Androids, Brain Waves and Movies, Oh My!: A Trek through the Uncanny Valley

Masahiro Mori was a robotics engineer who developed a theory based on his work. In, 1970, he found in his work and various expos like “the World Expo,” (Mori, 1970) that the more a product appeared human, the better it would be received, until a point. After some undetermined point, the familiarity would be shy of perfect, and result in disturbing product. The “product,” usually a manmade object trying to mimic factors from nature, can often be million-dollar endeavors such as animated movies and prosthetics.

Both are complex undertakings, which are affected by this theory: the Uncanny Valley –the concept that a product may seem human to such a degree that its mimicry is not exact, and by extension, upsetting. Where an animated movie may attempt some kind of accuracy, most animators avoid it all together because of the difficulty the Uncanny Valley poses. And this is just the animation aspect; prosthetics have far more fatal complications. Prosthetics have both societal and cellular rejection, posing serious health risks. Both the medical and entertainment industries have to funnel millions of dollars into making suitable products.

We Will Write a Custom Case Study Specifically
For You For Only $13.90/page!

order now

Even then, these works can still fall into the Uncanny Valley. Although advancements in technology have allowed scientists to develop realistic prosthetics and artists to mimic human physicality in animation, the resulting uncanny valley phenomenon keeps society from recognizing these advancements as positive and within the realm of a societal norm. The Uncanny Valley is caused by overstimulation of key parts of the brain under different factors such as movement, temperature and light. These factors elicit different reactions; people react differently toward movement than stills. This is first articulated by Masahiro Mori himself, who sketched a graph of his observations. As this helps explain the point, it is as follows: (Mori, 1970).

Notice the differentiation between the moving dotted line and the solid still line. The peaks and valleys are much more extreme in the moving line because theoretically, moving objects are more alarming. One clear distinction is that of the corpse versus the zombie. In most cases, they look the same. They are both dead. However, when the corpse begins to stare, move, and eat humans, it is unnerving.

Mori explains this idea as the ability for human connections. While death is unsettling, it is not uncommon. However, to challenge death the way a zombie does threatens human concepts like mortality. It is mortality which scares people, and this is the root of issues with movement. Specifically, movement can be broken down into speed and angle. These sub-factors affect the familiarity as well.

Around Mori’s time, advancements in robots were progressing –not at the rate of today, but rapid nonetheless. One example of this is a robot Mori observed at the World Expo. It had, “29 artificial muscles in the face to make humanlike facial expressions. According to the designer, laughing is a kind of sequence of face distortions, and the distortion speed is an important factor. If we cut the speed in half, laughing looks unnatural.

“(Mori, 1970) That is still true today, though the 29 muscles can now be manipulated simultaneously to a more human-like degree, businesses struggle with these same problems. Affetto, a more recent endeavor by the company Asada, shows this. By manipulating “moving actuators” Affetto can show over five different emotions (Boyle, 2011). Though meant to be a nonhuman subject for training and studying caregivers, it is a perfect example of the Uncanny Valley. The reason for this is that though the robot is advanced, not only is it still too slow to be believable, but Affetto is actually a robot baby head made of unrealistic materials. “Its eyeballs are equipped with cameras and it has microphones and tactile sensors embedded in its silicone skin, lending it a pretty realistic appearance.

” (Boyle, 2011) But “pretty realistic” is not enough to save this creation from the Uncanny Valley. If anything, it puts Affetto deeper into the valley. The dilemma with the Uncanny Valley is that the harder businesses try to improve, the deeper they sink. It is the quicksand of prosthetics and animation. But this is neither the businesses’, nor the consumers’ fault. The reason people react this way is because they are unable to understand and therefore predict the movements of the object in question.

The most remarkable ability if the brain is its ability to learn, adapt, and predict. By knowing how a person, animal, or object may act, it is easier to react. The Uncanny Valley restricts this process by presenting the observer with an unknown: something impossible to predict. This has been observed in multiple experiments. It can be concluded that, “for the Android… there was a mismatch between the human-like appearance and the mechanical motion, leading to a larger prediction error, manifest as activity in relevant brain regions.” (Saygin, et al.

