In Our Own Image: Do We Need Humanoid Robots?

Science fiction is full of things you don’t want to think too hard about. Why do starships with transporters have brigs with forcefields? Why not just beam a prisoner into an enclosed space?  Why do Cylons fly ships with human controls? Why not have a plug in their… well, you get the idea. For that matter, why do Cylons (and Kaylons, and Gort) even look human at all? Why aren’t some Cylons just ships?

Of course, the real reason is so we can identify with them and actors can play them with some cosplay gear and makeup. But real-life robots that are practical rarely look like humans at all.

No one is going to confuse a robot factory arm or a Roomba with a person, yet they are perfectly suited for their purpose. Yet we are fascinated with human-looking robots and continue to build them, like Nadia from IHMC Robotics in the video below.

It is something of hubris to think that the human form is ideal for everything. Look at nature. While some animals look like us, others are totally different because they fit different ecological niches. If you are going to make a smart machine, why not make it fly, jump, swim, or even stay in one place, if that’s what it needs to do? Robotic fast food fry chefs, for example, tend to ride on a ceiling-mounted rail, and why shouldn’t they? As much as we like to envision androids, an army of metallic workers asking us if we want fries with that would probably be a little unnerving.

No trouble picking out which is the robot. CC-BY-SA-4.0 by [Nicholas-halodi]
There have been calls lately to not arm robots. We had to chuckle at that because robots are already armed. What is a cruise missile or an autonomous drone with a weapon but an armed robot? Self-driving cars are weapons all by themselves. A bullet isn’t any more dangerous than getting hit by a 3,000 kg vehicle. Yet it seems that robots that look like people carrying weapons make people more uneasy than non-anthropomorphic autonomous weapons.

One use case might be in the field of human interaction. Perhaps you really do want your robot nurse or translator (C3PO) to look sort of human. A robot made to fill in for a person sometimes might need to appear somewhat human, but generally speaking, that’s not an efficient approach. (Did Star Wars get it exactly right?) What use cases can you think of for human-looking robots that aren’t pretending to be people?

Let’s face it. Robots like Sophia, ASIMO, or RoboNaut/FEDOR make good news stories. Flippy ROAR might not be as sexy, but it is a lot more practical and you are more likely to encounter one in real life.

Not that humanoid robots don’t seem cool. Want to build one? There’s an open-source one out there, at least half of one, anyway. If you want the whole thing, check out Poppy.

105 thoughts on “In Our Own Image: Do We Need Humanoid Robots?

  1. This is the difference between a tool and a servant. There is a huge amount of venture capital being poured into voice assistants and humanoid robots because the wealthy of this world would like to have their slaves back, but without all the messy parts where they demand their rights. In practice that really only works to hide the humans being exploited behind a metal mask.

    1. I suspect the wealthy AND the non-wealthy will have their robots, just like they have their cars and their 55 inch TVs right now. Not necessarily a good thing. But perhaps you can program a sentient “slave” robot for it just to be able to demand its “rights”.

      Politics, politics everywhere!

      1. Those with the cars and TVs are still the wealthy. The non-wealthy are those working in the factories producing them, conveniently out of sight. Just like they would be conveniently out of sight producing and programming the robots. All the technology does is hiding the workers conveniently out of sight of the consumers.

        And if by “politics” you mean anything inconvenient, then yes, it’s everywhere.

        1. I suppose you write on a machine made by in-plain-sight people. Please tell us the capitalistic sinners where to buy it.

          Our world is, and always will be, full of suffering. It’s the price of being free. Only in a world of ants -with its own privileged “queens” telling the “proletariat ants” what to do and what to think- everybody would be “equal”. But I won’t like to live in it.

          Your “ideal world” has other name: slavery for all.

      1. But so much of our technology and infrastructure is built for, and will continue to be built for human beings, and if you want robots to be able to easily use it a crab shape would not be the way to go. Hell people in wheelchairs have a difficult enough time with infrastructure, imagine a man sized crab?

        1. The robosimian is pretty close to a crab shape, and has no physical problems operating in human environment – in fact, the problems mostly come from programming, as there are too many ways to do things for it, and the software has to choose the best one somehow.

