# Robo-grunts in the army: viable?

Around 100 years from now society has entered a multi planetary society, with mines in space bringing in resources to Earth in abundance.

Utilizing this new material wealth, Earth Nations start to, metaphorically, buy the Rolls-Royce cars of the new hip military tech: soldier drones.

These humanoid robots have basic, dumb AI, superhuman strength, and no empathy whatsoever in whatever they kill. They excel in physical and emotional stats compared to regular soldiers, but can be easily dispatched with smarts and lots of lead.

Thus, the current military now consists of the common Robo-grunt for large scale combat, and the Operators: special forces soldiers equipped with exo-armor and full control of said robo-grunts (including self-destruct, when they get hacked)

Would these units be considered effective if they were to exist in modern combat alongside regular soldiers? Will they offer an advantage over humans, or just become an expensive broken iPhone for the military to deal with? (soft science is intended for this answer, no need for extensive research. This is just whether or not there are some loopholes in this system.)

• Reminder to Close-Voters: The OP can't fix a problem if he is not aware of it. Please state the problems you see with a question when voting to put it on hold so that these problems can be fixed fast and the question reopened or not even put on hold in the first place. @WorldCraftTrainee: You currently have one person voting to temporarily put this on hold as "primarily opinion-based". I don't know why. Maybe because of the not-exact future which leads to "How they fight their war (tactic-wise) is unknown"? – Sec SE - clear Monica's name Nov 28 '17 at 13:22
• I haven't voted to close (indeed I can't), but reconciling It is the future (no need to specify when) and will these kind of military units be effective in modern combat to make the setting more definite would help. – walrus Nov 28 '17 at 13:35
• If these wars are not fought over resources (as they are coming in abundantly from the outside), what are these wars about? – Daniel Nov 28 '17 at 14:07
• @Daniel If you trust my friends, wars are fought over social issues such as rights and class divisions. Just because resources are abundant does not mean they are evenly distributed, as can be seen by modern humans. – Muuski Nov 28 '17 at 14:14
• Just don't do them fully humanoid please. Something spiderlike will probably make a lot better universal weapon platform, if you anyway are going to solve problem with walking – Artur Biesiadowski Nov 28 '17 at 16:01

Having ground based robotic troops assumes that the strategies of combat don't evolve with the tech and with refined strategic approaches to war and that's simply untrue.

Drones have taken a lot of the intelligence and combat workload from more traditional human based combat solutions; for those countries who have access to them and can afford them. The idea of putting drones in the air that can strike from a distance and are much harder to detect coming makes a lot of sense for strategic, intelligence based strikes, which is what we're moving to in combat. Why send an army of robots against your enemy's army of robots if you know where the controllers are and you can take them out with a missile strike?

Even as early as WW1, Lieutenants and other field officers had horrific mortality rates because snipers on both sides were deployed specifically to take them out in trench warfare whenever they stuck their head up. If you were an Allied Lt in France during that time, you were more likely to die than survive. Why? Because taking out the officers was more effective at stopping your enemy than taking out privates because there was no-one to command the men in the field. One death, many ineffectives.

This kind of strategy has evolved even further, and drones are the current way to strike at your enemy's weakness (their control centres, commanders, etc.).

So, robot soldiers, while an understandable extrapolation, aren't the weapon of the future. Long range, 'smart' but controllable strike platforms are the order of the day. For now at least, drones fit that bill perfectly.

