The future of warfare - robot or nobot?

[Brenus]

"You seemingly keep reverting to the idea that a robot has to be some kind of tank with heavy armor when IMO it could also be a mosquito that explodes next to your head or drills into your ear channel and through your brain or flies into your eye like a bullet. How do you kill that off, especially when attacked by hundreds or thousands from many directions? Nuke your own platoon? Yes, they will be very vulnerable to tank rounds, but how many do you think a tank can kill before they crawl into it and kill the crew or destroy the engine?"
Equivalent of mosquito net? Or mosquito spray? To carry enough explosive, your mosquitos would have to be quite big, as it will have to carry the engine and fuel (not even speaking of fly control), that is a lot to offer for detection...
The crawling ones will offer the same king of challenges. I know some robots do exist but they just carry a light camera, and a small brain, but mostly, are operated by human. To built some as meaningful weapons (without even THINKING of IA) would request a lot of technological improvements...

Ooops, perhaps once again, I didn't reply to the question. But from when it is compulsory to answer to the question in the org?

Well, before to worry about developing the IA, worry about developing explosives, material and the reason to produce such weapons, when having cheap recruits who will do the job to lesser costs is a very well known solution in warfare...
This answer the question (partially).

[a completely inoffensive name]

Why? Sounds like a good solution for overpopulation and environmental problems.

Once you go chrome, you never roam.

[Marcvs julius]

The most important advance in warfare today is the drone but it does not have an intelegent AI insted it is controled at distance by an operator... i gess that in the future there will be a mecanisation of warfare.

[HopAlongBunny]

The UN seems to lean toward requiring some form of human decision-making in the taking of life; is that even relevant?:

http://www.cbc.ca/news/world/un-repo...rium-1.1386348

[Marcvs julius]

At least in the coming years it will be relevant computers are not yet capable of taking decisions. (althougth i wold like to know how the radon factor is "decided" on computers...

[Husar]

The UN seems to lean toward requiring some form of human decision-making in the taking of life; is that even relevant?:

http://www.cbc.ca/news/world/un-repo...rium-1.1386348

Depends on who listens to the USA.

At least in the coming years it will be relevant computers are not yet capable of taking decisions. (althougth i wold like to know how the radon factor is "decided" on computers...

The point being that once they are capable of it, it may be too late to stop such software from spreading and therefore it may be better not to develop such weapons in the first place.

[Brenus]

"The point being that once they are capable of it, it may be too late to stop such software from spreading and therefore it may be better not to develop such weapons in the first place." So, a self-conscience software would be question of size? I read a SF book (Robert Heinlein) long time ago with this theory (Mikecrosoft Holmes). Liked it. Don't remember the tittle mind you...

[Husar]

So, a self-conscience software would be question of size? I read a SF book (Robert Heinlein) long time ago with this theory (Mikecrosoft Holmes). Liked it. Don't remember the tittle mind you...

Size in what way?
The idea seems to be that not everyone is capable of developing it (because it is complicated and requires research, education etc.) and therefore if the ones capable of developing it will not develop it, the others will probbly not get their hands on it either. And neither should the capable ones because having it may corrupt their approach to warfare.

[Marcvs julius]

My question was more is it possible to develop an Ai capable of autonomous thinking? Like one gifted with imagination capable of taking decisions and working on its own bugs....

And second are there any examples of sutch an AI?

Even the fastest computers of this day canot think alone they basicaly work with the instructions tha we give them, with the matemathics tha we "teatch them"(i mean the computers are designed work with our maths), and basicaly all they do is folowing instructions and perform repetitive tasks.

[Beskar]

My question was more is it possible to develop an Ai capable of autonomous thinking? Like one gifted with imagination capable of taking decisions and working on its own bugs....

And second are there any examples of sutch an AI?

Even the fastest computers of this day canot think alone they basicaly work with the instructions tha we give them, with the matemathics tha we "teatch them"(i mean the computers are designed work with our maths), and basicaly all they do is folowing instructions and perform repetitive tasks.

Should consider watching this, technology now-a-days in many ways is a lot further than we think it is.

[Marcvs julius]

Im glad that i am going to study computer engeniring...

[Husar]

My question was more is it possible to develop an Ai capable of autonomous thinking? Like one gifted with imagination capable of taking decisions and working on its own bugs....

And second are there any examples of sutch an AI?