          1. Never seen the Robosimian before, interesting, but it couldn’t get into a small elevator, sit in a car, etc. I’m not saying all robots in the future will be humanoid any more than all animals living in a city are humans but I am saying in an environment made for humans a humanoid robot can be quite handy.

        2. We often build robots to overcome our own limitations and those are often related to physical constrains (I often wish to have third hand, another pair of thumbs or even eyes) therefore they should not be much as we are. People on wheelchair are people with more physical constrains. A crab may have some difficulties (depending on design) but octopus with palms not necessary even if every palm would have only 3 fingers. Not to mention software interface in many machines allows to control them wirelessly so even snake could drive a car while sitting in trunk.

    1. Er, no, that’s conflating causation and correlation. Quadrupeds and, yes, crabs have occurred an uncountable number of times in hundreds of millions of years across almost every animal family while the humanoid form is about a dozen species in one very small group over just the last couple of million. Our other attributes allowed us to survive in spite of it, not because of it.

      1. From a human perspective a human body is far more practical than that of a crab. Can you imagine being a human mind trapped in a crab body? Pinchers for hands? It would be almost like Edward Scissor Hands!

        We are the only species known (to us) to have created a technological society. Sure, evolution tends to lead towards the crab body plan. Evolution tends towards whatever phenotype is best at reproducing it’s genotype in a given environment. Big deal. Is that what you value most about being human? Your ability to keep the human genome going? I’ll take dextrous hands which can manipulate tools and materials over a crab’s environmental fitness advantages any day.

        Not that we are perfect. Even better body plans might be possible. For example, perhaps something like a centaur body could blend the advantages of walking on four legs with the advantages of standing upright and having dextrous hands at the ends of long arms. Such a creature might also need a somewhat larger brain to both coordinate that body and still have human intelligence.

        Then again, how does a centaur wipe it’s butt?

        A few non-primate species do also have dextrous hands. Take for example rats and racoons. Being able to climb like a racoon could be cool. Being small like a rat would be great for space travel. But none of those bodies are going to support a brain large enough for a human-like intelligence.

        As far as I can tell, a better body to inhabit has not evolved on this planet so I’m pretty glad to be human.

          1. Bird brains actually have about 4 times the density of mammal brains, if I recollect correctly. So birds could in theory have the same intelligence in a 4 times smaller brain than humans. This would be an indication that the human body does NOT have the most optimal shape and size. Our heads could theoretically be 4 times lighter. Eliminating the spinal issues that the human body designs suffers from.

            It would probably also allow for more blood flow to the brains, and allow for it to grow even more neurons, presumably giving humans even more problem-solving capabilities than we already have.

            Imagine a human brain with 4 times the amount of neurons, and thus 4 times more brain potential of what humans have today.

            Or, inversely, a human brain that’s 4 times smaller in size but has the same capacity. Humans could theoretically become 4 times smaller, which again would solve many of the problems our physical body tends to have (again, mainly problems with our spines (including our neck)).

            A smaller and more compact brain will also has shorter propagation delays because the connections between the neurons are simply shorter. So there is also a potential of increasing the speed of our brain processes.

            Of course I don’t expect that we can extrapolate it that simple. ;) But for sure there is a lot of room for improvement to the human body.

            The main problem is that we have no choice other than to wait a few hundred of thousands of years for these improvements to be ‘invented’ by nature.

            Or, we could experiment by mixing human DNA with lizard or bird DNA and immensely speed up the trial-and-error methods of Nature. But that’s a rather controversial topic. ;)

      2. “crabs have occurred an uncountable number of times in hundreds of millions of years across almost every animal family[…]”
        Erm… Unless you’re referring to crab lice, you’re talking out of your hat. Crustacea is “it’s own” family (phylum) & does not appear “across almost every animal family”. Not on this planet, anyway.

        Crustacean mammals – what a thought…

    2. True the humanoid form is pretty practical, but lets be realistic, at best humans are “okay” at a lot of things, while there are animals that are much better at specific things, but they just wouldn’t know how to handle most things we see every day. Do our robots actually need to be able to handle all those things?

