According to this NYTimes article, the military is planning on spending "tens of billions" of dollars on robots that can kill people. Currently, they are remote-controlled, but the military hopes that within a decade, the robots will be able to think for themselves (obviously without all the hang-ups humans have about killing other members of their genetic pool).

I'm working on background for a story that may include these (not in the actual story... only in the background) so I ask everyone here interested in responding, what do you think?

I can see them being used to deny territory to an opponent, but so can minefields, and mines continue to be cheaper to produce and distribute. If I were the commander of a mixed forces group, I would use them to prepare the battlefield in my favor, and to set up killing fields.

1. Where's the shut-off switch?
2. Which people do the robots kill and how do they identify them?
3. Where's the shut-off switch?
4. Do they understand or respond to surrender?
5. Where's the shut-off switch?

ON KILLING
by David Grossman

The author isn't exactly the most polished wordsmith, but the info in this book is very interesting.

Robotized convoys -- sure; still need some work, as the Times article pointed out, but probably doable within the next few decades (assuming that your convoy never has to deal with enemy fire, anyhow). Robotized air power isn't far off. It's already going up for surveillance. Most airial combat today depends upon the weaponry, with the airplane simply a platform for launch, and the pilot a guide to get it there, anyhow. The guy with the better missile will win most dogfights. Robotized artillery isn't unreasonable either, as long as a human is around to tell it where to point and where to drive. And we're already seeing the value of robots in bomb disposal and information gathering.

But truly autonomous robotic cavalry and infantry are probably a long way off, if they are ever practical. First, to be really effective, they have to be essentially autonomous. Any method of communication can be jammed. If your infantry/cavalry depend on someone telling them what to do, they are useless as soon as the signal is interfered with. And they are terribly dangerous if someone else is able to trick them into taking the wrong orders. (There's a story for you -- computer hackers becoming the elite combat forces, as they try to break into and take over the opposition's forces, while protecting their own. Hmmm...)

Second, movement is an issue, especially for infantry. The human body is an all-terain vehicle practically without parallel. But the engineering involved in the human body is generations beyond what modern robotics can accomplish. And the more complex the robotics become, the less reliable they are, more fragile they are, and harder to maintenance they are.

Maintenance is the next major problem. The article presents the fact that robots don't need to eat, ect., as a strength. But they still need either fuel or electrical charge. And robots cannot heal themselves -- even the most minor damage will require a trained technician to deal with, as well as a constant flow of parts. I suspect that maintenance and supply issues would actually increase with the use of battlefield robots.

Finally, there is simply no substitute for human intelligence and initiative. While Big Blue showed us that within a simple fixed set of parameters, and with superlative programming, an artificial intelligence can equal or even beat the best of humans, that is far from showing that even the best of AI's could compete with trained, intelligent humans on the battlefield (especially since, contrary to what movies generally show, the humans will probably be more mobile and more resistent to injury that prevents them from functioning than the robots are).

Robotics (last I heard anyway) was leaning more toward emulating insects and the way they move. If robots weren't made to look like humans or move like humans, if they were like a swarm of insects, would that be more effective in combat?

Specifically, I'm thinking of a cloud of locusts effect.

As for maintenance, that's a pretty big problem. The money saved on medals probably wouldn't cover the cost of repairing the machines. It would be even worse with smaller, insect-like robots unless they were considered disposable from the start.

Minister, I love your story idea.

Edited to thank Corpsegrinder for the book recommendation.

No answer needed; the value is in the consideration of the question.

I think that the greatest problem is moral/morale.

quote:

---

Pentagon officials and military contractors say the ultimate ideal of unmanned warfare is combat without casualties.

---

You'll note that this is our primary justification for developing such a system. But this idea is incompable with fighting and winning a war, even a completely defensive war. Once the people making military decisions are making them on a basis of avoiding casualties at all costs, the nation is doomed to lose in war. A nation that is consistently losing wars is not in a good position to make a major military-technology advance. It soon won't be in a position to do much of anything at all, in fact.

It is only after listing the real reason that we're moving to robots that the "economic" justification is offered.

quote:

---

The median lifetime cost of a soldier is about $4 million today and growing, according to a Pentagon study. Robot soldiers could cost a tenth of that or less.

---

This is well and good in some ways, but the truth is that the listed pricetage of how much it costs to field a soldier is artificially inflated by our cultural attitude towards the risk of taking casualties. The economics of this only appears to hold up because we've already factored in the morale issue by means of programs to cover the moral cost of casualties economically.

