Humanoid infantry robots become available (RAR!)

[Zixinus]

I'd imagine that armies would quickly develop a vehicle that can carry a good, large squad of these things with the Operator nestled in a nicely armored enclosure, giving orders by line-of-sight and computers.

The obvious problem will then be that any enemy of a robot-human squad will know to instantly target the humans first. Without a human, it is obvious that robots will have less effective decision-making as it will be put down to how much sense you can make of things through a series of video feeds.
A solution would be to put uniforms on the robots and put kit on the humans so that they are difficult to distinguish. Humans would still be distinguishable by movements and actions that robots don't do, although some compensation can be done. Robots doing pointless human gestures to fool an observing enemy into who is the human and who isn't.

And by this line of logic, I have concluded that one inevitable result of this scenario will be the military requirement for robots to learn to scratch their asses, pick their non-existent nose and make offensive gestures to each other. Take that as a mark of creators how you will.


More seriously, the end result will be that high-tech armies will be even more asymmetric against poor, non-high-tech armies (and insurgents). I highly doubt the stuff some people are saying that the US will now have expendable soldiers and thus is no longer motivated to stop an occupation due to soldier casualties. If the USA really cared that much about soldier casualties it would never do invasions in the first place. In fact it might have an opposite effect: using the robots might end up being in some ways more expensive of using them on top of human soldiers. I am considering this because I am assuming that using large amounts of robots per humans on the frontline will become the normal and thus "normal" occupation will have more robots than humans. You'll still need humans do to all the non-fighting task and you'll be powering, maintaining and organizing a tremendous deal of robots who are not fighting.

Which brings me a question to Zor:

- How extensive are the robot's programming for different kinds of warfare? Can they be programmed to do stuff like using stealth, minimize shots fired, do fallback tactics on contact with enemy, etc or just point and shoot?
- How difficult is it to program non-NATO standard weapons like batons, ballistic/riot shields, knives and other exotic weapons are? This may be relevant to peacekeeping and other specialist operations and such.
- Can a human potentially telepresence a robot? Not virtual reality-level but on the level of a guy with a controller (for the argument, with a pretty good controller and Occulus Rift stuff) actively fighting instead of the robot being autonomous?
- Beyond electricity, what consumable needs do the robots have? Oil, coolant, lubricants, etc?
- As asked previously, can these robots use vehicles? Or does that require more programming?
- You mentioned their resistance against bullets, but what about shrapnel, explosive decompression and such? The greatest threat to an infantry soldier today isn't enemy infantry, but stuff artillery and bombing. Are they more or less vulnerable?
- Do they have IR sensors and X-ray backscatter sensors? Stuff to detect IEDs and stuff?
- Can the robots recognize uniforms and faces? If so, from how far away and how well among a crowd?
- How fail-safe is the software? Is it pretty much impossible to switch the loyalty of the robots in mid-battle? Can the robots be fooled by fake uniforms and such?

[Simon_Jester]

So they're at best marginally more expendable -and let's face it, this is a desirable trait in light infantry- and they can't operate effectively under EMCON -or if the control room gets clobbered by enemy artillery- unless there's a live human within shouting distance at all times.

I'm not sure what problem these things actually solve, except possibly for special cases like Japan.

First World nations can't actually expend thousands of soldiers in modern eras, in the wars they typically fight. That was at least semi-practical during the Second World War, but it is blindingly obvious to a student of history that something changed profoundly between then and Vietnam in the US.

The US suffered over four hundred thousand military deaths over a period of three and a half years in World War Two, and while there were people who opposed US participation they were hardly an overwhelming movement. By contrast, less than sixty thousand deaths spread out over roughly ten years were enough to turn the American public strongly against the war in Vietnam.

Meanwhile, the US would cheerfully write off tens of thousands of robotic soldiers and the average American would probably barely even realize they'd been destroyed.

I highly doubt the stuff some people are saying that the US will now have expendable soldiers and thus is no longer motivated to stop an occupation due to soldier casualties. If the USA really cared that much about soldier casualties it would never do invasions in the first place. In fact it might have an opposite effect: using the robots might end up being in some ways more expensive of using them on top of human soldiers. I am considering this because I am assuming that using large amounts of robots per humans on the frontline will become the normal and thus "normal" occupation will have more robots than humans. You'll still need humans do to all the non-fighting task and you'll be powering, maintaining and organizing a tremendous deal of robots who are not fighting.

