The year? 2050.

The United States of America, French Republic, Japan, and the Republic of the United Korea have pooled their resources and technical knowledge to solve a problem:

Armed conflict is still, unfortunately, a very real thing. While the nations decide that human soliders should still remain the on-site face of their militaries to civilians, the nations would prefer to remove soldiers from traditional roles of warfare. Why? Well humans are imprecise, introduce issues of morality, and are progressively expensive as compared to robotic alternatives. Most importantly they simply think robots can do a better job than humans can.

While a great number of military tasks and roles have been automated (such as air drones, armored vehicle drones, unmanned ships, and many aspects of logistics) the world hasn't tackled the niche human soldiers fulfil until now.

Upon the decision to move forward, Isaac Asimov began rolling over in his grave for the blatant disregard of his laws, but... times change.

While some visionaries began to create sketches of androids, the engineers of the various nations realized that there's no direct reason for robots to look human (after all, humans will still interface with civilians), and the engineers decide that they should consider alternatives.

The robots should:

  • Act semi-independently
  • Be able to be grouped effectively to increase coverage and threat to any opposition
  • Navigate human structures and devices (cars, doors, stairs, etc) capably and efficiently
  • Operate human devices (machinery, cellphones etc) as the need arises
  • Identify, engage, and kill threats as quickly, accurately, and efficiently as possible (enemy humans shouldn't have time to react before the robots take their lives, and it should be difficult for humans to gain advantage of angle of attack or element of surprise)
  • Provide ground reconnaissance in urban environments
  • Serve miscellaneous roles in their environments as needs arise (similar to the human soldiers they replace)

The nations are willing to deploy more than one design to the field simultaneously to provide specialization, but in order to keep the scope of this question sufficiently tight, answers should only provide a maximum of three designs to work together to replace human soldiers. A single design is prefered.

What design(s) and what form(s) would these soldier replacements take?

  • 2
    $\begingroup$ This site is not meant to act as an idea generation platform for authors lacking inspiration, so asking people to simply list ideas for you feels out of scope. $\endgroup$
    – AndreiROM
    Sep 27 '16 at 15:33
  • 2
    $\begingroup$ @AndreiROM I disagree. MonicaCellio once gave me the advice of (heavily paraphrased) "Provide two of the background/context, the method, and the outcome/goal. If you only provide one it's brainstorming." In this case I provide the background/context (using robots to replace human soldiers in a near-future Earth context), and the goal (a list of bullet points to achieve), but I'm struggling with the best method (design). I'm looking for a practical and efficient design, not just any design, so I'm not aiming for brainstorming. $\endgroup$
    – Ranger
    Sep 27 '16 at 15:41
  • 1
    $\begingroup$ I would submit that you are missing a key element in your design specifications: safety from hacking (prevent hacking from turning your own robots against you). Perhaps one of the reasons for Asimov's laws. $\endgroup$
    – David
    Sep 27 '16 at 15:48
  • 1
    $\begingroup$ @David Excellent consideration. $\endgroup$
    – Ranger
    Sep 27 '16 at 15:52
  • 3
    $\begingroup$ I think you're taking Monica's advice out of context. You're providing us with a list of requirements and asking us to create a design for you. This is known as idea generation, and is a close reason. If, for example, you were asking our opinion on a design of your own, that would be different. $\endgroup$
    – AndreiROM
    Sep 27 '16 at 15:52

I suppose, given your requirements, our Company CyborgCorp (CCC) can deliver what you need.

Introducing: CCC Combat Cyborg Model "C". (CCCCCC)

The new CCCCCC is a 100% human-shaped bipedal combat robot, completely with artifical skin and hair. The shape was chosen specifically with interaction on human devices in mind. Cellphones, cars, elevators and many more objects can be manipulated exactly like a human could operate them. The shape also allows to utilize human combat equipment to be used if picked up in the field. This way compatibility between our CCCCCC and regular footsoldiers is given. You do not need to deploy different weapons and equipment. The familiar shape also causes reluctant reactions and emotions of pity and mercy in enemy human soldiers, reducing their combat efficacy. Civilians feel less threatened, and tend to follow CCCCCC orders 20% better than with non-human shaped combat robots.

Of course, CCCCCC outperforms human soldiers in many ways. Beneath the SmartSkin cover, we plated the body with carbon nanotube plates, lightweight and able to resist 95% of all infantry-operated firearms. Physical capabilities are upped by over 200%, allowing sprint speeds of 50km / h, without any tiring effects, and an amazing lifting capability of almost one metric ton. Carefully hidden additional sensors allow 360 degree vision, heat vision, radar and radio wave vision. CCCCCCs senses are a manifold sharper than human senses, and especially auditory input is directly wired to a huge database, identifying enemy footsteps and gunshots, tracking possible enemies and fire arcs. All CCCCCC models are connected wirelessly, so every enemy spotted is immediately projected as a 3d holo-picture into every other CCCCCC units field of vision. if footsteps or gunshots are reported by more than one model, the enemy can be pinpointed and will be marked on the internal mapping system.

