in robots they'll keep building them, (robots building robots...) and find ways for a massive depolyment. Countries that won't develop them will either unite with the others or get walked all over on. Basically it gives more power to the people that already have it.

4a. What happens if these robots aren't classified as troops? And the president (or whoever they give the power to) can deploy them wherever he wants for as long as their battery, and ammo last...

Those "laws" were made up by mr Asimov to make a good story, not for any other rea

These laws were used my mr Asimov to display a higher meaning. That is what possibly makes it a good book, not a story. I've never read it personally, but if it tells a story like Lord of the Flies than it is a good book.

[femoimal]

Mister Singo, i like a technological optimist when i see one :D

Its hard enough for robots to drive around in the desert without stumbling after 100 yards. They can barely see, and cannot walk without tripping. In a combat zone, you got confusion by the ton: rubble, dust, smoke, civilians (by mistake or intentionally). Pression shockwaves by explosions, screams, derelict structures and ground churned up by explosions and artillery. Its hard enough to walk there if you are human and agile. And we have millions of years of evolution behind us.

Robotic sensors are waaaay below par when it comes to having a quick assessment of human behavior. We can spot a small change in the walking pattern of a human being, a suspicious move or even a facial expression. Machines can scan you faster in ideal lab conditions, say. But in the battlefield ? with dust particles, reflections, hostile lighting ? People get camouflaged in battlefields: branches, color patterns, fox-holes...The owners of machines will deem it safe to keep on the trigger-happy stance. Who wants to loose a 20 million clockwork piece to a 2 dollar grenade ?

Machines have 2 more likely roles. Intelligence and support.

Small micro (and then nano) machines, very much like insects, will be able to fly and increase battlefield awareness in the "house-to-house" battle. Small flies with electronic eyes and insect-like wings will hover and perch, transmitting information bursts for small distances. That will totally change the battlefield: anything flying or crawling can be a spy ! Small portable lidars coupled to small automatic lasers could get rid of small flying buggers like that, but nothing in a dusty environment would be efficient.

Support robots will be like mobile tanks. Legged or wheeled, they will be SUV-sized, to be able to follow the troops in any terrain. Large sensor arrays, immediate response weapons (chain MG, or lasers), good comms and a good battery array will make them useful as an assault support roles, especially by night. But on their own, they are quite vulnerable.

And, as a small side-note: machines are only as good as their programming is. Remember the iranian Airbus shot down by the USS vincennes in late eighties ? Well, the cruiser is an AEGIS ship, bristling with weapons. Bad thing is that its software got confused when 2 radar trails intersected: it didn't know which aircraft was which. So put microsoft in a terminator, and watch it reboot every 10 yards :)

Machines are not given rights anyway, their programmers do. If a robot is conscious and has rights, i doubt it would be very much interested in slicing iraqis to pieces. It would rather compose a symphony or dedicate itself to nature preservation.

[Neo]

Not to break it to you or anything but:

Most robotic or automated things used by the military/armies around the world, don't usually include an extensive AI. As such the "laws" don't really count for much.

Also, did you see or read I, Robot? It kind of through laws right out the window =P Besides that, The laws of robotics are fiction. If we were truly programming AI that had weapons, then the smartest thing to do would be to have built-in mechanical-based switches that could shut it off at a moments notice, or hard-coded (ROM-Type) programming that it couldn't "get around" or "reprogram"

The whole "OH MY GOD ROBOTS!?" sentiment really comes from Terminator-Like doomsday movies and stories, and less from reality.

To be honest, what I find much more interesting is nanotechnology. WIth Nanotech, it basically means everything is obsolete. We'd have no need for doctors at all, because we'd each have a nanopackage that would scrub our bodies and keep us healthy -- also effectively increasing our life span.

We'd have no need for builders, nanotech could take some raw material and build it. We'd never be hungry, same deal.

We could even use nanotech to enhance ourselves.

The major drawback is that if the nanobots aren't programmed well enough, we might see something like the grey goo affect, which means we're all dead. But seriously, one would hope the people behind nanotech aren't being stupid.