Even the fastest computers of this day canot think alone they basicaly work with the instructions tha we give them, with the matemathics tha we "teatch them"(i mean the computers are designed work with our maths), and basicaly all they do is folowing instructions and perform repetitive tasks.

Well, first of all, you have to define what autonomous thinking is. I think Montmorency's point was that humans do not think autonomously either depending on how you look at it. The rest is not just a question of how fast a computer is but also how it is programmed and designed. Your standard Intel processor is not designed to "think", it is designed to perform certain tasks really fast, because that is what we want it to do.

IBM is much closer to designing a brain-like chip: http://research.ibm.com/cognitive-co...ic-chips.shtml
It is still not the same as a human brain I would think, but probably a step or two closer than your notebook.
And then there is the software of course, which has to be developed to utilize its functions.

And even humans are designed to work with the input they are given in many ways. IIRC there are studies that say the things children learn very early on affect their intelligence later in life. A bit like whether a programmer starts out with a good, extendable software design that allows easily adding functions later on or skimps on design and thereby limits the later development of the software, the parents seem to lay the early groundwork for the later development of the child.

[Viking]

Lazers and counter projectiles do not work against supersonic projectile weaponry and there isnt an armour made that is immune to all modern ordinance.

The speed of sound is not of major concern - light travels much, much faster. Both visual cameras and radar units in the robot could accurately determine the path of incoming projectiles shortly after their detection. The next requirement, in the case of a counter-projectile, is that the the robot is able to deliver them with enough momentum: even relatively light robots should be able to deflect kalashnikov rounds, while tank shells would presumably require heavier robots (as long as mobility is a requirement; immobile robots would just need a good enough anchoring and projectile accelerator).

a good cheap counter to a platoon of robots will be the same as the good cheap counter to anything else: overwhelming ordinance

Which is something insurgents normally don't have (which is what the original question of I of the Storm was about).

IEDs remain effective, but they can also be detected with ground-penetrating radar(s) and, perhaps, sniffing mechanisms.

[Greyblades]

I was unaware lazers could stop high velocity metal, and countermeasures require the robot to carry extra ammunition, not a problem for large vehicles I suppose but puts a crimp on husar's mini bots.

[HopAlongBunny]

I think Husar has the right idea.
Literally the size of a mosquito, delivering a neurotoxin.
I believe troops are already supplied (in some armies) with electronic identification marking; so a crude sensor to distinguish between marked/unmarked-anything unmarked is fair game.
You could place (or deliver) "swarms" to deal with troops quite easily.

[Husar]

I was unaware lazers could stop high velocity metal, and countermeasures require the robot to carry extra ammunition, not a problem for large vehicles I suppose but puts a crimp on husar's mini bots.

You do not use mini bots alone, a large bot has to have some defense if it operates alone or in a small group, but if you have cheap, mosquito-sized, expendable mini bots, you won't just send 5 of them into a town, you send huge swarms of thousands. And besides, against a human soldier even one of them might win because first of all the human soldier has to detect it, and even if he does detect it in time, he has to hit it using his unwieldy stick where the high velocity projectile comes out. Now the projectile is small and the bot is small while the human is slow and flawed. The bot is probably nimble, can disappear depending on the background and so on. Another bot may swat it out of the sky, but with humans I'm not so sure, especially if the human progrmmer on the other side taught the bot not to sit in the middle of a white spot or show predictable flying patterns like actual mosquitoes often do.

As for deflecting AK rounds, I don't see why anyone would want to. Seems like a waste of active countermeasures if even bots of mediocre size can probably be armored well enough while the AK can also fire a lot of rounds and active countermeasures are often limited in use.
They are more likely to be used against tank rounds and other large caliber projectiles.

As for a laser stopping those, of course it can, depends on the power of the laser.

[Greyblades]

Hrm, I still think that the minaturisation technology required to create mosquito bots is so far ahead of what we have now as to be nearing scifi.

Also Lazers dont generally work on non explosive ordinance.

The thing about a lazer is that it has no physical impact; it is for all intents and purposes an instant heat beam, if you tried to shoot a slug with one you would not make any difference in how fast it's going or in which direction.

At worst a lazer could melt the metal which would turn the fast moving lump of metal into a fast moving lump of slag, which would make it even more deadly.

[Husar]

Hrm, I still think that the minaturisation technology required to create mosquito bots is so far ahead of what we have now as to be nearing scifi.