      And im pretty sure that’s the very point the article is trying to make, just because the humanoid form CAN do a lot of things doesn’t mean it SHOULD & it also doesn’t mean the humanoid form does those things in the best possible way. Look at the best swimmers (fish?) best runners (quadrupeds?) best jumpers (fleas/insects?) etc, all of those are “purpose built” for that “most important task” in their lives/habitat, so they can excel at it, we as humans just stumbled around a bit for a few thousand years slowly learning to do things “okay” trough trail-and-error-repetition over MANY generations

      So yea, why would (for example) a robot chef need to be humanoid? its not gonna be practical because it doesn’t have to use stairs, open doors, run to catch the bus, go for a relaxing swim or anything like that, so it makes more sense if (like the examples from nature) its purpose built for its “most important task”, so just an arm on a rail then.

      Dont get me wrong, part of me would like to see humanoid robot servants everywhere too, but its just not logical, instead of trying to make “1 robot to rule them all” just make 200 robots for the 200 different tasks we dont wanna do ourselves, that’s going to be better, cheaper and less wasteful.

      Fast forward a few hundred years then yeah okay sure, “one humanoid to rule them all” would be cool and possible, but today? not a chance. I dont need some humanoid looking ‘all purpose’ robot to cook my burger and then slip on a slice of tomato and kill its human co-worker because oh yea oops we didnt test that and it just so happens to be a few 100 kilograms. I’ll just take the burger from the robot arm thanks.

      1. I think that one of the greatest challenges of making a “universal” robot able to operate in human environments is the fact that we are so unnaturally huge, and so the robots need to have ridiculously long reach. Building a small robot is easy, but scaling it up is hard – you get a cube law, meaning that the weight and the strength of the actuators and the strength of the materials grows proportionally to the third power of the size. Make it two times bigger and it weights eight times more, and needs to be eight times stronger.

      2. I’m pretty much in agreement, with one small exception; there’s no such thing as a robotic chef. Chefs are artists, using all their human senses to make exceptional sensory experiences out of food. Robots that cook are appliances.

        (I’m married to a chef)

        1. Chef is almost as abused a term as Engineer.

          IIRC the technical definition of Chef is ‘Someone skilled in all aspects of operating a commercial kitchen.’

          Artist isn’t in the definition. But the best are, same an Engineers. Even smaller % are good artists.

          1. “IIRC the technical definition of Chef is ‘Someone skilled in all aspects of operating a commercial kitchen.”

            Yes, that’s part of being a chef. But chefs with good accounting skills and mediocre food don’t get very far. ;-)

        2. We have AI that can create pictures, and movies, and stories, which are rapidly approaching the point where they are equivalent to those produced by a human. How long will it be before this is applied to food.

          Your artisan chef is not immune from the coming robot age.

          1. Meh… prepare for another AI winter.

            Generative AI just mixes human input and makes permutations by statistical guessing. The trick and the mechanical turk in the machine is that there’s always a human in the loop to vet the input and weed out the wrong outputs, and to pick the good results. The AI doesn’t actually make anything truly new, and since it doesn’t know what its doing (the data is meaningless to the program) it won’t do any good without human involvement.

          2. What Dude said. AI currently can only mirror what we put into into it. Granted – if you can put the collected directions and feedback from the top 400 chefs into one machine, it could probably kick out some good food. But it’s just playing back the food “recordings” from those chefs.

            The point of robotics is to take on all the menial crap (eg cleaning the kitchen) so that us meatbags may do nothing but create, experiment, and enjoy.

            Meanwhile: “Tea. Earl Grey. Hot”. I don’t think a robot could screw that up.

          3. “Meanwhile: “Tea. Earl Grey. Hot”. I don’t think a robot could screw that up.”

            You haven’t read “Hitchhiker’s Guide to the Galaxy”?
            I.e. Sirius Cybernetics Corporation

          4. KenN: “What Dude said. AI currently can only mirror what we put into into it.”

            If AI has access to knowledge, has an interface through which it can access that knowledge, and has the capability to process that knowledge, then AI can easily mirror what it puts into itself. All we need to do is to put some basic kickstart into it, after which it can theoretically teach itself anything it likes.