Hell, with nanotech wars becomes pointless. If you can send out a nanotech package and simply disintegrate any ammunition/explosives...

-Neo

[singo]

Mister Singo, i like a technological optimist when i see one

Heh, studying history does that, I LIKE the long term :P


Its hard enough for robots to drive around in the desert without stumbling after 100 yards. They can barely see, and cannot walk without tripping. In a combat zone, you got confusion by the ton: rubble, dust, smoke, civilians (by mistake or intentionally). Pression shockwaves by explosions, screams, derelict structures and ground churned up by explosions and artillery. Its hard enough to walk there if you are human and agile. And we have millions of years of evolution behind us.

Oh yes, as of NOW they are a bit crap, but give it ten, fifteen years time and they will be shit hot.



Robotic sensors are waaaay below par when it comes to having a quick assessment of human behavior. We can spot a small change in the walking pattern of a human being, a suspicious move or even a facial expression.

I was using the example of a battlefield where the only human soldiers are the enemy. Thus, if it is carrying a gun and has two legs, its fair game. I would hope that with the amount being spent on them the things can at least recognise limbs and assault rifles.


If a robot is conscious and has rights, i doubt it would be very much interested in slicing iraqis to pieces. It would rather compose a symphony or dedicate itself to nature preservation.

Why? Humans dont.


http://www.foster-miller.com/literature/documents/DS08-006-MAARS-Robot.pdf

In all honesty, my argument does boil down to "I cant help being in favour of something that looks as amazing as that".

[femoimal]

...which is the best argument ever ;), and the one that made me join a Defense Minister think-tank too. I prepped a massive file on unmanned vehicles (did you know the first one was by Tesla?), that sadly never made it through the editorial pipeline.

10 or 15 years is a bit optimistic, for battlefield-environment sensor management. I reckon oceanic, aerial and open-country (say desert) theatres will see the most likely emergence of autonomous AI warriors. (especially the underwater one, where its not possible to issue or transmit orders because of the water). Urban environment is last.

We have been overly optimistic about sensory data integration and decision-making in robots, and for a long time. We can attain the hypothetical IQ of a bee, in programming, and are reaching currently the lizard's level. But their sensor data are way better managed than ours. Perhaps neural networks self-training programs are a solution.

... or many little robots using a gestalt AI.


...nanos are way more scary indeed. IT could be like targeted chemical/bio warfare, I suppose. I think the future of warfare is small: imagine a small fly attaching to your neck, or arm, and injecting you a neurotoxin :)

[Toucan]

I have actually been very surprised by everyone's responses in this thread.

The problem that I see with machines like these is there no morality or emotion with them. A few individuals could potential decide the fate of millions. Possibly even with few in the world knowing what happened.
With a large number of human soldiers there really isn't any way to hide what is going on, every one of the soldiers families knows they have been deployed somewhere. And war remains a terrible thing to avoid, rather than creating robots that will give those that wield them the belief that they can commit acts of war without consequences.

Personally I believe the international laws that oppose the landmine should be modified to oppose these machines as well.

[singo]

Emotion in a combat zone can be just as bad as its lack, example, the treatment of prisoners. If a person has just seen their mates get dismembered by machinegun fire by the people who he now has at gunpoint, their sense of reason may well be out of the window.

Remember how often "shot whilst trying to escape" has been used.



And there is no chance of few knowing what happened. Someone needs to fly the robots out to the combat zone and switch them on, even if they are practically autonomous from there. And the OTHER side will certainly know.


As for the international laws opposing landmines. It isnt that landmines are not a valid weapon of war, they are banned because they lie around armed and virtually undectetable for decades and blow legs off twenty years after the war has finished.

[Toucan]

No robot can conscientiously object to its orders. This final fail safe to ensure actions to be taken are warranted or legal should not be removed.

Robotic planes already exist. Building larger ones, is just a matter of building larger ones.

And although the other side may know, if a large robotic force capable of killing with machine efficiency is already there, its to late.