, 2011) This prediction error is the measurement of this inability to predict. As the error grows larger, people become more erratic. The inability to overcome this error affects our decision making processes, and people are more likely to react dramatically and/or negatively to compensate for this instability. Even in terms of daily activities, these properties can be seen. Human’s struggle against the unknown can be simplified thusly: “They are most likely understood based on visual description of the observed events and inferences of their consequences and/or goals.” (Buccino, Lui, Canessa, Patteri, Lagravinese, Benuzzi, Porro, & Rizzolatti 2004).

The “they” is ambiguous as it applies to so many things: humans, the actions themselves, and the reactions. It is important to understand the way movement adds depth and complexity to the Uncanny Valley before this paper moves inward, as these topics resurface again and again as the paper continues. Now, to the specifics. From a prosthetics standpoint, the ability to create a usable prosthetic is relatively simple; it just has to take the place of a missing body part. However, for the recipient to accept this prosthetic mentally, for it to move and act like the part it replaces, and for it to be perceived as “normal” by onlookers, engineers have to use extreme detail.

This is best explained by Hsu (2013) “‘For example, we could be startled during a handshake by its limp boneless grip together with its texture and coldness. When this happens, we lose our sense of affinity, and the hand becomes uncanny.'” The various levels of silicone, synapses, alloys and etc. are necessary. But this is expensive and extraneous. Though Hsu is concise, it is important to see the source of this thinking.

All roads flow back to Mori, though he wasn’t a philosopher, his ideas were so profound as to be reflected on years later. His take on the prosthetic hand is based on the products he saw on a trip to Vienna. Once again, the degree of detail is astounding, but nonetheless uncanny. He theorized, “the prosthetic hand detects the current by means of electrodes and amplifies the signal to activate a small motor in the prosthetic arm to move the fingers. This hand can move in a way that causes some healthy people to feel uneasy. If you shook a woman’s hand with this hand in a dark place, the woman must be shocked!” (Mori, 1970) Though it may seem dramatic, these intense reactions aver even more visible in animation.

The textures focused on can be felt, but it can also be seen. Light plays a major role in animation, and thusly, the Uncanny Valley. Author Lawrence Weschler did extensive research on the subject in his pursuit to understand the Uncanny Valley. In a journey throughout Silicon Valley, California, he returned to write a book which succinctly explains this. “.

..if the surface is sufficiently opaque and reflective—metal, say, or plastic, or an antlike exoskeleton—the formulae and algorithms are all pretty straightforward and they don’t require all that much computer power to flesh out. Which is one reason the early breakthroughs in realistic computer animation involved plastic toys or shiny insects (or leathery dinosaurs). But photons of light behave differently upon contact with flesh… They don’t just ricochet off the surface; they penetrate, scattering about the inside there in confounding quantum fashion and then emerging at entirely different places and at all together different angles… Use the standard formulae [Newtonian billiard model] and skin will end up looking like plastic, marble like concrete, and a glass of milk like a column of chalk.

” (Weschler, 2011) While this does not pose the same physical repercussions of prosthetics, this better explain the complexities of the Uncanny Valley through something often overlooked and underestimated. In fact, the Special Interest Group on Graphics and Interactive Techniques (SIGGRAPH) holds a conference every year such ideas. While they often seem like a random assortment of graphs, they show the major qualities of light: reflectance, shadow, masking, absorption, and scattering. Such examples can be seen in the works of Pixar, where the complex algorithms and advancement create new textures and lighting. In terms of structure, complex algorithms often interact to create just one feature, like Merida from Brave’s curly hair.

Some of the senior animators involved in that movie explained it as, “…a mass spring system defined by a piecewise linear curve that captures the deformation of the hair over time… It gives artists an intuitive interface to define and modify finely detailed, complex geometry, as well as provides straightforward simulation targeting and art direction and easy integration of external forces. However, mass-spring systems have inherent limitations, most notably concerns with stability when faced with stiff systems.” (Iben, Meyer, Petrovic, Soares, Anderson, & Witkin, 2013) And that was just one feature! Imagine the complexity of a single scene.