The CCCCCC comes equipped with several internal weapon systems as well as our patented MOB-EX system. MOB EX is a small, shouler mounted laser connected to an IR camera. The camera tracks the reflection of human eyes when exposed to special light, and shoots low-energy lasers at these positions, effectively blinding everyone who looks at a CCCCCC for a short duration of up to 5 minutes (permanent inquire may occur, not recommended vs civilian targets). The arm-mounted minigun and grenade launchers can achieve effective combat distances of up to 1500 meters.

Our CCCCCC are also well-suited to be used in any environment. As their sensitive systems are sealed by the SmartSkin insulation, sand, cold and heat are no problem. The system is waterproof up to 20 meters deep and can operate underwater indefinately. The battery life of our model is up to 4 hours of operation, or two hours of all-out combat operation.

We at CCC are aware that in combat, nothing ever works as expected. And you can never prepare a robot for the totally unexpected, can you? So we also offer a holo-deck mode for all our CCCCCC units. At any time, human supervisors can link up to the robot, and percieve everything the robot sees, issue new tactical doctrines, or operate the unit manually via our new VR interface. This allows our units to e.g. enter secret passcodes and operate classified keypads without your data being transfered or needed to be stored inside the unit.

We are sure our product will satisfy you completely. Please contact customer support or sales departement for further information.

Your CCC. We sell the good stuff.

  • 2
    $\begingroup$ Hm. Rubs chin thoughtfully I C... $\endgroup$
    – Ranger
    Sep 27 '16 at 15:44
  • 7
    $\begingroup$ Is the AI written using C? $\endgroup$
    – Anoplexian
    Sep 27 '16 at 19:27
  • 2
    $\begingroup$ @Anoplexian C# would be more appropriate for weaponry. $\endgroup$
    – SethWhite
    Sep 27 '16 at 21:23
  • 1
    $\begingroup$ Your robots sound very 6C. $\endgroup$
    – DeanOC
    Sep 27 '16 at 22:26
  • $\begingroup$ news.gatech.edu/2016/02/29/emergencies-should-you-trust-robot indicates that people will already follow a bucket shaped robot's orders 100% of the time even if it least to a firey death, so your extra 20% seems redundant. $\endgroup$ Dec 20 '16 at 12:54

Robots will not be replacing human soldiers on the battlefield any time soon. At least not as autonomous fighting machines.

Sure, humans are imperfect, but the answer is not to remove them from the battlefield - it's to provide them with ever more advanced equipment, such as power armor, advanced weaponry, etc.


The biggest problem you face with deploying an autonomous fighting robot is software, not hardware.

The idea of entrusting software with life and death decisions such as pulling the trigger on a human being, or launching a missile strike is absolutely blood-chilling.

Software is rarely perfect, and is especially prone to acting unexpectedly in a complex environment such as a battlefield.

Trained human soldiers have the psychological flexibility to deal with very complex, and unexpected situations, adapt, and react to threats much better than a few million lines of code written by the lowest bidder.

Furthermore, deploying an armed force consisting of robots opens you to some very nasty consequences if a successful electronic attack causes your forces to act unexpectedly.


The other shortfall of a robotic fighting force is maintenance, repairs, and resupply. A human being can scavenge his own food, fix his equipment, maintain his weapons, and organize his resources. For a robotic force to achieve this complex level of self-maintenance and independence you're basically looking at creating AI-level software. This isn't likely to happen in the next hundred years, so having that in place, tested, and entrusted to kill other people on the battlefield by 2050 is ridiculous.

At that point you're looking at deploying good ol' human mechanics and techs to take care of your robots, and that's defeating the purpose of deploying robots in the first place.

Sure, techs, mechanics, and other support troops are still needed in order to keep human soldiers in the fight, but I think you'll find that a robot is more work to take care of than a trained soldier capable of servicing his own gear.


While drones and remote control robots will likely become more common place on the battlefield, I think it's naive to imagine autonomous killer robots going into battle any time soon.

  • 1
    $\begingroup$ "The idea of entrusting software with life and death decisions such as pulling the trigger on a human being, or launching a missile strike is absolutely blood-chilling" We already entrust software to make life and death decisions all the time. We get on planes with autopilot systems and nobody really bats an eye. I get that intentionally taking a life is a little different from a system designed to protect human lives, but it is totally doable, and not particularly revolutionary $\endgroup$
    – Kevin
    Sep 27 '16 at 23:07
  • 2
    $\begingroup$ @KevinWells - you're really comparing apples to oranges, as well as missing or trivializing the moral dilemma. The autopilot on a plane, while "advanced" is only trusted to perform a fraction of a pilot's responsibilities, and there is still a full flight crew operating said autopilot. Even then, over-reliance on autopilot still leads to loss of life. Ultimately, however, that software is meant to keep a plane in the air and thus save lives, not intentionally take life. An autonomous killing machine is something completely different. $\endgroup$
    – AndreiROM
    Sep 27 '16 at 23:13

May as well make them humanoid. The requirement to be able to use "off the shelf" human equipment means that a humanoid anatomy is virtually required. For example, in order to use a cell phone you need a microphone (to hear) and a speaker (to speak) in close proximity, so may as well have a humanoid head. You will also need human-like manipulators for typing, trigger actuation, etc.