• What you say is true, but humanoid robots are useful. First proof: Boston Dynamics wouldn't make them if they weren't useful, and AFAIK they are military funded. Second proof: drones are only deployed when the objective is to destroy. There are more possible objectives, like guard duty, psychological effects of being seen by civilians, taking over infrastructure without destroying too much of it, probably more. – Mołot Nov 28 '17 at 8:49
• I would not consider current drones as robots as they do not act (fully) autonomously. They are remote controlled and the kill decision is still done by a human. – luator Nov 28 '17 at 9:28
• @Molot; You're right of course - there is still a practical application for humanoid robots. I'm just saying that the battlefield is not as likely to be the place where that happens. Guard duty is a very useful task for these kinds of robots as it limits the friend or foe, battlefield context problems combat will ultimately raise. – Tim B II Nov 28 '17 at 9:42
• @luator; not all robots act autonomously either; car assembly line robots work to a specific set of actions and could hardly be called 'smart'. I'd also argue that the kill decision being made by a human is a good thing at this stage in our AI development because in combat, context can be everything. As early as the 60s the phrase 'it was a computer error' was a bad joke; the last thing we should be doing is giving that as an answer for a massacre caused by an onboard system matching the wrong pattern to a situation. – Tim B II Nov 28 '17 at 9:45
• @Mołot Boston Dynamics are developing humanoid robots because the US military thinks they might be useful in the future. Whether or not they actually turn out to be is a whole other question (which is of course the point of this question). – walrus Nov 28 '17 at 12:54

One of the hardest things in military training is to get one man to kill another just because he's been ordered to.

This was a key driving factor from conscripted armies to professional armies. Conscripts don't shoot to kill, that's if they shoot at all. This is not a problem for robots.

Politically they're also a lot safer. Bringing home most of a robot because it was ambushed has a lot less political fallout than bringing home most of a soldier.

Irrespective of which of these factors is stronger in your society, robots on the battlefield are ultimately inevitable. Cost is nothing in military spending.

• Three things here; modern combat strategy (and weapons) are not specifically designed to kill, they're designed to incapacitate (the military calls it 'soft kill'). Second, Teutonic charges on the battlefield are becoming increasingly rare; it's just not how war is prosecuted anymore. Finally, I'd have a problem with robots designed to kill autonomously until I can be absolutely sure they won't fire at me through some form of error. That robots kill without question is actually MORE concerning on a frenetic battlefield than a soldier scared witless, in my humble opinion. – Tim B II Nov 28 '17 at 9:53
• "Cost is nothing in military spending." Not for major symmetric conflicts. Except maybe for surprise attacks, I would say those are mostly about resources and efficiency. If you run out of fuel and bullets, you loose the war. If I can buy 100 bullets for the price of one robot, but need only 10 bullets to destroy said robot, I'll win in the end. – nikie Nov 28 '17 at 11:21
• @nikie, You can probably buy 1,000,000 bullets for the cost of the robot. It's getting those bullets into the robot that's expensive. – Separatrix Nov 28 '17 at 11:29
• I disagree. The hard part is getting a man to die just because he's been ordered to. To kill another people has never been that much of a problem. The main problems with killer robots are two: its absolute lack of fear to death may make them overexposing themselves, thus being easy targets; and, fundamentally, how do you make sure they are shooting at the right people (i.e., not you). That's why we're limiting to drones right now; they are used wherever there's no AA defense to bring them down, and with a human supervisor to shoot. And even in this case, theres plenty of collateral damage. – Rekesoft Nov 28 '17 at 12:00
• @Rekesoft there's a very interesting video here on just how rare it is for people to try and kill other people without being trained to it - the headline figure from the video is something like 10% of WW2 American infantrymen who had been in a firefight actually shot to kill rather than shooting to suppress(I can't re-watch it right now to confirm the exact figure). – walrus Nov 28 '17 at 12:51

Current airborne drones leak data and require nearly continuous command and control. They work pretty well in practice because they are used in asymmetric conflicts and these downsides aren't cheaply exploitable. In a more even conflict these might be serious enough problems to require workarounds.

Voice, handsign or near field communication commands from a local human are probably more robust to detection and interference by opposition. Things closer to the ground, operating colder, and radiating lower power communications are harder to detect. Supporting weight on the ground is much more efficient then on air, allowing lower temperatures, higher loads or longer runtimes. This might lead to human controllers embedded in ground based robot units.

Which probably leads to targeting the humans first. Which perhaps leads to making the robots mimic the humans. Human-looking robots with robot-looking human officers.

• This makes the most sense to me. I was trying to justify embedded humans, and you nailed it. – Stephan Nov 28 '17 at 20:02

You need to consider escalation. If one side gets these robo-grunts, other sides will probably get them just to stay on an even footing (the arms race of the Cold War is an example of this).