      3. “So yea, why would (for example) a robot chef need to be humanoid? ”

        I think that’s the wrong example. ;) Chefs also provide entertainment, in both Italian (Pizza) and Japanese (Teppan Yaki) kitchen.

        I would actually make a robot chef humanoid, if only for entertainment value.

  2. In the anime series Ghost In The Shell: Standalone complex, there’s a scene where a couple of Tachikoma–spider tanks–are talking to each other and ask this question. Why are the “Operator” gynoids built in the image of humans?

    As they discuss this, we’re treated to shots of Operators doing low level office work like making tea. The more knowledgeable Tachikoma explains to the other that they, the Tachikoma, are built for the specific task of urban paramilitary combat, and have all the necessary equipment and weapons built in, and their form reflects that. The Operators, on the other hand, are general-purpose machines made to do all kinds of work in indoor environments. Those environments, and all the tools found in them, are designed to accommodate the human form.

    So, that’s why we might want humanoid robots: because we’ve developed robots advanced enough to be *general purpose*. Flippy ROAR might be optimized for making fast food, but it’s poorly equipped to handle other tasks.

    1. That is a very common argument, and yet I find it insufficient. We live everyday lives in our “human environments” together with our dogs, cats, ferrets, and in extreme cases even children, and the seem to manage pretty well. And robots can have forms and shapes even more versatile of those of humans. A robosimian has no problem operating human tools or even driving a car. A robot that looks more like an orangutan would probably be much more efficient in human spaces than a human-shaped one. And legs, those are really only good in cramped spaces or difficult terrain, there is no reason not to have wheels in addition to legs, for moving efficiently, quietly and fast on all those flat surfaces that we tend to build.

      The real answer to the question was right there in the movie: because we want slaves who serve us tea.

      1. Toddlers definitely don’t manage that well, constantly having to pull drawers out to climb up to counters. Little people and people in wheelchairs also have issues in an environment built for a standard human.
        Wheels are great until you reach stairs, which we pesky humans tend to love adding to our buildings. The Tachikomas actually had a unique take on locomotion, legs with wheels that became feet when not in use. I believe it’s adaptations like that, that have no biological equivalent, is where the future is.

          1. The only reason they have wheels is because they can’t do passive dynamic walking, so they can’t move efficiently and have to patch up that shortcoming by “cheating”.

            Having to have both modes of locomotion is actually a handicap, because it’s more complicated and adds mass to the robot that would be avoided if they just walked better. For long distance travel on roads, ride a bicycle.

    2. Battlestar Galactica made a similar argument that the cylons were made to be able to operate human equipment. But that argument gets tenuous when there are no humans or they’re trying to destroy all the humans!

        1. I like that in the Mad Magazine parody of BSG, a Cylon warrior says “how do they ex-pect us to hit an-y thing with on-ly one eye that boun-ces back and forth a-cross our heads like a ping pong ball?”

          1. Neither of us should remember that. We could use a way to recover that storage. Tequila and high THC pot isn’t working.

            Did you know that Netmare 2 could lose a file service process when you added RAM? Do you know why? I wish I didn’t.

  3. InMoov is another cool open source humanoid project that deserves a mention, but what really deserves a mention is that all 3 (Reachy, Poppy & InMoov) come from France, i have a strong hunch there is/was some government funding in France for these kind of projects? possibly related to art and/or fashion? If anybody knows more (and/or more French open source humanoid robots?) please share! ^_^

  4. From a pure mechanical and reliability point of view, the human form is fairly inept.

    One can also question what benefits the human form would supposedly have for a given application.

    Climbing trees/cliffs is where I see the human form as decent. But our feet are fairly poorly made for this application.

    For general loitering about, a quadropedal form is advantageous due to more inherent balance and in part better load carrying capacity. (And likely why Boston Dynamic’s Spot meant for strolling about and collecting measurements isn’t bipedal.)

    From an emotional attachment standpoint. Well, people get emotionally attached to pets all the time, even stuffed toys/plushies, sometimes even homes/buildings. Not to mention long term (often hobby) projects, etc. Is any of these human in appearance? No, most often they aren’t in the slightest.

