What Is the Uncanny Valley?

Artist Sadie Clayton (R) is seen with "Ai-Da", the first ultra-realistic robot artist.

Leon Neal/Getty Images

What Is the Uncanny Valley?

Androids, avatars, and animations aim for extreme realism but get caught in a disturbing chasm that has been dubbed the uncanny valley. They are extremely realistic and lifelike — but when we examine them, we see they are not quite human. When a robotic or animated depiction lies in this "valley," people tend to feel a sense of unease, strangeness, disgust, or creepiness.

The uncanny valley is a term used to describe the relationship between the human-like appearance of a robotic object and the emotional response it evokes. In this phenomenon, people feel a sense of unease or even revulsion in response to humanoid robots that are highly realistic. 

You’ve probably experienced the feeling before — perhaps while watching a CGI animated movie or while playing a video game. The animated human might look almost real — but that slight chasm between looking “almost human” and “fully human” leaves you with a feeling of discomfort or even revulsion. 

The phenomenon has implications for the field of robotics and artificial intelligence. Devices and online avatars that are made to help mimic the human touch may actually end up alienating people who are using such tools. 

Origins

The term was first coined and described by the Japanese roboticist Masahiro Mori in an article published in 1970. In his work, Mori noted that people found his robots more appealing if they look more human. While people found his robots more appealing the more human they appeared, this only worked up to a certain point. 

When robots appear close but not quite human, people tend to feel uncomfortable or even disgusted. Once the uncanny valley has been reached, people start to feel uneasy, disturbed, and sometimes afraid. 

“I have noticed that, in climbing toward the goal of making robots appear human, our affinity for them increases until we come to a valley, which I call the uncanny valley,” Mori explained in his seminal paper on the topic.

Mori used a number of examples to clarify this idea. An industrial robot has little human likeness and therefore generates little affinity in observers. A toy robot, on the other hand, has a more human likeness and tends to be more appealing. A prosthetic hand, he noted, tends to lie in this uncanny valley — it can be highly lifelike yet generates feelings of unease. 

Examples

The uncanny valley has been observed in a variety of contexts, from highly realistic robots to video game characters. Some of the best-known examples of the uncanny valley can be seen on film. These include:

Final Fantasy

The 2001 movie Final Fantasy: The Spirits Within showcased some of the most realistic CGI animation ever used at that time. Despite efforts to make the animated characters appear super-realistic, the movie was a flop. The film’s commercial failure at the box office is often attributed to the uncanny valley. Simply put, people didn’t want to watch the movie because they found the animation disturbing.

Shrek

The early test screenings of the film Shrek elicited unexpected feelings of anxiety in children in response to the character Princess Fiona. She was simply too lifelike, causing kids to feel unnerved and even frightened, many crying whenever she appeared onscreen.

Based on the responses and feedback, the filmmakers edited her appearance prior to the film's theatrical release to give her a more cartoon-like appearance in order to prevent the uncanny valley effect.

Cats

The 2019 film-adaptation of Cats featured humanoid felines that many people found unsettling. While some reportedly found the effect humorous, others reported feeling downright revolted. On social media, many viewers described the film as weird, creepy, and even nightmare-inducing. 

The fact that the film relies so heavily on movement may have also played a role in the audience’s reaction. Mori has proposed that the addition of motion can amplify the uncanny valley effect, and in this case, the film includes both human-like movements as well as cat-like motions that serve to further confuse the viewer.

In these and other examples, being close to human in appearance does not produce affinity with the characters among viewers. When characters fall into the uncanny valley, people instead perceive the characters as cold, vacant, and soulless.

Possible Explanations

There have been a number of proposed explanations for why people experience the uncanny valley effect, but no clear cut consensus has emerged. Some theories suggest that the phenomenon is biological, while others suggest that there are cultural explanations as well. The following are some factors that may play a role.

Ambiguity

A 2016 paper suggested that the sensation of being “creeped out” is often caused by a sense of ambiguity. When we see things that are almost but not quite human, it creates a tension that feels unpleasant.

Such effects are sometimes exploited to heighten the horror or creepiness in movies. Horror films, for example, often infuse human characteristics into non-human entities including dolls (Annabelle) and clowns (It) to terrify audiences.

Mismatched Elements

Research has also shown that people tend to be disturbed when elements that normally don’t occur together are combined. For example, a paper published in 2011 found that while people are not disturbed by robots with “robot” voices or people with human voices, they do feel creeped out by robots with human voices. 