With all of this combined, observers can see how animators manipulate light to create various dimensions. It is important to understand this difficulty and how it plays into the Uncanny Valley so it is easier to understand the effect human reaction has on this process. When a person uses their senses—sight, smell, touch, taste, and hearing—nerves send this information to the brain which then processes it and reacts. The Uncanny Valley is the brains reaction to this information when it is inhuman, so it is expected that visual and emotional processes would be activated as the brain tries to make sense of the situation. Previous experiments like those done by Ayse Pinar Saygin, Thierry Chaminade, Hiroshi Ishiguro, Jon Driver, and Chris Frith, (2011) confirm the neurological results of the Uncanny Valley in the tracked responses in the brain. In the experiment, they showed participants three clips: human, android, and robot, all performing human actions.

Because the robot is seen as a nonhuman object, there is little activity as it is easily identified and dismissed. The human video activates different parts of the brain associated with identifying other humans. Seeing as this is more difficult, it takes more brain power. Most interesting is the android clip where there is activity in all the sections seen in the robot trial and the human trial, but intensified. This is most likely the result of the brain struggling to identify if the android is human.

However, this was not the first time such results were recorded. Thierry Chaminade earlier with Jessica Hodgins, and Mitsuo Kawato (2007) found similar results while testing detail in computer animation. The idea was that in recording reactions to increasingly rendered characters, participants reacted less to the least rendered, which were merely dot frames. The reactions in this and various similar experiments continued to draw the same conclusions. The overstimulation caused by unknown objects was verified by the parts of that brain that were specifically affected.

The parts of the brain overstimulated were the superior occipital, inferior frontal, lingual, and fusiform gyri, and the head of caudate nucleus, anterior insula, temporo–occipital junction, anterior calcarine sulcus, and cerebellum (Saygin, et al., 2011). Further research showed the relevance of these section: The various gyri are responsible for different forms of facial and body recognition, the caudate nucleus focuses on response based on memory and motion, the anterior insula is responsible for self-awareness and cognition, the temporo-occipital junction (also known as the temporoparietal junction) assists in processing and comprehension, the Anterior calcarine sulcus is the home of the primary visual cortex, and the cerebellum assists motor coordination, fluidity, memory, and cognition (Myers, 2010). Overall, that means that when participants in these experiment recognized something not quite human, and then struggled to remember anything similar, whether it be their actions or others they had seen. This draws back to the earlier ideas of prediction.

The Uncanny Valley is simply an error in human cognition; fear of what people cannot understand. But imagine a world where we did understand; a world without the Uncanny Valley, allowing for the futuristic technological harmony that people often fantasize about. Concepts like household robots and more enjoyable movies would be highly attainable. Firstly, they are already all around us. Rowley, Petty, Sains, Aston, & Reinhardt (2004) define Lego Mindstorms, Blackberry’s, the “Intelligent Parking Assist System” of a Prius, TiVo, and even the microwave as robotic.

This is “because they contain sensors, microprocessors, and rudimentary artificial intelligence that allow them to do repeated tasks without human intervention” (Rowley et al. 2004). Robots have so many applications and abilities, the possibilities are endless. Household robots could assist the elderly, watch the children, do chores, etc. In terms of industry, there could be robotic secretaries, astronauts, any tasks difficult or harmful for humans. In fact, in some countries, we’ve already began to branch out in this manner, “…relying on robotic systems for repetitive chores such as vacuuming or mowing the lawn, guarding our homes, and driving our cars.

In the future, we’re betting, robots will watch over the elderly, tend to the sick, and help us stay in touch with friends and family over vast distances. With one-third of the planet expected to be 65 or older by 2050, the potential market for robotic companions is huge.” (Rowley et al. 2004). In terms of potential market, the entertainment industry would also benefit from the Movie makers can focus more on content than the visual aspect. One example of this is the film The Polar Express.

The movie had more human characters, but at the same time, many people found their soulless eyes and milky textures creepy and unnerving. This affected sales, profits, and now the styles used in that movie are not used in others (Noe 2012). Though the plot and idea behind the movie was great, it was diminished by the animation. Without this hindrance, the movie would have been a success. Besides that, eliminating the Uncanny Valley would allow artists to bush boundaries.

They could create photorealistic movies rival live action movies. Rather than trying to humanize nonhuman characters, they could portray real human effectively. Movies like Avatar are a success because they portray humanlike characters who are introduced as nonhuman. This is like when a puppy nudges an empty bowl toward you to be filled. It is surprisingly human from something that is not.