    Another problem of the more human appearance is the uncanny valley. And dipping into this generally makes people far more uncomfortable than anything. Perhaps good if one wants to scare people, but beyond that it isn’t of much use.

    In the end.
    I rather question of why one would want to make a machine in the human appearance.

    Other than for situations where one needs to augment a human, like in crash testing, or puppeteering in the entertainment sector, or make impossible stunts/scenes for movies etc. (And most of these don’t need full on augmentation either.)

    But beyond that, there is almost no practical use for it. But I guess a lot of people just lack imagination and simply copies themselves.

    1. Well, there is one huge industry that would embrace submissive realistic humanoid robots with a huge enthusiasm, and that’s the sex industry. Ironically, even in porn the purpose-build specialized robots seem more common than humanoids, though.

      1. Yeah I was about to say… this is one of the bigger reasons they’re making a push for humanoids. And they understandably don’t wanna talk about it. Rather use the PR of “butler” until they finally perfect it and put a wig and cat ears on.

    2. A quadruped like a horse can stand and loiter around more efficiently than a man – they can even sleep standing up because they can lock their knee joints – but you don’t see horses climbing trees. A monkey can climb a tree better than a human, but its hip and knee structure makes it inefficient at moving on level ground and it uses four times as much energy to walk.

      The human form is general purpose, not too optimized at the expense of other functions. We also have other adaptations that these animals don’t have, like opposable thumbs and hands free for carrying stuff, and two legs is actually more efficient than four legs to the point that people can out-walk a horse over longer distances.

      We can walk, run, swim, jump, climb, operate tools, we punch, kick, grab and throw… all the other animals do just one or two things well – we do everything. That is not ineptitude.

      1. Counter point. As a species we generally dont do any of those things well.
        Some people cant swim. Some cant fight.
        Some are olympic standard.

        The difference with animals is we have the ability to choose our environment and thus dont have to specialize if we choose to. Some tribes have lived the same way for thousands of years until we “discovered” them.
        This is due to having a larger brain and evolving beyond cave dwelling and clubbing things to a situation where food is no loner scarce, and our social structure has concepts such as bartering for our skillset.

        1. >Some people cant swim. Some cant fight.

          That’s a different argument. Of course if you eat yourself into a 400 lbs blob, then you can’t do pretty much anything. That’s not an argument against the humanoid form in general.

        2. And of choosing your environment, it’s rather that we’ve evolved to survive in any environment because we had little choice, and having done so we then spread across the entire globe.

          Even today you have people living up trees in the Brazilian jungles, or nomads trekking across deserts, and people who live by diving for fish off of floating islands made out of reeds. We can live up on top of mountains or under the ground, we hunt, forage, farm, and eat just about anything in all niches from the equator to the poles, on land and water. We’re like all animals in one.

      2. I think you forgot that fact that a robot can still have arms.

        And opposable thumbs isn’t really an all that exclusively human trait. Nor even exclusive to primates either.

        Also, out walking a horse is the wrong statement. Humans can distance wise outrun a horse. And even over exhaust one too if continuously chasing it. But pure walking distance there isn’t really a difference. It isn’t like cavalry has historically been the ones lagging behind the general army. And it isn’t like travel by horse has generally been the faster and more long distance approach to travel for centuries before more mechanized forms of travel arrived.

        “but you don’t see horses climbing trees.” Yes we don’t.
        But there is plenty of quadropedal creatures that climbs trees with rather good proficiency. Often times greatly outperforming humans at the task. The most obvious candidate here is the squirrel, but a fair few cats are on the list too, as well as a whole bunch of reptiles. However, all of these mainly have the advantage of better grip thanks to claws.

        One can argue that humans have dexterity. And this is true.
        Few animals shows similar degrees of finess when it comes to handiwork. And here birds are the current closest competitor, even as far as overall intelligence is concerned.

        But humans are still far from all that ideal.

        And if one wants to talk about efficiency of movement, then wheels are superior to legs with the only exception being terrain. However, no reason one can’t have both.