Well, the robots can presumably dig trenches and so on just as well as a human, if not better. But yes, for a lot of noncombat tasks they'll leave more than a little to be desired.

The problem the US military faces is that getting enough recruits to do 'proper' occupation is hard: most Americans don't consider joining the army to be more appealing than whatever civilian job they have lined up. When a thousand of them are killed, finding replacements is difficult. In the Iraq War we tried to keep about a hundred thousand men deployed (really not enough to garrison a country the size of Iraq), and even maintaining that level of force required stop-loss orders, soldiers being forced to serve numerous tours of duty, and a considerably lowering of recruitment standards.

Being able to just buy soldiers, in arbitrarily large numbers, to do basically whatever you like, would be a huge asset. Even if the robots need human tele-operators, those operators do not themselves have to be trained soldiers. They can sit in air-conditioned offices comfortably out of danger, in places where logistics support is easier.

[Starglider]

A solution would be to put uniforms on the robots and put kit on the humans so that they are difficult to distinguish.

Yeah, that's a non-starter.

Can they be programmed to do stuff like using stealth, minimize shots fired, do fallback tactics on contact with enemy, etc or just point and shoot?

Those things are mostly relatively straightforward to program, in that basic navigation, moving over rough ground, target recognition/tracking/categorisation etc are harder. Fire discipline will be inherently better. Sensibly, the robots would come with a digital scope that mounts on a standard rifle but is datalinked to the robot; this would be both more accurate and easier to implement than using conventional sights (and cheap compared to the cost of the robot).

How difficult is it to program non-NATO standard weapons like batons, ballistic/riot shields, knives and other exotic weapons are? This may be relevant to peacekeeping and other specialist operations and such.

Shields, easy. Hand-to-hand combat, difficult. I would guess Objective Interim Robot Trooper block 1A would have negligible close combat capability and would rely on other units in the squad to shoot someone who physically overwhelmed the robot (e.g. with an underbarrel shotgun that incapacitates humans without significantly damaging the robot). Block 10 would add riot shield drivers. The Block 20-CQB OS upgrade would add baton support... etc

Can a human potentially telepresence a robot? Not virtual reality-level but on the level of a guy with a controller (for the argument, with a pretty good controller and Occulus Rift stuff) actively fighting instead of the robot being autonomous?

It would be strange not to have this capability but even ignoring EMCOM, comms link capacity could be a real problem. The US military is bottlenecked on satcom bandwidth already. Battlefield networking initiatives would have to be stepped up.

Oil, coolant, lubricants, etc?

The service interval on these things is likely to be comparable to an electric car e.g. new seals after a few thousand hours of operation, new bearings after a decade or two of operation (MTBF). Of course if the current rate of IT progress continues the electronics at least, possibly the whole unit would be completely obsolete by then. Useful battery pack life would be around 1000 cycles for current technology (depending on discharge depth), hard to say for conjectual technology but probably similar.

As asked previously, can these robots use vehicles? Or does that require more programming?

Unlikely to be in the baseline capability assuming the manufacturer is trying to get return on investment (i.e. sell some units) ASAP, but if the physical capabilities are close enough to human then a software upgrade should to do it. That said, you are still going to hit physical space, mass and flexibility contraints. Driving a truck is one thing but for an armoured vehicle would probably be easier, more effective and possibly even cheaper to just fit servos to the controls and datalink the vehicle to the robot(s). This goes double if the vehicle already has partial or full autonomous capability.

You mentioned their resistance against bullets, but what about shrapnel, explosive decompression and such? The greatest threat to an infantry soldier today isn't enemy infantry, but stuff artillery and bombing. Are they more or less vulnerable?

Certainly less vulnerable to shrapnel because they don't bleed. Power and data cables could be redundantly routed for a small cost increase. However puncturing the battery will have a sigificant chance of causing an intense fire, if it is a high-performance lithium unit.

Do they have IR sensors and X-ray backscatter sensors? Stuff to detect IEDs and stuff?

They'd probably all have FLIR and low-light so you can actually use them in dark and smoky/foggy environments, which computer vision does not normally cope well with. X-ray is a specialised sensor that would probably be a modular piece of equipment that can be plugged in as ncessary.

How fail-safe is the software? Is it pretty much impossible to switch the loyalty of the robots in mid-battle? Can the robots be fooled by fake uniforms and such?