When a character is introduced as a human, it is expected to live up to a set of human expectations. This is reiterated in research that suggests ” anthropomorphic characters, for example those animated from the movements of real actors’ recorded using motion capture equipment in the movie The Polar Express (© Warner Bros Entertainment Inc.) ‘feel more uncanny’ than the stylized heroes moving unrealistically in The Incredibles (© Disney/Pixar)” (Kawato et al. 2007).

This draws back to the predictive coding mentioned earlier. Though both The Polar Express and The Incredibles both present humanlike characters, but because the Incredibles were not tethered to the normal constraints of human movement and were not expected to (Elastic-girl was far more flexible than the average human), the movie was better received. Unfortunately, the Uncanny Valley cannot be beaten because of this fact. This consistent conclusion, that the Uncanny Valley is not a conquerable concept, has been tested by various experts in the field. When asked why, these experts responded with resources, quality, and time.

Some believe that with enough money, or advanced technology, the Uncanny Valley can be beaten. One example of this is a statement made by an anonymous Pixar animator ( 2013), “Some would say that VFX can make any picture you want, it is just a matter of cost.” VFX being visual effects, the animator’s thoughts were that with human creativity and the money for the best software, processors, etc., any project would be able to best the Uncanny Valley. However, others thought we still had a ways to go. Animation companies are always trying to stay up to date when it comes to computer animating technologies, so while animators wait for the next big thing, it’s already here.

In 2007, Pixar animators Tony DeRose, Pushkar Joshi, Mark Meyer, Brian Green, and Tom Sanocki, raved about the next influential tool in animation, “Modern high-end systems, most notably SoftImage XSIR and MayaR , offer a variety of articulation methods such as enveloping [Lewis et al. 2000], blend shapes [Joshi et al. 2006], and chains of arbitrary deformations.” Then the next fad in 2011 is articulated by Kenrick Kin, Tom Miller, Bjorn Bollensdorff, Bjorn Hartmann, Maneesh Agrawala, and once again Tony DeRose, “The interface of Eden (Figure 3), our multitouch set construction application, is composed of a main view, a virtual drawer, and two columns of buttons.” And now, Pixar has developed their own software called “RenderMan.

” Though you can purchase case-by-case “Tractor” licenses and task completers for $120 so non-Pixar animators can create works of similar detail, the full software in its most basic setting is $2000 dollars. (Pixar, 2013) Even then, there are training courses which teach you how to use them. (Pixar, 2013) All of these attempts to remedy the situation, while making some money on the side, but still fall short. In regards to quality in movies, some believe that it is not just the animation that bothers people, but the story too. One of the most complete, metaphoric expressions of this is by Alva Noe (2012): A movie like The Polar Express traps us in a kind of rhetorical contradiction.

In so far as the characters resemble living human beings, we are invited to take an interest in them; we feel the impulse to watch them; we are invited to take up the stance to them that would be appropriate to live-action movies. The Polar Express is often used as an example of the Uncanny Valley for this reason. While some might argue for its plot and others for its animation, there were so many issues taken against the Uncanny Valley that various theories have been built around it. This is seen in the rest of her article, Here, then, is what I propose: the uncanny valley yawns not when animators fall short in their rendering of the human body — even if in fact they do — but rather when they get confused about what kinds of stories they are telling: Are they inviting audiences to play, or giving them an opportunity to watch? In this confusion, the story dies, and with it, the light in the eyes of the characters. As long as new theories and processes emerge centered on the Uncanny Valley, it cannot be beaten, because it cannot be forgotten. And just like in animation, even with the best technologies, prosthetics can still fall short.

This can be seen in an extended time frame, just like that of animation. Masahiro Mori (1970) noticed the limitations of advanced Viennan prosthetic fingers, “This hand can move in a way that causes some healthy people to feel uneasy.” They were so advanced that they could detect and respond to electrical currents of a recipeint’s muscles by means of amplification and small motors. Then, 40 years later, Professor Resnick (2010) of the University of Pittsburg found similar limitations, “Today, state-of-the-art prosthetics are mechanical limbs controlled by nerve impulses and microprocessors. While these enhancements can make life easier for amputees, a cutting-edge limb alone will not suffice.