The point to remember is that the thing that is better than a troop of robo-grunts can be two troops of robo-grunts, oooooor something that effectively neutralizes or destroys said troop of robo-grunts.

In an arms race, there will not just be a race to have more/ better of the same, but also to come up with something to make the robos obsolete.

So in an arms race, robo-grunts will only be useful until they are made obsolete by something better.

As a means of peace-keeping or show-of-force for the sake of the little people, they are likely to remain effectively scary far longer.

They may also be useful against rebel militia-types who lack their own robos to fight with, or against a massive alien force where sheer numbers might help win the day.

What I'm asking about is whether or not will these kind of military units be effective in modern combat.

I would say yes: There are already dumb robots in military (e.g. EOD Bots). There's a lot of jobs that are dangerous and only require some guidance but not intelligence. The Operator would probably just send a bot into a potentially dangerous place first. "Alpha6 move in on potential EOD, direct video feed to command!"

Will they offer an advantage over regular soldiers, or just become an expensive broken iPhone for the military to deal with?

That depends on how expensive they are going to be. But think about how expensive a human soldier is! He/she needs extensive training, equipment (you mentioned exo-skeletons) and care if wounded. Now if those bots are dirt cheap the operators are going to just send them in and not care if they are destroyed. If they are rather expensive but still cheaper than human foot soldiers operators are going to be more careful and maybe use them for protecting human soldiers but not just sacrifice them to clear a mine field.

As a side note: If you have the tech to make those robots you can probably also figure out a way to make a human control them maybe with a advanced VR setup so they have the intelligence of a human with the strength of the robot.

• Welcome to WorldBuilding lukas! If you have a moment please take the tour and visit the help center to learn more about the site. Have fun! – Sec SE - clear Monica's name Nov 28 '17 at 8:47
• Remote controlled robots may however have an issue with lag and feedback. FPS games regularly show the importance of frames and reflexes that are measured almost in milliseconds. And recreating more of the senses would drive cost up, not to mention the data it would take to send it to a platform, and then send control gestures back again. – MartinArrJay Nov 28 '17 at 9:29
• If there are expensive robots for advanced tasks, there are surely also cheap robots for mine-clearing etc. – Daniel Nov 28 '17 at 14:10
• Many years ago I think I read/watched something about some soldiers that put a 50 cal machine gun on a bomb squad robot and used it for patrols in Afganistan (I think). If my memory serves me then we already have robot ground troops. – Muuski Nov 28 '17 at 15:10
• Yep, if you lose a grunt it is also a socio-economical loss for your country - less workforce and less purchase capacity. As I understand it is not much of a deal in the USA but in Russia with only 145m peeps and constantly declining over past twenty years it is one of the factors of adapting drones and casualty reducing vehicles and equipment. – Nick Dzink Nov 29 '17 at 20:24

Having robot soldiers marching around provides several advantages over conventionals armies.

