Way back in 1919 Sigmund Freud postulated his concept of the uncanny. In the (cleverly named) The Uncanny, Freud explored a problem of aesthetics—when something is both familiar and unknown the experience of viewing it can be strongly unsettling. Fifty years later, roboticist Masahiro Mori presented his own work on the uncanny. Drawing heavily on his predecessor's work, Mori developed his "uncanny valley" hypothesis.
The deceptively simple chart progresses from the decidedly not-human (a factory robot, for instance) to something that would be indistinguishable from ourselves (whether it's a perfect computer graphic or these guys). One might guess that as the objects in questions become more human, they become more familiar to us, and so it goes. Until, that is, we hit the uncanny valley. Once the robot or graphic or toy or not-quite human creature becomes too human like (and yet not human like enough to be indistinguishable), we can't handle it. Viewing these things stimulates a feeling of repulsion (they're familiar! Yet unknown!) not readily eradicated. If we hope to make a future where robots are folded into the mainstream, the argument goes, we must be mindful these potential obstacles.
But enough theory!
In this first episode of our newest series, The Science of YouTube, John Pavlus and Christopher Mims of Small Mammal wrench open our psyches and explore the reasons why, right about now, you're feeling so unsettled.
nice one, i especially love how they accept the fear of death as the cause of uncanny valley... i know i believe that.
Maybe artificial likenesses of us are creepy not because they are man-like, but because they seem to be humans that are not man-like. I don't know if that comes off quite how I mean it. I struggled with the best way to say it for a bit.
I agree with ghost....but who came up with the name uncanny valley? it sounds ......very odd.
Maybe this finally explains why the Governator is, well,...
.. uncanny =oP
nice one perkdog
I would also like to apologize for my prior comment, the one asking who came up with the name uncanny valley?....yep, I obviously didnt read the story below....
I think there's a simpler explanation- if they're close enough to fool you into thinking they're human, but aren't good enough at it, they look like a sick/injured/mutated human, and there's a reaction to stay away to avoid that happening to you too.
I'm going to have to agree with solace and say I believe it’s instinctual for us to fear anyone who does not look of the norm, are minds could interpret them as being diseased or mutated and I’m sure we could have evolved instinctual fears over time to avoid breeding with others who resemble having these disorders.
Dolby, I would take it a step further than that. Suggesting we fear anyone who does not look the norm sounds superficial (although it is indeed true as well). In regards to robots and animation I would say as a matter of instinct we fear anyTHING that seems unnatural.
As humans, up here on the top of the food chain, we feel pretty comfortable being in control of almost any natural thing, or at least we know what to expect from it.
If something is unnatural, we don't know what to expect from it or how to control it so fear is the natural response (or in the case of not being directly threatened, discomfort or wariness).
We know many things about robots (which in turn leaves us with knowing the things we don't know =): As machines they can be made to wield great power (more than we could control if turned against flesh and bone). Most are not currently programed to have the capacity to reason like humans (as such we might be caught off guard if they are not prepared to receive our communication in a specific way leading to unnatural communication situations).
As humans we also crave a connection to an emotion or consciousness in our communication and relation to other objects. Through our knowledge of the inanimacy of an object or recognition of that consciousness or emotion, we form our reactions.
And there's always the deception factor. Until robots get to the place where we can't tell if they are human or not, we will always know there was a human behind it doing the programming and teaching. You are then presented with a creation that is made to express itself as something it is not and I believe humans are particularly wary of that as well. This is not the same as questioning the raising of a human. As humans were are taught what kind of human we should be. To teach an animate object to be something it is not (i.e. program a robot to think it is human so it plays the part well) is a form of deception that I think would cause trust issues between humans, robots and the creators.
I think Rick R. is on the right track here - our brains try to classify things, and people are one of the most important categories. So upon seeing a humanoid robot, our brains try to fit it into the category of "human". However, we are perceptive enough to notice that there is something "wrong" with this human. This makes us uneasy, for the same reasons that we instinctively fear the unknown - we don't know what to expect. Additionally, I think it triggers our distrust systems because it seems like someone's trying to "trick" us. The combination I think lends a great deal to the creepiness.
I think it something more basic that "weirds" us out when we see an artificial intelligence that looks mostly real. Maybe more like "what makes us human?" and "could humanity be artificially created?". If a robot could be truly human, we would suddenly seem to lose our uniqueness in this universe, and thereby our significance. As more and more sci-fi becomes reality, I think it leaves a creepy feeling of "if we succeed in recreating ourselves, where does that leave us?"
Humans have five basic ways to take in information. We see, hear, smell, touch or taste everything to know what we can about it. There are other things in the universe that we cannot directly detect (like radio waves). We have used our imaginations to develop ways to detect these other things through machines. If we create a humanoid robot that can think and reason like we can (or faster), does not age, is more durable, and can detect things in the universe directly that we can only detect through devices... then we have created a superior creature. If there is a creature superior to us then our species is in danger in a way that it has not been in recorded history. Instinct is often thought of in terms of a collective memory of our ancestors. Maybe this has something to do with the uncomfortability. Maybe this is a completely different uncomfortability for me.