” While Mori talks about a certain uneasinesss, Resnick addresses the next topic, the inflexibility of the human brain. All these obstacles may be avoidable or preventable, but that still cannot alter a human brain. Though the brain can be tricked or distracted by drugs and other stimuli, it is near impossible to change the inert mechanisms of every brain. This applies to the Uncanny Valley as it plays to some of the more primitive yet complex areas which are harder to isolate and neutralize. It also deals with prosthetics in how the recipient must mentally accept the prosthetics.

But most unconquerable concepts—death, love, and the Uncanny Valley—are all given that classic solution: time. It all comes back to time. We need time to amass the money to buy the equipment for advanced technologies. We need time for our brains to evolve into what can accept these technologies .As an anonymous Pixar animator (2013) attributed advancement to more resources, they also remarked, “I’m confident that further advances in technology will make it possible to make synthetic actors. It is anybody’s guess how long that will be.

” While the animator was quite confident, research shows that the Uncanny Valley only worsens with time. Not just in case by case exposure, but in an overall humanoid sensitivity. In the passing Uncanny element such as watching clips of the Repliee 2, shows that at firs what it was, “mistaken for a human being, but existing evidence indicates an uncanny valley experience with more prolonged exposure (Ishiguro, 2006)” (Saygin et al. 2011). This drawsback to the earlier claims of prediction guiding the Uncanny Valley.

To predict is not one isolated part of the brain. It includes visual, auditory, olfactory, and other sensory factors to build memories, patterns, and ultimately causes reactions. This is further explained by Saygin (2011), In a predictive coding account of action perception, the android is not predictable. An agent with that appearance (human) would typically not move mechanically. When the nervous system is presented with ‘the thing that should not be’ [Lovecraft, 1984 (1936); Hetfield et al., 1986], a propagation of prediction error may occur.

..” These “predictions” are what help react to situations, and these errors what cause overstimulation and overreactions. Thusly, it is our inability to predict the uncanny that causes eeriness. While this inevitability may seem like a lost cause, the Uncanny Valley is something companies should still be aware of, as the consequences of ignoring this phenomenon are severe. Movie makers go to all kinds of extremes to produce successful movies.

One example of this is Shrek, the story of a grumpy ogre who falls in love with a princess. Those who have viewed it have found it funny and light, not at all scary or peculiar as the uncanny valley suggests. That is because they changed it. Before the movie was released, like any product, it was tested on a small group. In this case, it was the expected viewers: children.

This is where they stumbled into the Uncanny Valley, “The animators of ‘Shrek’ were so good, so sophisticated, that they were scaring their intended audience. Why? The princess had fallen into what’s known as the uncanny valley,” (York, 2010). A 60 million dollar movie was almost scrapped because one character was too uncanny. In order to avoid this, they have to exaggerate every detail. The process creates waste.

In one scene of the movie Ratatouille, there can be over 500 million hairs (Ryu, 2007). Think tanks, test group, etc. are all common parts of research and development of a product. However, this is usually done for the prototype. Pass, it is produced.

Fail, there is always another project. But it is not the same for prosthetics. There are multiple materials and techniques for prosthetics that rack up costs as well. Each year there is more advancement, “such as silicone gels and urethanes, which act like a layer of fat next to the skin. These materials absorb the sheer forces that would otherwise scrape the skin. Riley says more advanced materials are being developed that will sense the electrical charge of the body and change its shape to accommodate physical fluctuations in the stump (Resnick, 2010).

” All this adds up to hundreds of millions of dollars over something that could still be rejected by the recipient. All these projects, these examples, these amazing stories, bring the future to the forefront of our imaginations. The possibilities of a world without an uncanny valley are truly astounding: artificial intelligence, household aids, extensive accepted prosthetics, and in a way, a greater social atmosphere. The uncanny valley is a physical and psychological construct built to maintain our survival. Originally, the reaction to something irregular, to even the smallest degree, could be the difference between predator and prey, as there are often species that mimic others to stay alive.

If such species were exactly identical, they would have no predators and overwhelm the population of those they mimicked. But they do not; it is this degree that our senses allow to identify. Though seemingly an exaggeration, as it is no longer used for that today, it is still an impactful factor in human interaction. The Uncanny Valley constricts our perceptions and makes it harder to be different and accepted. The same way an animated character may seem inhuman, we often disassociate from other humans for their idiosyncrasies.