These machine will still be around 20, 50 or perhaps even 100 years after they where constructed. Keep them well maintained and they will still be in perfect working order when they become obsolete. Their not going to retire, they will remain a security issue from the time they are constructed till the time they are destroyed.

[singo]

When was the last time a bit of machinery lasted 100 years in perfect working order?

Now consider that their job involves getting shot at and blown up.


I really think everyone here has been watching Terminator and the Matrix a bit too much, they do not have a working AI, and can ONLY follow their orders. Just because they could be misused is no reason to object to the technology itself.


Lets wait and see what happens instead of objecting on the grounds of what Sci-fi novels and films have told us.

[Toucan]

Any machine will last if its maintained. Thats why there are 100 year old cars that still run.

It's not there job to be a target, its there job to destroy targets. I don't see why an engagement would bring about more inoperative robots then fallen soldiers.
Most soldiers in the western worlds armies don't get killed, most soldiers will live on to do some thing else.

And your right, these robots are not that smart, there is no such thing as a truly intelligent AI yet and we all know it. And we are going to trust these things to decide what targets should be destroyed?

[Tech]

The way I picture military robots in the near future is as security and defensive machines only. Femoimal is right about sensors and the complexities of moving around a modern battlefield, it'll be a long while before we can come up with a robot that can not only efficiently move through and active battleground, but also make decisions how to kill the enemy, and who they are.

Yet the purpose of replacing humans with robots is to save human lives, and currently the most lives are lost on routine patrols by IEDs and at checkpoints blown up by suicide or car bombs. So the most obvious roles for the first generation of combat robots would be in those areas. So a robot may not be able to sprint from cover to cover in urban combat, it can definitely stand on a street corner or at a checkpoint scanning for weapons and explosives. Seems like simple enough programing to me, if the robot is shot at or if someone runs a roadblock then it has authorization to kill. If explosives or weapons are detected then it would have to request human authorization on how to proceed. Obviously nothing in robotics is simple, but in comparison to other tasks that human soldiers would do these would be the most likely to be replaced by robots.

Then as far as robots being a threat until they're destroyed, that's obviously not true. What danger is a robot that's out of ammo, or who's battery is dead? By the time we figure out a way to put weapons on a robot with infinite ammo and a mobile power source that never runs out, I'm pretty sure we'll have the technology to control the things properly. Heck, by that point we will have probably achieved world peace (or destruction).

[Toucan]

Then as far as robots being a threat until they're destroyed, that's obviously not true. What danger is a robot that's out of ammo, or who's battery is dead?
It would be extremely dangerous if an enemy was able to capture the battery dead or currently unarmed robot. Batteries can be recharged, weapons can be reloaded or replaced.

The idea of saving human life by creating machines to kill more efficiently is crazy. I can see a real potential for escalating warfare in our world with these machines. Not to mention much more desperate measures being taken by those governments that do not have matching technology. (Iran for example)

Robotic mobility is increasing in leaps and bounds, there are bipedal robots today, 10 years ago such a thing was impossible. The current designs are built like small tanks, caterpillar tracks, heavy armor, a lot of weight.
They would have no problem on any desert battle field.

No weapon of war's creation has ever decreased death. They are designed to kill, period. Although I don't think there has been a weapon invented that its inventers didn't claim it would reduce mortality rates on the battlefield.

[singo]

It would be extremely dangerous if an enemy was able to capture the battery dead or currently unarmed robot. Batteries can be recharged, weapons can be reloaded or replaced.

Yes, but as long as the owning party keep track of where it machines are, they would know about it, and then they can shoot any robot that is where no robot should be.


I can see a real potential for escalating warfare in our world with these machines.

No more than any other new military application of technology.



No weapon of war's creation has ever decreased death. They are designed to kill, period. Although I don't think there has been a weapon invented that its inventers didn't claim it would reduce mortality rates on the battlefield.

They decrease FREINDLY death rates.

[Neo]

Toucan you seem to be fixated on this terminator-esque idea that these machines will all have a high level of AI and be completely logical and cold, thus turning on their creators and/or humans totally.