I just came back from Universal Studios, where I went on the Terminator "ride." Part of the storyline is a demo of the new soldiers being developed by Cyberdyne: the Terminator robots. The dialogue and the real-life news reports are eerily similar.

Right now, remote piloted vehicles (RPV's)like the Predator posess limited degrees of autonomosity. For example, if communication is abruptly lost, the RPV has the ability to choose between several potential rendezvous/loiter/recovery destinations. The RPV then flies to the selected destination and awaits further instructions.

That may not sound very impressive, but the Predator is also armed with a pair of Hellfire antitank missiles.

To my knowledge, no "autonomous" Hellfire kills have been made by Predators, but I wouldn't be surprized if it has already happened...

Anyone unimpressed by Hellfire missiles can volunteer for target practice for all I care. I certainly don't go about my day to day business with the firepower to take out a couple of light tanks should the occasion arise.

However, the Predators do not currently have the ability to choose their own targets. That would require at least a next generation advance in their onboard visual processing capabilities for them to even be able to pick out a ground target for themselves, let along decide whether it would be a good idea to blow it up. Actually, I'm not sure that the Predator has any ability to use the visual information they collect, they probably do their semi-independent navigation by GPS and instruments for the most part. But I wouldn't be surprised if they can "see" some things. And ground targets would be a useful first step, that way they could run basically unmonitered until they "noticed" something interesting. But they aren't to that stage just yet.

We've already seen this with nuclear weapons; we use them because they give us more destruction at less cost. Reducing the need for human labor is the name of the game.

High velocity, mass, and human delivery are the watchpoint now for bombs.

I think that I would notice a barrel sized robot moving at a rapid walking pace, be it day or night. You'd have to disguise it as something else.

“Well, there's a certain diversity of opinion in the Pentagon. But overall there is a noticeable creep of "let's find ways to get paid without getting shot at" attitude. That's a perfectly sensible attitude for most lines of work.”

Well, no. I take that back. I DO mean to pick on Survivor, but only because he’s so much fun to debate with—I mean that as a compliment. No, really!

Anyway, the “noticeable creep” that Survivor talks about is actually one of the distinguishing characteristics of democracies at war. Democracies are always less willing than totalitarian states to expend large numbers of casualties, for obvious reasons. But on the other hand, this characteristic also means that a democracy will only declare war if it thinks it has a good chance of winning. Also, the armies of democratic nations tend to be more lethal than those of totalitarian states. In other words, soldiers from democratic societies are very efficient killers—I’m not just talking about technologically advanced nations like the US; this is a trend that goes all the way back to ancient Athens. The Athenians, after all, invented the concept of the “pitched battle”.

The common fallecy that drags down the military of a totalitarian state is Clausewitzian thinking. Basically, that you can make your soldiers fight best by inspiring them to fear their own officers and government (and the related idea that the best way to get the enemy to stop fighting is to make them fear your military). This idea persists because you can get soldiers to go out and fight by inspiring fear of their own side. But the judgement of history is very clear, normal humans fight best when their fear (and consequent hatred) of the enemy is the fundamental motive.

This is the basic reason that democracies are able to field more lethal, better motivated armies in the wars they end up fighting, because the individual soldiers are eager to fight the enemy. Every soldier on a battlefield that is thinking first and foremost about how to evade the wrath of his own officers and government is basically a minus when compared to any soldier that is there because he personally wants to kill the enemy.

Note, in none of this thinking is there ever a calculation of whether casualties will be heavy or light. The urge to avoid taking casualties is unrelated to the general desire--always present in a democratic army fighting a popular war--to inflict casualties on the enemy.

Perhaps you said it best, "the armies of democratic nations tend to be more lethal than those of totalitarian states." It is the individual soldier's will to kill the enemy that makes such armies successful.

Anyway, to the issue at hand. I don’t think you’re correct when you assert that soldiers of democratic armies are better motivated to KILL THE ENEMY than those of totalitarian states. In point of fact, during the Civil War, World War One, and World War Two, American and British soldiers were among the most poorly motivated to kill bad guys. (Then, of course, there were the French…we won’t go there.) The soldiers of Nazi Germany, Communist Russia, and the Imperial Japanese were much better motivated to kill the enemy—especially when they were fighting against a foreign invader on their own soil, as happened during the later stages of WW2. As far as individual soldiers were concerned, the German and Japanese soldiers were the best.