“We must not see any person as an abstraction. Instead, we must see in every person a universe with its own secrets, with its own treasures, with its own source of anguish, and with some measure of triumph,” (Wiesel xi) Despite our own advancements in civil rights and tolerance, much like the constantly innovative technologies of movie animation and prosthetics, we still sink into this valley. But do not be discouraged by humanity’s flaws. Awareness of a person’s limitations is a sign of our humanity, and this humanity is our most precious quality. References Boyle, R. (2011, February 11).

‘Affetto,’ A disembodied japanese baby head, takes uncanny valley to new depths [Videoblog post]. Retrieved from Popular Science website: Buccino, G., Lui, F.

, Canessa, N., et al. (2004). Neural circuits involved in the recognition of actions performed by nonspecific: an FMRI study. Journal of Cognitive Neuroscience, 16, of 114-26.

Hery, C., & Villemin, R. (2013, July). Physically based lighting at pixar. Retrieved from Pixar Online Library database.

Hoffman, N. (2013). Background: physics and math of shading [PDF]. SIGGRAPH 2013 Conference proceedings. Retrieved from http://blog. Hollywood eyes uncanny valley in animation [Radio episode]. (2010, March 5). In J.

York (Producer), All things considered. National Public Radio. Hsu, J. (2013, November 13). Prosthetic Hands Trigger Uncanny Valley Sense of Creepiness [Blog post].

Retrieved from IEEE Spectrum website: Iben, H., Meyer, M., Petrovic, L.

, Soares, O., Anderson, J., & Witkin, A. (2013). Artistic simulation of curly hair.

Proceedings of the 12th ACM SIGGRAPH, 63-71.

2485913 Joshi, P., Meyer, M., Green, B., DeRose, T., & Sanocki, T.

(2007). Harmonic coordinates for character articulation. ACM Transactions on Graphics (TOG) – Proceedings of ACM SIGGRAPH 2007, 26(3). http://dx.doi.

org/10.1145/1276377.1276466 Kawato, M., Hodgins, J., & Chaminade, T. (2007).

Anthropomorphism influences perception of computer-animated characters’ actions [PDF]. Social Cognitive Affective Neuroscience, 2(3), 206-216.

1093/scan/nsm017 Kin, K., Miller, T., Bollensdorff, B., Hartmann, B., Agrawala, M.

, & DeRose, T. (2011). Eden: a professional multitouch tool for constructing virtual organic environments. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, 1343-1352. http://dx. Mori, Masahiro (1970) The Uncanny Valley. Energy, 7(4).

Retrieved from Myers, D. G.

(2010). Psychology (9th ed.). New York: Worth. Noe, A. (2010, January 20).

Storytelling and the ‘Uncanny Valley’ [Blog post]. Retrieved from 13.7 Cosmos & Culture website:

7/2012/01/20/145504032/story-telling-and-the-uncanny-valley Pixar. (2019, November 18). Uncanny valley project [E-mail to the author]. Rowley I., Petty A.

, Sains A., Aston A., & Reinhardt A. Ready to buy A home robot? [Blog post]. (2004, July 18).

Retrieved from Popular Science website: Resnick, B. (2010, November 3). The problem with modern-day, high-tech prosthetics [Blog post]. Retrieved from Popular Mechanics website: http://www. Ryu, D. (2007, August). 500 million and counting: hair rendering on Ratatouille. In ACM SIGGRAPH 2007 sketches (p. 51).

ACM. Sayin, A. P., Chaminade, T., Ishiguro, H.

, Driver, J., & Frith, C. (2011). The thing that should not be: Predictive coding and the uncanny valley in perceiving human and humanoid robot actions. Social Cognitive and Affective Neuroscience Advance Access, 7(4), 413-422. Weschler, L. (2011). Uncanny valley: On the digital animation of the face.

In Uncanny valley and other adventures in the narrative (pp. 1-20). Berkeley, CA: Counterpoint. Wiesel, E. (2010).

[Introduction]. In R. Skloot (Author), The immortal life of Henrietta Lacks (p. xi). New York: Crown.