It sounds almost like a phobia to be honest.

Someone could "capture" one, but without knowing the technology underneath it won't do a damned bit of different -- conceivably a tribe in the amazon could "capture" a plane or something, that doesn't mean they could use it.

It takes time to reverse-engineer, and it's unlikely you'd just sit by while you know one of the machines is missing.

The different here to is that most things you mention, from my understanding, are simply remote controlled -- they aren't actually automatic, nor do they have an AI (stupid or not). So there is still someone "piloting" them, a human, in the background.

Now granted their are Computer-Controlled weapons, but again, we aren't talking about AI, we are talking about a computer controlling a missile so that it reaches it's actual target -- precision. Not something like a human who might miss the target, or a pilot who might miss that military installation and instead hit a grocery or school or something.

Yes AI way machines are scary, but I strongly doubt they will overthrow us and/or cause us to be more lax about war or something. Anything that can save lives is a good thing. For our side, our theres. Ideally a robot, if it was ever automatic and not RC, would have a directive, for instance, to only target people with weapons. Ideally it would also be better at that then normal troops, especially in a heavily populated area, for instance, Iraq. If a troop see a large group of people and a couple off to the have guns and are trying to shoot him, there's probably very little time to think "oh well, those others are probably innocent" (and/or troops ARE human, and get fatigued, or emotional - "Sorry sir, but I thought they were the enemy" or "They are all the enemy, every last one of them!" or other sentiments).

With an AI controlled (Even stupid ai) this becomes moot. An AI won't randomly suffer emotional break downs, or gun down civilians because they couldn't tell the difference.

Still, this is all moot. The future is nanotech, not military robotics. Seriously, you could send out a nanotech package and simply dissolve ammo, disarming your enemy without ever firing a shot.

-Neo

[femoimal]

good points and interesting ideas thrown all around. I'm glad this discussion is that dynamic :D

Oh we are certainly influenced by 2001 and terminator. I personally love robbie the robot. We persistently overestimate their abilities, so far, the only operational robots are mere toys, flying in the air alone (wowww groovy, it can fly on its own ! autopilots do that everyday), roaming at amazing (1mph) speed through the (flat) battlefield, mostly telecontroled. Its quite pathetic, robotics-wise: most of the trajectories are pre-programed,and machine decision making would not really shame a shrimp.

That's that: modern military robots are merely toys. From small ones to large ones, the latter being some very expensive autonomous platforms that exist solely because the cost of latrines on-board is too high.

Why do robots exist on the battlefield ? Well, first we gotta ask ourselves what the other weapons are.

Consider a weapon.
This weapon is fragile, naturally prone to breakdown, takes about two thirds of its useful life for maintenance, takes years to manufacture and lasts only about 2 decades, at most. Its non-standardized, may refuse to function without notice and might even turn upon its user. Its used in millions everywhere, but the loss of a single of these weapons can cause complete political turmoil. A capture of a single dozen of those weapons can paralyze or cause an entire conflict.
What general in is right mind would choose to field such a weapon ? Well they all do, because that weapon is man.

Man is full of shortcomings. It can adapt and take decisions (wrong ones as well), and its cheap (especially if you enroll people from lower social strata). Nevertheless, it tires easily, its morale can break, its sensors are limited, it needs to sleep, to be trained, brainwashed, and only fights when between 18 and say 35. Its a very lousy weapons from many points of views.

The first way would be to improve man from within. The US army distributes speed pills and drugs heavily its air force pilots. Too bad it changes the behavior of those pilots. Speed pills and sleep pills (gogo-pills i think is the name) are just continuation of the WW2 german habit of drugging its grunts. It impairs judgment and lowers the life of the user. You could use implants and such, but the cost here rises a lot, and you've got the issue of retiring those augmented people afterwards (plus the very likely accidents).