• They can be mass produced with quality. You want a gigantic army? Then all that matters is having an important industrial capacity and money. On the other hand you can not mass produce humans, they need much more time to perform the same basic tasks a killer robot would do.
• They put less strain on your economy, at least provided your bots are dumb enough to not be able to be part of the workforce of a country. A developed country with a lot of its economy oriented toward service couldn't use bots to work in place of humans. Thus couldn't send much of its workforce to war, without exhausting a lot the economy of the said country. Remember when the third Reich mobilized something like 8 million soldiers out of 70-80 million people? Such things weigh a lot on a country economy, even though plundering other countries helps alleviate such costs.
• Human's lives matters. For politics, it would be very important to show that your fellow human citizens are worthy enough for the government to spend billion replacing human soldiers with bots.
• At the end of the war, if you don't need those bots, you just dissasemble them, or mothball them. As far as I can tell you can't mothball humans, and sending back to your country soldiers who fought for several years can be a difficult tasks. They must adapt once more to civilian lives, and depending on how well the war turned out, it might be very hard for them. Remember the movie "Das Wunder von Bern", or the Miracle of Bern, when the father comes back from its prisoner camp. And how well other germans think of him. Now, deal with 8 million conscripts going back from a world war, with no jobs at all and sometimes with hate from the public opinion (remember Rambo!)
• A wounded soldier costs more to its army than a dead one. You must take care of him, several people need to support them. And apart from physical injuries there are also mental ones to deal with. The horrors from war caused many trauma, especially in the first one. Do you remember when the Allies never wanted to wage a world war again after that one? Remember how people let fascist do what they pleased, hoping to avoid another slaughter? That's the cost of having humans bear the weight of a full out war. Your bots on the other hand will never ever complain.
• When a bot dies, you can replace easily it. When a human dies, you can't do that so easily. For a small and rich country, it would be the perfect fit to counterbalance a low manpower (or low commitment to the army). Counterbalancing manpower with technology was attempted several times in human history, and sometimes it worked like wonders: industrialization of agriculture for instance, which saved enough manpower to send more men to war without starving to death. This is also why China or India couldn't send tens of million soldiers against the Japanese/Germans, because most of its workforce was needed elsewhere.

So what more do you need? Invest in automated war bots, saves human lives, enhance your war economy, and show everyone how good a ruler you are.

• Good points and "As far as I can tell you can't mothball humans" => +1 – Sec SE - clear Monica's name Nov 28 '17 at 9:51
• "you can not mass produce humans" Have you never heard of Genghis Khan? ;P – Muuski Nov 28 '17 at 15:13
• @Muuski afaik Alexander the Great did the same thing. – Nick Dzink Nov 29 '17 at 20:26
• @Nick Dzink All I did was do a Google search for who had the most children and GK popped as the winner by a long shot. I don't know the accuracy of the list. It was a really lazy joke. ;) – Muuski Dec 1 '17 at 20:11

Joining in late but here it goes.

No the robots wouldn't be used in modern warfare simply because they have dumb AIs. What the robot needs to be viable is a smart AI. A robots ability to aim and shoot as well as walk around would all be directly tied to how smart and powerful its AI is. Just because a robot has detected movement, doesn't mean it is smart enough to identify the movement as a enemy, friendly, civilian or just the environment around it. It would be more effective to have the robot grunt remotely controlled by operators (Which I assume is possible because you have a fully autonomous robot and exo skeletons).

Next in terms of raw power/destruction and beefiness, it would still end up being a foot soldier. More effective methods such as drone/planes/ships/armored vehicles/tanks/missiles/etc would render the robots ineffective. This ties into the dumb ai as well as the robot would likely be unable to react to sudden attacks, not stay under cover, walk in the open and so on as it is supposedly "dumb". The pure development cost and production costs of several robot grunts would be better spent developing larger scale missiles/drones fitted with better weapons which would rain fire onto the robot grunts.

The main problem I see with a robot grunt, is that any amour or weaponry it is assigned would be better and more cheaply be placed on a person. The robots power supply, processing unit and motor sizes would all be fairly large and bulky, and if they were more effective than a human, it would still end up being a large bulky robot, which would lead to more armor and more weight and an endless cycle.

What would be very effective is a smart robot that follows orders to the letter. Basically the reverse situation. That would allow the robot to utilize its superhuman strength, lack of emotions, as well as its computer brain which would allow it operate and move with perfect(or close to perfect) accuracy, react faster than humans can comprehend and allow it to predict human behaviors and tactics with cold hard logic.

• Very realistic answer! I was actually expecting these kinds of responses earlier instead of the positive-ness the question got. – WorldCraftTrainee Nov 29 '17 at 6:59

From a military point of view, robot infantry have a large number of upsides.
Separatrix mentioned one of the most important factors, namely obeying the order to kill. This is not an issue for a robot. And should it be, fix the programming.
Separatrix also mentions the problem of relatives and the general public not valuing crippled heroes very much, and how this doesn't apply to machinery.
Then there is fear, and misjudgement: I seem to remember that during Operation Desert Storm, most wounded US soldiers were victims of friendly fire. Even if that wasn't true, friendly fire is always an issue in a chaotic scenario, where shooting first and identifying the target second might be a decision of life and death. A robot has no life and doesn't die, so even if it was capable of fear, it wouldn't fear death or pain, and it should be highly capable of distinguishing friends from targets.
As a result, a robot should be more efficient in many ways.

