[Body 1] Background Summary
Brief Explanation of the Uncanny Valley
The uncanny valley is an adverse psychological reaction that is triggered by seeing products that have strong similarities to humans. The term uncanny valley was first used in the 1970s by Masahiro Mori, a Japanese robotics engineer. According to the theory behind the uncanny valley, if a product holds 60-70% similarity to the physical appearance of a human being, it can gain a favorable impression. If it goes beyond that, however, it immediately becomes unlikeable and comes off as unnatural, especially when a product is made poorly and loses its original features. The reason behind this type of reaction is because of cognitive bias. For example, when we see an industrial robot, we immediately try to search for some of the similarity it holds with humans, like “it uses two arms like a human.” With the increasing number of commonness, the level of likeness also increases. On the other hand, when we see an artificially intelligent (AI) robot that looks like a human, we start searching for our differences first. For instance, how its facial expressions are unnatural and the mouth is not synchronized when it speaks. Hence, Professor Masahiro Mori concluded that in order not to experience the uncanny valley, the product must either look totally different or more like a human. However, with the current level of technology, it is unable to imitate the physical attributes of a human perfectly.
|The Graph of Uncanny Valley (IEEE Spectrum)|
The Area Where the Uncanny Valley Arises
1. Field of Artificail Intelligence
Out of all of the artificially intelligent robots, Sophia is the most widely known. Sophia was first activated on February 14th, 2016, and was first revealed to the public in mid-March. Sophia is famous for being the first AI robot to earn citizenship in Saudi Arabia. However, she is also infamous for triggering the uncanny valley in onlookers. Sophia was invented by Hanson Robotics and was inspired by Audrey Hepburn; Hanson Robotics have imitated a total of 62 expressions of the actress. When Sophia first visited Korea in January 2018, a young girl was asked to touch Sophia and share her thoughts about the robot. After touching the robot, she replied by saying: “She's like a person”, “I'm scared.” She was conveying her natural repulsion towards the AI robot. Humans are bound to concentrate on the differences more when faced with products that resemble them, which is why missing out on small details could be a critical issue.
|The First Visit of Sophia to South Korea (The Korea Times)|
2. Artistic Field
The other field where the uncanny valley is usually triggered is in the area of art. The term uncanny valley was formerly only used for robotic engineering. After the advancement of technology, however, the term has started to be used in computer graphic animations, also. Two famous movies for triggering the uncanny valley are Beowulf and Cats.
Firstly, for the filming of Beowulf, the actors were told to have sensors attached to their bodies and went through meticulous monitoring for motion capture. Even with meticulous rendering, however, the audience received an unpleasant feeling. Manohla Dargis, a film critic, said, “To be honest, there’s no point in using motion capture when it fails to portray realistic-looking human forms. It misses out on the small details and ends up showing the cladding but not the soul of the character.”
Secondly, the movie Cats, which premiered in December of 2019, is an animated film adaptation of the musical of the same name Cats. Unlike the highly praised musical version of Cats, the animated adaptation has been faced with criticism, especially in regards to the uncanny valley. They overused the CG effects and made the characters look like hybrids of cats and humans. In the hope of sending a vivid flow of emotions to the audience, it instead brought the reverse effect of fear.
The uncanny valley is not merely the aftermath of overusing CG and other technological skills. For instance, the movie The Curious Case of Benjamin Button and the character Gollum from The Lord of the Rings were both also made by motion capture, like Beowulf. Unlike Beowulf, these films were highly praised and loved by the audience. This proves that the triggering of the uncanny valley in the artistic field is because of its lazy imitations. For example, depending on how the eyes of a character are portrayed, it can make the character more animated and can pull the audience into the character.
[Body 2] Theories Regarding the Uncanny Valley
The Theory of Thalia Wheatley
At this moment, there is no clear explanation of why such a psychological reaction occurs in humans. However, the most accepted theory to this question belongs to Thalia Wheatley, a psychologist at Dartmouth College. Her theory proposes that the human brain has a stimulation evaluation system. The stimulation evaluation system unconsciously categorizes things people see, and when it senses something that looks off or distorted, it immediately stands on alert, which makes the person wary. Due to these reasons, when people see AI robots or animations that have sloppily imitated the physical appearance of humans, they feel displeasure. Additionally, Thalia Wheatley has also stated that such a reaction is normal and is considered to be a natural survival instinct.