This effect does not just hold true for robots, but can also be seen in computer-animations feature people and animals.

Inconsistency

Even in highly realistic depictions, people are adept at spotting even the smallest inconsistencies in a robotic or animated humanoid. Even relatively minor divergences can make a character go from lifelike to uncanny.

In one of Mori’s original examples, a robotic character was suddenly seen as unlikeable and creepy when it smiled slightly too slowly.

Survival Response

Mori and others have suggested that the uncanny valley is an aversive, evolved response to the potential threats of death and disease. Because something is human-like but not quite lifelike, it may evoke the same response that people feel when they encounter something that is dead or dying.

Category Uncertainty 

Theories also suggest that the uncanny valley may exist due to the difficulty in determining what category an entity belongs to, including whether a figure is a real living human or a computer-generated three-dimensional model. Human and nonhuman represent two separate and mutually exclusive categories. 

So when something approaches a point where it seems to transition to one to the other, it can trigger feelings of cognitive dissonance. When people hold conflicting beliefs, they tend to experience feelings of psychological discomfort.

In this case, there is a conflict between the belief that an entity is human and the belief that it is not human. Something that looked human might abruptly appear nonhuman, or it may even shift back and forth as the viewer observes it.

The artificial representation is realistic enough to almost fool you into thinking it is alive, but it falls short of reality just enough that it clashes with your expectations of how a real living person would behave.

This mismatch between what you are seeing and your expectations may make you feel ambivalent or even threatened. In other words, it creeps you out.

Research

While Mori first proposed the theory in 1970, formal empirical investigations did not begin until the mid-2000s. Some research has supported the existence of the valley, although findings of how and why it happens are mixed.

Likeness and Eeriness

A 2013 study examined the relationship between human-likeness and eeriness and found evidence supporting the existence of the uncanny valley. The researchers found that there was a linear relationship between likeness and eeriness when manipulating facial proportions and realism.

Children's Responses

A 2014 study found that children between the ages of 9 and 11 were also prone to experiencing feelings of uncanniness in response to human-like virtual characters.

Virtual human-like characters were seen as being stranger and less friendly. Interestingly, these feelings of unease were more pronounced in instances where the characters lacked upper facial expression. Startled facial expressions were also more likely to be perceived as uncanny.

Other Names

While some research has supported Mori’s original hypothesis, others have characterized the phenomenon as more of a cliff or a wall than a valley. In other words, rather than ever rising up to the other side of that valley, likeability may simply drop off once a robot reaches a certain degree of realism.

Can It Be Prevented?

And not everyone agrees that the valley itself truly exists. For example, one of the earliest scientific investigations on the phenomenon was conducted in 2005 and concluded that the eerie feeling people experience has more to do with poor design and aesthetics, something that can occur at any level of realism.

While people certainly experience a sense of the uncanny in some cases, the research proposed that the valley could be overcome with good design.

Implications

The uncanny valley has a number of implications in various fields. These include:

Robotics 

As people rely more and more on robotic technology, it is important to design devices that do not create uneasiness or distrust. This is particularly true in the development of assistive technologies designed to help people with disabilities perform tasks and interact with their environments.

People are more likely to be receptive to designs that are both useful and appealing. Designs that fall into the uncanny valley are likely to be poorly received and utilized less frequently. 

Digital Avatars 

These representations are used in a number of areas including online customer service and online therapy. In the field of online therapy, digital representations are often utilized to facilitate online communication between therapists and clients, particularly in situations that involve online chat or email communication.

When they are effectively used, avatars may help promote the therapeutic relationship, but overly realistic depictions may interfere with the process.

For example, one study found that robots that look too human-like were often rated as being not only less likable but also less trustworthy, which would pose considerable problems in a therapeutic context since trust is so important to the success of treatment.

Film

As blockbuster films increasingly rely on CGI effects, filmmakers have continued to work toward developing realistic computer-generated animations that blend seamlessly and don't provoke the uncanny valley.

While many animated films are often criticized for their unrealistic depictions of the human form, such designs featuring overly large eyes and other dramatically exaggerated features, this may often be an intentional strategy to avoid the uncanny valley.

Game Design

The uncanny valley can also have an impact on how players react to realistic characters in video games. In some cases, designers may actually take advantage of the uncanny valley to create a sense of dyspathy for villainous characters.