Lifelike robots freak me out too.
Humans rely heavily on facial expressions and gestures to pick up on what a speaker is trying to infer. Have you ever noticed how distracting it can be when a person has a small tic, goofy laugh, a repeated "crutch" word, etc? Humans are such sophisticated creatures that we pick up on these things however small, and sometimes without realizing it. Take Charles Manson (ugh) for example. Would you agree that he has "crazy" eyes? It's interesting how many pick up on that, but can't quantify it. How wide would someone's eye's be before they look "crazy"? As it's said, "the eyes have it", or "the eyes are the window to the soul". If the eyes and mouth gestures are not spot on, we will notice it.
We are just too smart to be fooled by the relatively crude machines that exist today. Technology may never get there, or will have to progress light years.
I think it all boils down to environment and learned behavior. If I think back to my childhood influences, if I were to stumble across an Ewok while taking a stroll through a forest, my instinct would be to somewhat cautiously reach out and try to befriend the cute little fella. Why? Because my exposure to such a creature was one skewed toward seeing them as friendly. If you look at how robots were depicted in early Sci-Fi, they were utilitarian, subservient to humans, and in general helpful…”Danger Will Robinson!” The more recent depiction of a more humanoid robot is typically that of a futuristic bounty hunter sent to destroy the human race. I would suspect that the folks that develop and work around these humanoid robots do not find them uncanny/creepy, as they are intimately familiar with them. In contrast, the rest of us must rely on our fight or flight response system, whereby our brain must determine whether something is friend or foe, based on the information available at the moment we first encounter it. Since our view has been tainted toward the negative, fear/uncertainty is the “gut” response. I would suspect that if someone were to conduct an experiment to gauge young children’s responses to such a robot, they would not be repulsed at all. Just my $.02 hypothesis.
bec. AI will have a dominion power and some how the sci fi films affects natural human on what they see in that film
i don't know. i think it'll come down to one of two types of robots. your basic terminator/matrix bot, or your typical asimov robot. depending on their types of programming, they're either gonna try to take over by killing us all, or take over by "protecting" all of us. either way, we're screwed. as long as robotics keeps getting more detailed, that is. my question is this. WHY would you want to create a human-like robot? what is the point of it all. are we trying to play god? or what? let's leave the robots where they belong...in the factories and only "smart" enough to do a specific job. because i don't care how many safeguards you put in place, robots will eventually be smarter (i.e. more superior) than humans, and the bots will eventually realize that.
I don't actually get the "creepy vibe" that most people seem to get about those humanoids. However that may be because my mind doesn't quite work like that of a "normal" person.
Nonetheless I have to go with dolby and solace on this one. Now, I obviously don't have any research to back this up, its only my own hypothesis.
I would suspect that the more instinctive parts of the human brains identify this "person" (the humanoid machine) as somehow deformed. The visual clues are subtle, most people would probably be unable to quantify what is "wrong" with the face of a humanoid, but nonetheless instinctively identify it as being incorrect. The viewer is automatically repulsed by this deformed "person" in the same way that he is repulsed by a real human with physical deformities, for example: a person with Down's syndrome or a withered limb.
Why would the viewer be instinctively repulsed? Because such conditions of the body are usually the result of a genetic abnormality or a disease. In a more primitive era it would have been important that the deformed person not transmit his disease or pass on his (apparently) defective genes to others. So instinct teaches us to shun this unhealthy person. Survival of the fittest, in action.
While a humanoid is not actually deformed, the instinctive part of the brain identifies this "person" as "deformed" and rejects it.
I agree with engineer wife and genXer. As humans we constantly strive to do more, BE more. There is always something more to accomplish. Sometimes, our vision is clouded by dreams of greatness, of outmatching a rival until we put aside the thoughts of the complete outcome. Now, I'm not saying that everyone thinks like that because it's not true. As has been portrayed in movies such as, "I, Robot", mankind fears being conquered. Whether or not by humanoids it doesn't matter. We often take advantage of our superiority, and now that we are getting closer and closer to undistinguishable robots it's unnerving.
Just look at that movie A.I., Teddy was way cooler then any of the humanoid robots.
Just look at that movie A.I., Teddy was way cooler then any of the humanoid robots.
BTW generalsnozzie is right.
I have to disagree with this article that the robots are creepy/uncanny because they are "familiar yet unknown". I think solace hit it right on the head. Say for example that the researchers were trying to build a life-like dog or any other kind of animal. Regardless how close they got, it would still be less creepy than the woman robot showed in the video. We perceive the the wanna-be human robots as having something wrong with them.
I don't see how it is creepy either. By the way talking desktop is pretty cool