S.L.A. Marshall calculates that in combat conditions, only 1/6 of individual American soldiers fired their weapons at the enemy, when given a clear target. (Marshall has his critics, to be sure, but so far no one has managed to directly disprove his thesis. Furthermore, the works of earlier researchers tend to support Marshall.)

And that's my essential point. The Clausewitz/Prussian model, when applied to soldiers from a democratic society, is fundamentally defective. It worked far better for totalitarian armies. And even though reducing the soldier's fear of the enemy by allowing him to kill with little risk to himself is somewhat different from threatening him with dire punishment should he fail, it is still an expression of the same model of motivating soldiers to fight.

Marshall's estimates are hard to reduce to a single, neat figure. There were a variety of different circumstances posited for what percentage of soldiers were carrying out what level of combat behavior. For instance, in the circumstance where an enemy was in sight and an officer or non-com was present, most soldiers would discharge their weapons in the direction of the enemy, and the percentage went up further if the officer actually reminded the soldier to do so. But the evidence suggests that many of them were not actually shooting to kill or even hit the enemy. On the other hand, soldiers not being directly supervised would often engage in such "combat activities" as constantly reloading or adjusting their weapons over and over rather than firing at all.

Yeah, Marshall himself floats around between 25% and 15%. He never really boiled it down to a repeatable constant.

"...the figure drops to below a tenth for most WWII engagements and something far less than that even for many of the Civil War battles..."

I can believe that! Several years ago, I participated in a Civil Reenactment in Missouri. There were approximatly twenty thousand uniformed participants, 100 mounted cavalry, and at least that many full-sized artillery pieces. The battles were all pre-scripted and very tightly controlled. Anyway, as I stood in the front rank with several thousand other Yankees, reloading and firing blank cartridges from my musket, I wondered how in hell ANYBODY could actually do this in actual combat. The noise and smoke had to be experienced to be believed--and we were shooting blanks with only thirty grains of black powder! A typical Civil War cartridge had seventy grains. And on top of that, a minnie ball makes quite a bit of noise as it cuts through the air. (I know that from standing in a target trench at a shooting range, raising and lowering targets while other shooters fired black powder rounds over my head. I simply can't imagine what cannister must have sounded like...)

The same thing is true in an actual game of cards, as it so happens. Not that that has anything to do with anything.

For example, American artillery routinely fired TOT barrages (Time On Target--that’s where hundreds of guns scattered over many miles would fire in such a way as to allow their shells to land on target within a few seconds of each other). German TOT barrages were extremely rare, and also very small. Also, the Allied counterattack after the Ardennes Offensive was spear-headed by a massive barrage of shells with radar proximity fuses. These things were designed to explode approximately fifty feet above the ground, showering everything below with shell fragments.

Getting back to the original topic, there was one particular robotic weapon deployed during WW2—the Mk 24 acoustic homing torpedo. This was an air-dropped torpedo that ran a pre-programmed search pattern for a period of about fifteen minutes. It had a vacuum-tube signal processing system that determined when to deviate from the search pattern and home in on a potential target. If it didn’t locate a target by the time its batteries ran down, it would automatically sink to the bottom.

It was a very successful weapon, by the standards of the day. Out of approximately 340 Mk 24’s launched in combat, 37 German and Japanese subs were either sunk or damaged—that’s a hit ratio of about 11%.

That element is clearly necessary to an entity that is not just going to be responsible for hitting the target, but for making the decision whether or not to fire in the first place. For the time being, developing a non-expendable automaton that can be entrusted with its own survival while carrying out a mission is the expression of that challenge. We might have those in under a decade, or not. It depends. I know there are bright people with military experience and electronics engineering training, and there are more in the pipe. Whether there are enough, along with the other necessary factors, is still an open question.

TOT barrages...you can see why the Germans didn't feel they needed them. TOT barrages are a method of concentrating firepower in order to weaken a defensive position prior to an assault. They had no real place in the (highly effective) attack strategies developed by the German command, and they are nearly useless for a defender faced with steady losses of territory. There was simply no phase of the war during which the Germans would have had much reason to use them.

It's actually interesting where Marshall's research, artillery, and military robots meet. While I don't approve of the concept of "bloodless combat", I do think that there has to be a fundamental understanding that the vast majority of humans (the kind that can be substantially affected by conventional artillery, for instance) have no place on the battlefield (except as hostages, impromptu or otherwise). But it is important that we not forget that there are other humans that are...different.