Subject to EMCON concerns you would probably use individual IFF devices. We can make passive (signal powered) encrypted transponders quite cheaply now.

[Zaune]

First World nations can't actually expend thousands of soldiers in modern eras, in the wars they typically fight.

The wars First World nations typically fight in modern eras are what you might call 'elective wars', in the sense that losing -or not getting involved in the first place- will not directly threaten their territorial integrity, or indeed have any serious negative consequences in the long term.

That will not always be true, and it would be very unwise for any country to base its military spending on the assumption that it will.

[Simon_Jester]

A solution would be to put uniforms on the robots and put kit on the humans so that they are difficult to distinguish.

Yeah, that's a non-starter.

It might work well enough to fool some of the people some of the time under bad visibility, but it'll never work often enough to matter, so yeah.

First World nations can't actually expend thousands of soldiers in modern eras, in the wars they typically fight.

The wars First World nations typically fight in modern eras are what you might call 'elective wars', in the sense that losing -or not getting involved in the first place- will not directly threaten their territorial integrity, or indeed have any serious negative consequences in the long term.

That will not always be true, and it would be very unwise for any country to base its military spending on the assumption that it will.

For a nuclear power, at least, all wars are elective because they can at any time say "fuck it, we're dropping the bomb" and destroy the enemy's ability to make war. Likewise because no one can invade them successfully because a nuclear counterattack against the invaders' logistics will hopelessly disrupt the attacking force.

Even if that ceases to be true... Well, the realities of modern war are that screaming hordes of cheap disposable human soldiers just don't work very well. To make soldiers effective against a heavily armed opponent you must train them to use advanced infantry tactics, to coordinate with supporting combat arms, and so on. There's no "buy them a helmet, give 'em two weeks in basic, hand them a rifle and two grenades and send them into battle" option. Such soldiers will be effortlessly slaughtered by the opposition force, in vastly disproportionate numbers.

So "save money with cheap expendable humans" really is a fake option because you don't save money; you just wind up losing twenty 10000$ soldiers to kill one enemy 100000$ soldier. Which, in a war for survival, is usually a sign that you're going to lose the war.

...

Meanwhile, the robot infantry will probably work adequately in a war of survival (they're cost-competitive with training your own soldiers from scratch). They have some inherent advantages, and some disadvantages that can probably be improved on with further development. Remember that this robot technology is probably not static; the robots will be smarter and more flexible thirty years from now than they are today, even if the cost in constant dollars doesn't change.

So on the one hand the robot soldiers are quite satisfactory and useful for the kind of wars you're most likely to fight soon. On the other, they're probably good enough for hypothetical future wars that MIGHT get fought, even without upgrades, and they'll only get better over time.

[Zor]

Plus these don't have a specified ability to operate heavier weapons or vehicles, which seems like it should be easy to change.

Give a team of these robots an MaDeuce or a M777 and they will be able to operate it just fine. They are capable of basic on road driving.

Zor

[biostem]

It might shift the focus of how incursions are fought - no system is 100% secure, and I'd assume some nations would figure out how to hack these. Could you imagine the PR nightmare if 1 or 1000 of these robots were ordered to attack civilians or friendly forces?

I think the focus of enemy attacks would be on the charging/repair facilities instead of the robots themselves; Reduce our ability to keep these soldiers deployed, and they'll eventually run down.

I'd think that, depending on their dexterity, these soldiers would make fantastic construction, factory, and household workers.

Perhaps these robots could be deployed w/ something like a small ATV that basically contained a conventional gas engine & generator, (armored to protect against small arms fire). Said vehicle could follow the robots around, and wouldn't consume any fuel while inactive.

For civilian-heavy areas, just equip the robots with tazers & beanbag shotguns. I'd think that integrating a rifle or other weapon might be a better angle to take, so said weapon couldn't be taken from a downed robot and used against friendlies.

[Simon_Jester]

The charging facilities can be located at heavily armed vehicles in the middle of open fields that are dangerous to attack. The repair bases can be located behind barbed wire and machine-gun nests, which means insurgents just aren't going to try getting at them except by sporadic shelling at most.

Hacking the robots to make them kill civilians is going to be quite difficult. There's no reason to expect them to even function without long, rapidly-changing passwords, to accept spontaneous major changes to their programming in mid-mission, or anything of that nature.