        1. > no reason one can’t have both.

          Cost, complexity, mechanical compromises to make both work, and unnecessary additional mass. Imagine yourself, but having to carry a pair of roller skates absolutely everywhere you go because they literally grow out of you.

        2. >humans are still far from all that ideal.

          Depends on what is ideal. There are many options for doing one or two things better, but then a whole lot of other things become much harder. If there is one shape that has to do it all, then humans are quite unbeatable.

          1. Specialization vs generalization. That’s ultimately what most of the arguments revolve around. Now all we need to do is deal with the costs and benefits.

  5. “Science fiction is full of things you don’t want to think too hard about. Why do starships with transporters have brigs with forcefields? Why not just beam a prisoner into an enclosed space? ”

    Plenty of good reasons for this, a brig with a forcefield is more of a fail-safe system in terms of getting the prisoner out if the technology that controls access to the cell fails (although not so fail-safe in terms of security). You’d also need to add a ventilation system for the fully enclosed cell while the forcefield seems to allow gas exchange. Also Starfleet might see putting prisoners into a claustrophobe’s worst nightmare as being potentially inhumane – the mirrorverse’s Terran Empire might see that as a feature though :-) And may disagree on the prisoner safety/security tradeoff as well…

    1. With how much they used site-to-site transport in Star Trek, plus how often entire decks are breached into space by some encounter with the Borg or whatnot, I do kind of question why they have doors at all. Why is the ship not simply a honeycomb of sealed cells, using the transporter to move things around? I guess it would be a bit inhuman and not very cinematic (no walking and talking in the corridors!)

  6. That’s precisely what got me out of the robotic research / industry:
    For context, I was student in a Japanese university more than ten years ago, in the robotics department.
    A LOT of funds went in the development of the COOL and SYFY human looking androids.
    Many of my labmates were male teenagers inspired by ghost in the shell and other animes (no a judgement, just a fact, I was one of them), so you can easily guess why they were working passionately to achieve a human looking robot (so was I).
    The official purpose, accepted by everyone was to replace human labor by robots to solve the demographic crisis without having to improve the family/work balance that could allow many women to join the workforce (because you know, women…) or opening the country to immigration (because you know immigration…)
    That was bad but not enough to stop me because human looking robots are COOL.
    Fast forward one-year later, fucking huge earthquake, nuclear reactor goes nuts. Do not worry; with all the funds that went in robotics, it will not require to sacrifice poor people to intervene…
    Not even a RC car, nothing… they had to borrow some toys to the French CEA but ultimately: poor people with gloves and shovels.
    It forged my opinion that human looking androids are just a way to burn a lot of money on (questionable) boy toys, which don’t even deliver on simple tasks or assignments and comparatively underperform massively to the other form of robots.

    But they definitely look COOL !

        1. Would that be a primary reason for dating? I tried dating. It had its good points but became too much of an energy drain and getting in the way of my major to-do lists. Besides, I’m not into asking or expecting someone else to assume a daily chore just because I hate doing it; precisely why there should be robots for those things. And rarely do I order restaurant food; tasty but often unhealthy and overpriced.

    1. Great observations.

      The movies give us humanoid robots because it’s way cheaper to stick an actor in some plastic armour than to build and operate a more practical robot simulation.

      Thinking more about this… the pursuit of humanoid robots is because we want robots to do some of the stuff we currently do… but don’t really like doing. And also because it’s a pretty obvious benchmark for how robotics is “progressing”.

  7. I think you first have to try to answer the question “What is a robot”?

    My definition has long been:
    “A robot is an assembly of mechanical and electrical components that aspires to be a life form – anything else is just automated equipment”.

    So, no, we don’t require humanoid robots, but we NEED to create humanoid robots. It is part of what drives us — essential to humanity’s need to understand our world and our place in it.

  8. If robots are to be companions (think geriatric support for example), surely having a familiar form will have a positive impact on the mental health of its owner. Let’s face it, if you are living your final days alone, wouldn’t you prefer the last pseudo-living thing by your side look human or like a blooming dalek, even if the latter was functionally fitter for purpose?
    For that reason, depending on the purpose, I think trying to give robots a humanoid form is not a bad idea. And that’s coming from an engineer who values function over form.