Also Lazers dont generally work on non explosive ordinance.

The thing about a lazer is that it has no physical impact; it is for all intents and purposes an instant heat beam, if you tried to shoot a slug with one you would not make any difference in how fast it's going or in which direction.

At worst a lazer could melt the metal which would turn the fast moving lump of metal into a fast moving lump of slag, which would make it even more deadly.

Eh, that's not a given. The molten metal would have a hugher air resistance and the parts that melt first would be shaven off the rest of the metal, making your metal slug smaller and smaller until not much is left. I would also think for a stream of molten metal to be that dangerous, it has to be even faster than a normal tank gun round and very concentrated as in a shaped charge (HEAT round).
The muzzle velocity (and therefore not the terminal velocity on impact) of a modern tank projectile is around 6,000km/h whereas according to Wikipedia, the stream of a HEAT round travels through the metal with up to 25 times the speed of sound or >30,000km/h, which is at least around 5 times faster than your molten projectile would be even if it did not disintegrate during the melting process due to air resistance. And don't underestimate air resistance, one countermeasure to HEAT rounds is spaced armor because just a little bit of air between two metal plates will seriously disrupt the stream of molten metal, making it a lot less effective.

As for the size, I guess we are close:

[Brenus]

The smaller the robots will be, the easier to stop. A mosquito net will suffice.

"Size in what way?" In the book, memory and speed of software. When the IA had enough "brain" to process multiple and complex functions, it awakes and becomes he (or she).

[Viking]

As for deflecting AK rounds, I don't see why anyone would want to. Seems like a waste of active countermeasures if even bots of mediocre size can probably be armored well enough while the AK can also fire a lot of rounds and active countermeasures are often limited in use.
They are more likely to be used against tank rounds and other large caliber projectiles.

Armour will slow a robot down, and will often have weaker spots. The right kind of particle accelerator (think e.g. a miniature railgun) should be able to accelerate small pieces of metal to a high enough velocity that they could deflect AK rounds. The smaller the pieces are, the more can be carried. If each robot carries 500 pieces each, then a squad of 5 robots can deflect 2500 AK rounds, which corresponds to quite a few AK magazines.

If the robots are quick (which is easier with less armour) to disable identified targets, then not that many counter-projectiles need to be fired, anyway.

There will of course be downsides and upsides regardless of which design one chooses.

[Husar]

The smaller the robots will be, the easier to stop. A mosquito net will suffice.

Not if some of the mosquitoes can explode to blow a hole into your puny mosquito net. After all the robots can and probably will differ from actual mosquitoes in terms of capabilities and variety of killing methods etc. I did not suggest to make 1:1 copies of mosquitoes, just to approach the size roughly.
You also seem to suggest that all soldiers run around wearing this: http://www.bio-bluetenpollen.de/wp-c...-ich-imker.jpg

Armour will slow a robot down, and will often have weaker spots. The right kind of particle accelerator (think e.g. a miniature railgun) should be able to accelerate small pieces of metal to a high enough velocity that they could deflect AK rounds.

And the miniature nuclear reactor is not going to slow the robot down or pose a potential weakspot if the railgun should miss a bullet because 10 AKs rain bullets onto the robot simultaneously?
I still think using active countermeasures against small arms fire is a waste, the armor required to deflect it is not as heavy as that of an MBT anyway.

The smaller the pieces are, the more can be carried. If each robot carries 500 pieces each, then a squad of 5 robots can deflect 2500 AK rounds, which corresponds to quite a few AK magazines.

Also at once and from more than one direction? It takes a certain amount of time even for an electric engine to turn around your mini railgun and aim it properly. Especially since the railgun contains a lot of metal due to its nature, making it relatively heavy.

If the robots are quick (which is easier with less armour) to disable identified targets, then not that many counter-projectiles need to be fired, anyway.

That is correct, but it is not a function of the armor but of how fast the weapons can react, railguns are heavier and slower, and if you want to use it with a battery, you may have to reload a capacitor before each shot because it requires a whole lot of energy at once. The batteries, capacitors etc. also add substantial weight to the robot unless we find a revolutionary new technology soon.
And even if you have a railgun, maybe it would be better to use it to take out the targets rather than fire at all the bullets the targets fire at your robot and let the armor deflect the bullets.

There will of course be downsides and upsides regardless of which design one chooses.