Avoiding the Uncanny Valley

As robots become increasingly important in everyday life, researchers and designers are interested in finding ways to create tools that will not fall into the uncanny valley. This may involve making robotic devices even more realistic so that they move beyond that valley and appear more likable. 

Researchers have also proposed a number of design principles that may help animators and roboticists creating uncanny effects. This include:

  • Matching human proportions with realistic textures
  • Not mixing nonhuman and human elements
  • Ensuring that behaviors, appearance, and abilities do not conflict

Another approach is to create tools or devices that simply do not seek to mimic a human appearance. By relying on a non-human design, the device may be more appealing without running into the risk of alienating or even revolting those who interact with robotic devices.

In an interview with Wired, Mori states that while it may be possible to bridge the uncanny valley, he sees no point in trying. Instead, he advocates designing things that stop before they reach the point of uncanniness.

A Word From Verywell

There has yet to be a lot of research into the phenomenon of the uncanny valley, so more information about it, why it happens, and how to overcome it will surely emerge as time goes on. Mori himself has stated in an interview with IEEE Spectrum that his original observation was intended to be more of a guideline for designers rather than a scientific statement. 

As technology continues to advance, it possible that robots and digital animations could become so incredibly realistic that they simply blend with reality so people don't experience discomfort or anxiety. For now, robots, digital avatars, and online animations will continue to become an increasingly common part of everyday life, so it is important for designers to consider the audience's emotional response.

Was this page helpful?
Article Sources
Verywell Mind uses only high-quality sources, including peer-reviewed studies, to support the facts within our articles. Read our editorial process to learn more about how we fact-check and keep our content accurate, reliable, and trustworthy.
  1. Mori M. The Uncanny Valley (K. F. MacDorman & N. Kageki, Trans.). IEEE Robotics and Automation. 1970/2012;19(2): 98–100. doi:10.1109/MRA.2012.2192811

  2. NPR. Hollywood Eyes Uncanny Valley in Animation. Published March 5, 2010.

  3. McAndrew FT, Koehnke SS. On the nature of creepiness. New Ideas in Psychology. 2016;43:10-15. doi:10.1016/j.newideapsych.2016.03.003

  4. Mitchell WJ, Szerszen KA Sr, Lu AS, Schermerhorn PW, Scheutz M, Macdorman KF. A Mismatch in the Human Realism of Face and Voice Produces an Uncanny Valley. i-Perception. 2011;2(1):10‐12. doi:10.1068/i0415

  5. MacDorman KF, Chattopadhyay D. Reducing consistency in human realism increases the uncanny valley effect; increasing category uncertainty does notCognition. 2016;146:190‐205. doi:10.1016/j.cognition.2015.09.019

  6. Moosa MM, Ud-Dean SMM. Danger Avoidance: An Evolutionary Explanation of Uncanny Valley. Biol Theory. 2010;5:12–14. doi: 10.1162/BIOT_a_00016

  7. Burleigh TJ, Schoenherr JR, Lacroix GL. Does the uncanny valley exist? An empirical test of the relationships between eeriness and the human likeness of digitally created faces. Computers in Human Behavior. 2013;29(3):759-771. doi:10.1016/j.chb.2012.11.021

  8. Tinwell A, Sloan RJS. Childrens' perception of uncanny human-like virtual characters. Computers in Human Behavior. 2014;36:286-296. doi:10.1016/j.chb.2014.03.073

  9. Tinwell A, Grimshaw-Aagaard M, Williams A. The Uncanny Wall. International Journal of Arts and Technology. 2011;4(3):326-341. doi:10.1504/IJART.2011.041485

  10. Hanson D. Exploring the Aesthetic Range for Humanoid Robots. Published 2006.

  11. Rehm IC, Foenander E, Wallace K, Abbott JM, Kyrios M, Thomas N. What Role Can Avatars Play in e-Mental Health Interventions? Exploring New Models of Client-Therapist Interaction. Front Psychiatry. 2016;7:186. doi:10.3389/fpsyt.2016.00186

  12. Mathur MB, Reichling DB. Navigating a social world with robot partners: A quantitative cartography of the Uncanny Valley. Cognition. 2016;146:22-32: doi:10.1016/j.cognition.2015.09.008

  13. Katayama L. How Robotics Master Masahiro Mori Dreamed Up the ‘Uncanny Valley’. Wired. Published November 29, 2011.

  14. Kageki N. An Uncanny Mind: Masahiro Mori on the Uncanny Valley and Beyond. IEEE Spectrum. Published June 12, 2012.