Also, a wide variety of weapons are useless against robot. Psychological, biological and chemical warfare simply won't hurt a robot.

A robot doesn't even need air to breathe. And it only needs power while active. It doesn't engage in recreational activities, it doesn't become tired, it doesn't need sleep, but when off-duty, it needs nothing: no air, no food, no entertainment, no comfortable temperature.

You mention a multi-planetary society. Just think for a second about the difference in transporting a human army and a robot army. The robots can be stacked in the cold, dark and airless cargo hold, and stay there until they are needed. Humans have a weird tendency to not responding favourably to such treatment. Oh, and humans generally value being paid, preferably continuously.

Finally, there is the issue of training. Every human soldier has to be trained individually. This takes time, the only resource that cannot be compensated for. A robot gets his training via upload, which can happen remotely, even while in combat.

The downsides will probably be limited ability of decision-making. But that is easily compensated by remote commanders.

• "easily compensated by remote commanders." — not easy in the planetary transport situation you mentioned bit earlier. Operators on another planet would experience really hard lag. – Mołot Nov 28 '17 at 9:10
• @Mołot the commanders may well be on the same planet, or in this planet's orbit. They are removed from the effective range of the enemies weapons, though. – Burki Nov 28 '17 at 9:20
• "it doesn't become tired, it doesn't need sleep", "no air, no food, no entertainment, no comfortable temperature." - I'd dispute this - it has a limited energy supply in the form of fuel or charge, which needs to be replenished. And even machines have certain operating limits in terms of temperature. A robot could be just as resource hungry as a human, if not more so: while food may be relatively easy to supply, a robot may require a specific type of battery which requires specialised factories. I think, in spite of its need for entertainment, the human wins this one – colmde Nov 28 '17 at 9:20
• @colmde it definitely doesn't need sleep, and if it runs out of energy, it stops operating, until eventually more energy is provided. Try that same trick on an armed human. You might be faced with some objections. – Burki Nov 28 '17 at 9:21
• If commanders traveled to the same planet, you saved nothing of the cost of human transport. That's my point. – Mołot Nov 28 '17 at 13:05

Pros:

• Accuracy: Using machines and infrared sensors to aim your weapons could potentially make robots vastly more effective than human troops. Being able to reliably hit a target from insane distances would effectively neutralize most human troops.
• Perception/Reflexes: Machines would always be "awake", and would always be on full alert, processing 360 degrees of video looking for enemies. Given halfway decent cameras, they could also do this at huge distances, much further than a human could.
• Basically Everything Else: The stuff we care about in a good soldier (strength, agility, endurance, obedience, etc) are all things where machines either currently or will soon vastly outmatch humans. It's actually the "simple" stuff where machines start to show their flaws.

Cons:

• Friend/Foe Identification: Robots don't know that the United States and Russia even exist, much less which side they're supposed to be fighting on. Trying to explain to a robot which side to shoot at, and which side to protect, is something we still haven't really figured out. Recognizing uniforms/flags makes them easy to fool, transponders are unreliable, etc.
• Unexpected Situations: Robots are terrible at adapting to the unknown. This includes repairing damage, spotting enemies disguised as friendlies or civilians, dealing with unusual terrain, etc. Reloading guns and refueling can also be surprisingly challenging, though given your society's technological advancement that part is probably figured out.
• Hacking:This is probably a big reason why the US and other advanced countries haven't mostly replaced their armed forces with robot by now. Explaining to a robot who it's supposed to be fighting for is insanely difficult, and one hacker could potentially take down an entire army's worth of robots. You'll probably have to handwave this in your setting, or just say that the robots are not wirelessly networked to prevent this sort of thing. Instead, their orders are inputted physically when they are sent on their mission. This would make them even worse at dealing with the unexpected, as their missions can't be updated, but it would prevent hackers from reprogramming them in the field.