You can choose to improve man from outside. Stick him in an exoskeleton. DARPA i think is working on it. Because you have to know that in Afghanistan, the guys are carrying routinely 70 kilos on their backpacks, more than 15 being batteries if i remember well. 70 kilos, in the mountainous regions, under-fire, day and night. Good luck man. The exoskeleton will require work, especially in the battery department. It will increase the cost of the trooper but lower the general fatigue in operations. Plus it could integrate a better comms/sensory suite and greatly help integrate the soldier in the virtual battlefield.

You can choose to assist man. Robots can provide intelligence (very dangerous missions for humans), medical evacuation, sentry and picket missions. Robots are small, don't sleep and can switch off to save energy. Some could even use sunlight as a source. It's in that category that today's robots exist. They fly further, for a longer time because they don't have to cater with human fatigue. An AWACS could fly several days on end. Its currently limited by the capacity of its oil tanks, and by the crew collapsing on its consoles. Robots operate longer and at full capacity. Deprive a man of sleep for 36 hours and he sees Holy Mary talking to him.

You can replace man. Now that's something that's only working is space and under-water, for the moment. In air, it could come too in the next 2 decades. Assault robotic tanks are way off.

I don't think that, today, robots kill anything without the man pressing the button directly. Tomorrow, the button might be a software line and a rules of engagement. Bear in mind though that robots have serious limitations too. Battlefields eat machines for breakfast. Maintenance of even simple machines require armies of technicians. Robotic units would require ten-fold increase of support units, because of the complexity of machines. The more complex a machine is, the more likely it is to fail. Throw a stone at current robots, and you can hear the millions of dollars come crashing down. (this will change with time, though, especially if machines design machines).

I agree with Singo that automated surveillance by machines should decrease the number of casualties. As a matter of fact, robots would be the prime and more efficient users of non-lethal weaponry. Technology and armaments follow a continuous trend: substituting energy with information. Robots would lower even more the use of energy (explosives and such) with a more precise delivery.

I side with singo and want to be optimist. Until israeli micro-robots start beheading heads of Fatah leaders and robotic bulldozers burrow alive children and women.

PS: machines are much easier to fool than men. And those can really adapt fast to any machine: look at us defeating CPUs in starcraft. Hey we are hunter-gatherers for millenia.

[singo]

Throw a stone at current robots, and you can hear the millions of dollars come crashing down.

I am in complete agreement with everything except this bit.

http://www.foster-miller.com/lemming.htm

One was blown off the roof of a Humvee in Iraq while the Humvee was crossing a bridge over a river. TALON flew off the bridge and plunged into the river below. Soldiers later used its operator control unit to drive the robot back out of the river and up onto the bank so they could retrieve it.

If the buggers can be blasted off a bridge into a river and then DRIVE back out I reckon you would need one HELL of a big rock to stop one.

[mranderson]

TALON® military robots are powerful, durable, lightweight tracked vehicles that are widely used for explosive ordnance disposal (EOD), reconnaissance, communications, hazmat, security, defense and rescue. They have all-weather, day/night and amphibious capabilities and can navigate virtually any terrain.

It's amphibious, of course it came out of the river. So what would happen if it fell off the bridge, and oh say... slammed into the concrete, or rocks? I think it would break. You don't need a BIG HUGE ROCK. Stop following the propaganda. And I doubt it was "blown off." Took an indirect hit and went into the river.

[Toucan]

Toucan you seem to be fixated on this terminator-esque idea that these machines will all have a high level of AI and be completely logical and cold, thus turning on their creators and/or humans totally.
My first statement in opposition to these machine was that my major concern was "a few individuals could potentially decide the fate of millions." And I stated specifically that I think these machines will be to dumb to be trustworthy, how does that suggest I am expecting terminator-esque intelligence from them?
Not that there was about to be a war of man and machine. My concern is the political climate of real countries, not fictional forces.
And I'm not talking about amazonian tribes capturing them, a country like Iran capturing would be a serious concern.
And yes, all current robots on the battlefield are remote controlled. This debate is about the new designs.

I have never encountered a piece of machinery that would breakdown less than a human will become sick. Even technology as advanced as the space shuttle has been unable to see its career through without a couple of them just blowing up for (at the time) no apparent reason.