On the other hand, Ayse Saygin and some other robotic specialists, from the University of California, refuted Wheatley’s theory. They have pointed out that the technology is not fully developed yet, which possibly means that Wheatley’s theory has no scientific affirmation. Ayse Saygin has stated that when humans are observing characteristics, like facial expressions, our brain creates certain expectations. When it is not satisfied, dissonance happens, which causes the uncanny valley.
The Uncanny Valley and Its Psychological Relation
Humans are born with the recognition ability to guess what and who the counterpart is. An example would be how people are able to unconsciously detect the emotion of others through their facial muscle movements. Therefore, the absence of these small and detailed movements is what causes the uncanny valley.
In the case of animation, animators correct the flaws of the characters using the keyframe. A keyframe is a drawing that defines the starting and ending points of any smooth transition. For example, a character is running on the ground; however, because a proper detailed process is missing, the feet of the character might be like unnaturally placed and it will look like the feet are not touching the ground. This whole situation is viewed awkwardly by the audience and puts them off. Another instance would be when the proper human body ratio is not followed. The perfect example would be Sonic the Hedgehog, the unedited version.
|Movie ‘Sonic the Hedgehog’ Before Undergoing Editing (Entertainment Weekly)|
According to a study regarding the nerve mechanism, the reason behind people feeling repulsion towards such products is because of the cortex of vision, which handles the optical data. Additionally, the frontocortical area of the brain has a system that manages the judgement of the stimulation sent to the brain. One side of the frontocortical area judges humanity, while the other side judges the rate of likeness. Eventually, even if it has the appearance of a human, depending on how natural it acts, the rate of likeness can be different.
[Body 3] Possible Future Approaches
Reasons to Reduce the Uncanny Valley
It is important to reduce the negative effects of the uncanny valley because it can hinder the growth of technology. The uncanny valley can be an obstacle for people working in areas of artificial intelligence and the arts. To these areas, investments and support are essential for further development and research. However, the uncanny valley can disrupt the investments as it creates an unwilling sense towards future technologies. It is the reason why current scientists and psychologists are trying to reduce the uncanny valley effect.
Ways to Lessen the Psychological Reaction of the Uncanny Valley
So far, there are no accurate principles as to how the uncanny valley can be reduced, which is why the best option, for now, is to reduce the psychological reaction of the people gradually.
The most fundamental way to reduce the psychological reaction of the uncanny valley is to follow the basics. For example, proper eye contact, blinking of eyes, and the synchronization of the mouth and the words the model speaks are some factors to be noted. Solving these visual perceptions would be the first step to reduce the adverse reaction.
Additionally, according to robotic engineer David Hanson, the reason behind such a reaction could also be because of the aesthetic design. Although the facial muscles do not move as naturally, if the model is pleasing to the eyes, it can earn a favorable impression. This study is known as the theory of the Path of Engagement (POE). The theory of the POE states that if the model is well designed, it can come out of the uncanny valley as people feel more engagement than when it is poorly designed.
According to Chin Jung-kwon, a professor of German Language and Literature at Chung-ang University, he suggests taking the bull by the horns and delving deeper into the situation as a strategy can also be a way to reduce the uncanny valley.
Furthermore, in the artistic area, the uncanny valley can be used as a type of genre that the director could use. It would be another artistic expression for the director to utilize and explore. For instance, a director could perform a dystopian genre, in which the likes and dislikes are sharply divided.
Possible Future Advancement
After the reaction of the uncanny valley is reduced, it can help to bring improvements to our technological lives as the area of AI can be actively discussed and studied. For example, according to Karl MacDorman, a professor of School of Informatics and Computing at Indiana University, advanced AI could help develop the medical field, as they can act like real humans. Medical students would be able to gain more experience from having practical exercises using AI robots. In this way, students can enhance their skills and do better in real-life situations.
Also, for the artistic field, with the use of AI robots, directors would be able to save more time and shots to get the perfect action scenes as AI robots will lessen the number of “no good” (NG) cuts. They would no longer have to worry about the limit of the scenes in fear of the stunt double’s life. Currently, at Disney, robotic engineers are continuously developing AI robots under a project called the Stickman Project. The AI robots do aerobics and even fly like superman.
|Walt Disney's Stickman Project (Robohub)|
Humans are generally afraid of what is different. It is an instinct, and it is an issue that can be solved over time. Even historically, humans have fought for a long time to accept their differences. So far, it is clear the uncanny valley is another fear that humans face due to differences. Taking baby steps would be the best approach to solve the issue.
<저작권자 © THE SUNGKYUN TIMES, 무단 전재 및 재배포 금지>