  9. I think there’s robot blindness, we don’t see the millions of robots already in use, and go “A humanoid? That’s the last shape you should make a robot.” Yes.. it IS the last shape, there’s a myriad of other shapes in use and we forgot they’re robots. Robot storekeeper? Vending machine, and first one of those was 2000 years ago. What about a robot that does my laundry? You’ve probably got one, replaces all the labour of pounding stuff with a rock beside the river etc.

    Anyway, in a human built world that ensures manually operable stuff is manually operable by a large number of adult humans, then a robot guaranteed to be able to operate all of that will a simulacrum of an adult human. If you’ve got a constrained use case, built it to suit that constrained use case. Just don’t call a robot general purpose and not be human shaped unless you are willing to demonstrate ad nauseum x1000 that it is not physically prevented by design from performing all the actions a human can in a human environment, whereas if it is human shaped and mimics human radius of action in all joints then it’s obvious it should be. Software/AI is another matter of course.

    1. >>“A humanoid? That’s the last shape you should make a robot.”
      You might be right but…

      “we don’t see the millions of robots already in use”
      Yeah. We’re there. Time to finish the job and build that last form maybe.

      Call it a challenge.

  10. Form to function. Seems obvious. That should be the criteria (and mostly is now). But to sell well, a cute little human like device would be much better than that more efficient spider looking thing :) .

    Often wonder why NASA doesn’t have a ‘spider’ like robot to use outside the space station to scamper all over it, fix things, checking things, and plugs in to re-power itself…. Not a robot arm that runs on a rail and limited to that area.

  11. “Why not just beam a prisoner into an enclosed space?”

    Really? You understand the reason, I bet. I’d say it’s the major enemy of security in all forms: Convenience.

    If I’m a yellow shirt I don’t want to comm or walk down to the Transporter room, I just wanna flick a switch on a wall right beside the brig.

    1. Well, for the kind where you need to walk to the pads… The brig is always closer than the transporter room?

      How’s that work?

      For the kind where you call them and they just beam you from where you are.. If you don’t want to have to call then just give your yellow shirts a pack of transponder tags. Tag ’em, step back. Count to 5 and they are gone.

  12. I know the humanoid design is hubris, but I came to the conclusion that one kind of robot has to be that way. No consumer psychology even involved.

    Hah, not the “companion” bots. A lot of us dream of a sci-fi style household helper robot. If they’re going to take out my trash, or make my bed, they will be navigating an environment designed for humans (specifically me).

    The only other design that fits those requirements (and can scrub the back of the fridge) would still have to be about my size, and able to reposition its effectors through my range of height.

    Pretty much the only alternative I can imagine looks like a forklift with n-DOF arms attached. But then you have to solve stairs, curbs, short walls, etc anyway. That’s what humanoid form factors do.

  13. It’s essential delusions of grandour. God create man in his own image, so creating robots in man’s image is the goal. Unfortunately, humans are self-repairing/ maintaining, while machines are not. The greater the complexity, the more chance of something needing repaired or replaced. It’s not bad to have dreams and goals, even if they are ubrealistic, A whole lot is learned, new devices developed to tackle some of the challenges.

    I think we need a much better battery/power source, before untethered robots will get beyond specific tasks.

  14. Humanoid robots are useful because they can rapidly take on work currently being done by humans. The task they are designed for is to augment (or replace) human labor. The most versatile way to meet that need is to design the robot in the same form as the worker currently doing the jobs. By 2030 we are likely to have over 100M humanoid robot workers. The change is coming rapidly. We should start adjusting to the new world.

        1. And 100 M more unemployed living in poverty because the society no longer has any need for them. Or did you think the world governments would jump for joy to pay for these people to simply exist?

          The problem of displacing people with robots is that the people don’t go away – they still cost a living. They will turn up in some pointless service job that makes money by generating consumption, or bureaucracy, or as professional martyrs and social justice warriors and demagogues as they seek to find some way to wring money out of the society.