The robotic spyplanes currently deployed in Iraq cruise at 220 MPH, some what faster than 1 mph.

Man is full of shortcomings, does giving him another way to escalate warfare really seem like a great idea?

Current designs are able to roll through walls, take direct hits from explosives, lock targets autonomously (although they require human authorization to fire) and there are already armed designs in use of both the spyplanes and the ground based designs.

Anyone that is is going to site current robotic technology as a bases as to determine the future state of robotic technology is being foolish. Technology advances, technologies become intertwined. Why would the new designs not be armored with some thing as tough as the synthetic fibers that protect the abrams tank?

Now for the record, my concern with these machines not "OMG terminator is coming", its that these machines will make the world even more violent.

[Oblongato]

If we compare the level of directness of an attack on an enemy, there is a clear progression.

Most direct would be the fist, then the knife, the bullet, the bomb, the military robot, then the mine.

We are basically sending out hardware to do our will. When a bullet leaves the gun, we no longer control what it will do. Bombs, with their larger damage radius are even more imprecise, and we don't in most cases know exactly what the situation is where the bomb will fall, nor can we guarantee that it will fall where planned, though technology has improved tremendously in this area. Military robots could be programmed to recognize chips carried by friendlies, I suppose, making them more discriminating than mines, which are only geographically discriminating.

I think the comparison to mines was a good one. Mines will take out anyone who triggers them, and setting them is an indiscriminate attack with an indefinite duration. Bombs, it could be argued, are only more discriminating to a degree, as are bullets for that matter.

Ultimately, apart from concerns about the practical use of military robots, it seems to come down to our squeamishness about about the decision to take a human life being direct or indirect.

But here I think it is an illusion to believe that we did not long ago cross the line into indirectly taking human life. War has been by and large indiscriminate since the very beginning. Bombs set fire to cities and civilians die; bullets ricochet and kill the innocent. Indirect killing in some form has been with us for centuries.

This is where it becomes clear that war, in the end, is about survival. I do not believe that there is a country out there that would not completely abandon any concept of battle ethics when survival is at stake. Today, as combatants can no longer be divided into categories of "soldiers" and "women and children", our ethics will also have to be adapted. In the long run, it appears that being discriminating in wartime is an illusory ideal that will become increasingly difficult to pursue.

What the U.S. is even now asking of its soldiers is practically beyond human capacity: in a battle situation they are to decide who is a non-combatant and spare those people, a virtually impossible task. I predict that Iraq will be the grave of attempts at discrimination in a hostile environment and that our next war under such circumstances will not include attempts at interacting with the populace or nation building. Instead, we will simply incapacitate the nation as a whole from the air, hopefully reducing their military or terroristic potency. (No, I don't think that will work, either.)

[singo]

Well, the world has steadily been getting LESS violent for sixty years and, optimist that I am, I think that will continue even if robotic warfare takes off.

Even if terrorists with an unusual amount of technological savvy get hold of one and program it to...say... find a crowd of people and then get hugely indiscriminate, that is really just a variant of the suicidal maniac with exploding underwear. Only without the suicide element. Or the exploding. Or the underwear.



And as for entire countries getting hold of them. They will, its inevitable. Technology diffuses through the world like dye in water. BUT, just because a few of the....less savoury...nations of the world want them is no reason to put the entire concept under lock and key. Just because Iran want nukes is no reson to dismantle our own Nuclear Deterrents.

[Neo]

Toucan that's not logical at all. If we have the ability to send in machines to do the fighting, it wouldn't escalate the fighting, but rather snuff it out. Since machines (remote controlled or otherwise) could be manned 24/7 or never rest and such, there'd be no way that our troops would keep dieing in meaningless wars.

Yes it might make leaders quicker to launch an attack, but at the same time, it would also temper others responses. Isn't that what the cold war was about? We've all got nukes, but no one used them for fear of retaliation? If the major powers on the planet all had armies of machines/robots to do their fighting, I would imagine fighting would become pointless.

I don't know. I suppose in one sense you're right, but don't we risk this anytime a better way to kill stuff is created or improved?

-Neo