## Very

In the Vietnam war it is thought that only "on average fewer than three [out of ten US Soldiers] ever fired their weapons in combat". If your robots fire 100% of the time that makes them around $230\%$ more efficient than the average (American*) human soldier.

Your robots don't feel emotion, they aren't scared or homesick, depressed or uncomfortable...there is plenty that might effect a human soldier's efficiency. They can leave a team mate to die, sacrifice themselves or an ally the moment it becomes tactical to do so, things the humans couldn't do...things we wouldn't want the humans to do.

*Not saying it isn't true for others but the study was based on Americans, perhaps there are cultural effects.

• If you can find the statistics for all conscripts in conflict it'll be fairly universal. Many of those that did fire, fired into the ground. – Separatrix Nov 28 '17 at 12:17
• Is the figure fewer than 3 in 10 infantrymen, or fewer than 3 in 10 soldiers? The majority of a modern army isn't made up of front line infantrymen, but support troops and logistics personnel. – walrus Nov 28 '17 at 12:57
• @Lio Elbammalf looks like the link you attach corrects the rate from 3 out of 10, to a more reasonable 8 out of 10 – famargar Nov 28 '17 at 13:10
• Things humans always do in the controlled environment of games where fear of death does not exist. – Nick Dzink Nov 29 '17 at 20:29

One possibility in here: assuming that you don't trust robots on the battlefield, use the humans (even privates) as merely commanders.

Put a soldier out on the field with a designator. Make it something easy to point, perhaps a handgun form ractor rather than a full rifle. The soldier points and shoots, "painting" the target with a rapidly-changing laser that sends out an encoded message saying "I authenticate myself as PFC Barnes, and I order you to attack this until it is dead". Any robotic platform seeing this authenticates the user and unloads on it. Add an actual handgun to the designator for last-ditch defense, and consider dropping the rifle from the human's kit.

Unless the target sees the paint signal (the soldier "paints" the optical sensor), all the gunfire (and thus attention/aggro) goes to the expendable weapons platform. This turns the human into a Forward Observer and replaces big gun artillery with expendable autonomous artillery.

• interesting comment on lazer designation! I was leaning towards a "They shoot first, metal robots deflect, we shoot back" kind of thing, but this is a very nice solution to the targeting problem. Thanks! – WorldCraftTrainee Nov 30 '17 at 7:00

It is completely viable if executed in the right way. Current drones are most commonly controlled by an off board computer making their actions easily trackable even if they are being commanded through encrypted channels. So that is one flaw that could be easily fixed by giving said human like drones on board AIs as well as a targeting system able to tell the difference between friend or foe. And many people would also say the human like shaped would hinder their abilities. But these are robots were talking about. Machines that can do whatever we design them to do. So they can carry just about any kind of weapon or equipment imaginable as long as its a size that the chassis can handle. A little food for though when thinking of possible designs for such machines. Make them smart and capable.

Forget the money angle. Forget the training angle. Forget all the angles. The only question is how do I ensure they aren't turned against me in some way. The short answer is: ya don't.

Robots aren't loyal. You programmed them to recognise enemies somehow and shoot them. All it takes if modifying either the enemy recognition part, or the decision to kill part, and suddenly your army is worse than unless.

Robots are information leaks waiting to happen. They might hold battle plans, strategies, backdoors to your communication channels. There's no telling how much information your enemy can gather from even a dead carcass. Once they have hands on your hardware, you should consider the whole system compromised.

Robots aren't less vulnerable. A localised EMP would be bad news. You might be able to blind them with a simple, dirt-cheap IR blaster. Or maybe you can lock them out by DDoSing their network interface. Being superstrong isn't being better, particularly if you can field an equally strong human with a power armour.

You don't want an army of robots that has no allegiance to you. You don't want an army of robots you can't protect from prying hands. You don't want an army of robots that only redeeming quality is brute strength.