          1. “By 2030 we are likely to have over 100M humanoid robot workers.”
            “And 100 M more unemployed living in poverty[…]”

            100m in toto – not additional. There are already many millions already in situ.

            As to the cost of the displaced workers, well it’s about time that the non-employed receive a decent UBI (or Universal Basic Income) – as is being trialled in Finland, IIRC..

  15. If a robot have to interact with an environment built primarily for humans, a human-like form can be meaningful. Stairs, doors, handles at the height designed for humans. Of course, in a perfect environment built for robots (e.g., a fully automated warehouse or a factory), of course, wheels or more than two legs make much more sense.

  16. Humanoid Robot: a generalized form that is always worse than the best most efficient way of doing something PLUS due to all the additional degrees of freedom of movement, a maintenance nightmare!

    Sentient Humanoid Robot: a new species deserving of all the rights of a human.

    Where is the up-side? The former just reflects a subconscious desire for servants/slaves under the misguided thought they come with less of the challenges of human servants/slaves (wrong!). The latter engender all the challenges of human servants/slaves.

    1. The issue of making a robot for each task is that you end up with a lot more robots which are not efficiently utilized because they can’t work beyond their tasks, and the maintenance nightmare isn’t reduced one bit.

      Imagine a janitor, who sweeps the floor and changes the light bulbs. Instead of one robot that does both, you have two specialized robots that can’t perform each other’s tasks because one has wheels and can’t climb a step ladder and the other isn’t built to swing a mop. The light bulb swapping robot would also be sitting idle almost all of the time.

  17. The answer is to push the boundaries of technology. While specialized AI and robots designed for specific tasks will always outclass a more generalized AI/robot in that specific task, they are more limited in handling unexpected circumstances. By striving for a more generalized system that can handle a variety of tasks and can utilize knowledge/functionality for these variety of tasks, they can adapt to handle unexpected situations better. Achievements in the research into these generalized systems can be used to improve the more specialized systems. Plus, a generalized system is more of a challenge to build than a specialized system, and engineers love challenges.

  18. Early science fiction seemed to assume robots would be human-shaped and also intelligent (The play RUR, which I think actually gave us the term “robot”, certainly did… the robots were synthetic humans not mechanical machines) . I recall a couple of stories in old Astounding issues… late 30s or early 40s (thanks!) where an inventor had the radical idea of replacing general purpose “humaniform” robots with optimized “usiform” robots. The problem was that the humaniform robots were starting to go crazy for some mysterious reason. The inventor realized that the cause was boredom… a robotic elevator operator (remember, the 30s) was a reasoning, fully mobile machine that spent all its time standing in one place just opening/closing the elevator doors and operating the elevator controls. While a human operator would do the same things, they at least had an end of shift and an opportunity to then go home and do other stuff.

    His solution was “usiform” designs meant to do one job only. The same elevator robot would be reduced to just an arm and it was much happier (it had only one appendage which it could focus on using to fulfill its purpose).

    His trouble was unsurprising: bureaucracy and public perception. Robots were “supposed’ to be humaniform and were legally required to be so. The stories (I think there were only 2) were focused on his battle to convince others of the wisdom of the concept. It didn’t matter that the usiform robots were much more content with their situation and would happily tell anyone who bothered to ask.

    The author didn’t seem to go the final step to automation: the elevator operator was still an arm that manipulated human controls and was still fully reasoning rather than an integrated device that was simply hard-wired into the elevator.

  19. Why would a robot be using a mop or climbing a step-ladder or more generally accomplish a task the way a human would?
    Why do you need a stable of robots as opposed to a robot service brought in as required?
    And why would you relegate the tasks you mentioned to a robot when there are many, many humans happy to or needing to do that work? There seems to be a mentality that sees that kind of work as menial and ‘obviously’ only done by people who are servile/servants. In fact the only thing that makes those jobs at times difficult for those doing them, is the attitude _others_ hold for that work.

  20. re: comic panel

    yeah, well i’ve seen Millennials who couldn’t figure out how to work a manual door, so…

    no, seriously; just like the T-800 — couldn’t manage whether it was push or pull.

    then again, i’ve seen people who can’t manage an automatic door. #idiocracy

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.