The level of fluid and rapid pattern recognition you are asking for is FAR beyond what computers can do now. So these things are, for all practical purposes, sentient AI. To be able to traverse unknown and uneven terrain, identify and engage moving targets, and use objects found in the environment quickly and independently is a massive leap in computer performance. Being able to balance on 2 legs, align arms with multiple points of articulation on to a target, and track things using a combination of active and passive sensors seems like a trivial challenge in comparison.

Don't have to mimic human internal anatomy though. The head, for example, can just be a sensor platform, perhaps with an extendable neck, to allow for more effective recon. Optics mounted around the clavicles can serve as back-up sensors should the head get destroyed. The processing, cooling, and power apparatus can be safely stored in the torso. (unlike, apparently, the Cyberdyne T-800 models).

Barring significant advances in lift technology, I don't think a flying drone will be able to carry enough weight to house the processing hardware and weapons required of a fighting unit. But there can certainly be a "mother" android that has a clutch of drones assigned to it for purposes of target location, recon, and maintaining line of sight communication with other units.

I don't think other types of locomotion (tracked, quadruped, wheels, etc) will allow for the flexibility to navigate both outside and interior spaces. Maybe a beach ball sized "torso" with a series of telescoping rods it can use to propel itself, fix itself in place, and jump around (think of a puffer fish that can push out in various directions), but I think this will be too fragile compared to just two armored lower limbs when the artillery comes in.

Your real problem is that Russia, CHINA, India, as well as some European countries will see this technological consortium as a MASSIVE threat. Warfare automation basically turns combat into an economics equation, and it is pretty easy to see where the US GDP and military spending stacks up against everyone else. Take out the risk of human loss and the threshold for "going to war" is much lower, I think.


I believe you need exactly three designs to replace humans. None of them humanoid.

Quadruped fighter

Probably based on Boston Dynamics Big Dog or Wild Cat. Fast, relatively light, all terrain, may be a stable platform for weaponry. Already can lift more than human, have lower profile and a lot of autonomous decisions.


More advanced version of existing design of tracked robot would do. Maybe with a bit of Boston Dynamics quadruped, again. You want to add manipulators for interaction with devices designed for humans. You do not want to make them drive cars or use any weapons. But connecting to computers as USB keyboards would be feasible. Even using physical ones would be.


To move cars away etc you simply need heavier units, able to push them. Be fair, most of the time it's not the job of the infantry anyway. Not now, with booby traps being common. Low centre of mass and powerful engine, and moving a car by something substantially smaller can be possible. And these should mount heavy weapons, like anti-material riffles.


I know you asked for 3. And usual scouting can be done by quadruped fighters all right. But with all these little ones developed now, it would be a waste. That's one extra, because that's something modern infantry can't do anyway. See this, this and this for examples. Small robot, sturdy and light enough to be thrown by his big brothers, hard to notice, and cheap enough to really be disposable. Jumping or flying would be a nice bonus.

These are not replacement for humans, because they do what humans don't, and already meant to supplement humans on battlefield, so I'm adding them as a little extra.


I suggest you watch Star Wars - The Clone War and take a look at the Separatist Droid Army to get the ideal.

Basically, the B1 battle droid can completely replace the need of human soldiers (I mean any "sentient species" soldiers). These really-really stupid droids can be foot soldier, tank driver, pilot, ship crew. However, they are independent thinker.

Moreover, droid can be specialized to participate in planning strategy which replace the position of human (or/and any sentient species) operation leader, for example, tactical droid

With the set-up of Separatist Droid Army, the human general only needs to give order, for example, capture target X, then droid army would be able to plan and carry out their order by themselves without any further direction from human.

  • 1
    $\begingroup$ I don't see how the Separatist Droid Army would be considered ideal, even within the Star Wars universe. The whole point of the clone army (comprised of humans) is that the clones are "immensely superior to droids, capable of independent thought and action". (That's a quote from Lama Su, one of the characters that created the clones.) $\endgroup$
    – Null
    Sep 27 '16 at 17:34
  • 1
    $\begingroup$ @Null well obviously Lama Su is a little biased. You're forgetting how effective the droids were... at selling toys. With a robot army like in Star Wars, you can fund your war with the money from merchandising. The clones can't match that. $\endgroup$
    – Cody
    Sep 27 '16 at 20:06

Not the answer you're looking for? Browse other questions tagged or ask your own question.