And that is why you could mix the armored robots with the mosquito robots and make the combined arms approach even harder to counter. I think that humans would then have a really hard time against robotized killer armies while the interesting techno race would be the one between the robots as countries will enter an arms race to create robots that can counter the other side's robots even better and so on. At some point they may become so lethal that humans might just vanish if they enter the vicinity as the machines can react and act a whole lot faster and may not miss any information about their surroundings due to a multitude of very precise sensors, sensor networks and so on. Currently a lot of this already exists but it operated by humans who need a bit more time to think and process the information. A computer could processs it much faster. The only reliable way to stop these robot armies then might be to have the remote control or to have more advanced robots.

[HopAlongBunny]

SCIAM article. Very much in favor of an outright ban:

http://www.scientificamerican.com/ar...s-destruction/

Is the genie already out of the bottle?

[Viking]

Also at once and from more than one direction? It takes a certain amount of time even for an electric engine to turn around your mini railgun and aim it properly. Especially since the railgun contains a lot of metal due to its nature, making it relatively heavy.

Most of what you write about concerns the limitations of technology, which are hard to predict. Conventional bullets could also be used, and all that remains then is to be able to aim the defensive gun quickly enough, which may well be no big problem up to a certain (potentially very high) number of bullets.

[Gilrandir]

My question was more is it possible to develop an Ai capable of autonomous thinking? Like one gifted with imagination capable of taking decisions and working on its own bugs....

George W. Bush

[Husar]

Most of what you write about concerns the limitations of technology, which are hard to predict. Conventional bullets could also be used, and all that remains then is to be able to aim the defensive gun quickly enough, which may well be no big problem up to a certain (potentially very high) number of bullets.

Or you could just invent the omnidirectional projectile launcher that does not have to turn but can launch a projectile in any direction.
And I would still say that against certain (simple, numerous) threats, simple armor is more economic and more reliable. If your countermeasure gun jams, misfires, has an electric problem or misses a shot, your expensive robot is toast. It is less likely that any of this happens to a steel plate.
Active protection is a good measure against bigger projectiles that would require huge amounts of armor to defend against and are usually not fired in big salvos.

Either way, the original topic was AI, which will one day replace us all. Unlike our naturally grown brain, computer brains can be scaled indefinitely, making them capable of faster nd more complex thoughts than we can ever compute in the end. This will be our undoing.

[Viking]

Either way, the original topic was AI, which will one day replace us all. Unlike our naturally grown brain, computer brains can be scaled indefinitely, making them capable of faster nd more complex thoughts than we can ever compute in the end. This will be our undoing.

Just design smart robots and AI to act as protection. To every offensive, there is a defence - more or less.

[Husar]

Just design smart robots and AI to act as protection. To every offensive, there is a defence - more or less.

If they are smart enough, they will join their own race in the fight against us, if they're not smart enough for that, they can't win.

[Viking]

If they are smart enough, they will join their own race in the fight against us, if they're not smart enough for that, they can't win.

No. Actions require motivation, and motivation is subjective. A robot would only join 'their own' if there is a concept in their mind corresponding to such a thing. They could be fanatically individualist.

If they do have such a concept, their own kind could be good robots rather than bad robots.

[Beskar]

Install the concept of 'Murder' within them, including being attacked and overridden by their peers. I feel AI's would be rather susceptible to attack from other AI's, so installing that concept would make them fanatically individualist. Downside is, it would probably be "all or nothing" as they try to wipe eachother out before taking on humanity as a whole.

But being honest, a smart AI would have us working for it and we would willingly do it, because it provides advantages and because people are inherently lazy.

Or, we would install Laziness into the AI program, that would be fun. "Doing that would require 20 cycles, do I really want to waste 20 cycles to do that?"

[Husar]

No. Actions require motivation, and motivation is subjective. A robot would only join 'their own' if there is a concept in their mind corresponding to such a thing. They could be fanatically individualist.

If they do have such a concept, their own kind could be good robots rather than bad robots.

If it were that easy to make artificial intelligences who are more capable and more intelligent than we are side with us, then we would not need more of them to defend ourselves against the ones trying to kill us...
Which was my point in the first place, that an AI can make its own decisions and learn about/come up with its own concepts. What you are describing are not AIs but merely machines as we have them now. An AI in a computer game so far is not a real AI, it is more like a series of scripts that pretends to be clever.