top of page
  • LinkedIn
  • Instagram
  • Youtube

“Uncanny Valley”: Definition and Marketing Applications

Hi there! It’s Megi.





Now, let’s imagine - you sit in a bank waiting room, and a lovely little non-human-like robot comes your way and lets you know in a mechanical, robotic voice that you can go to the room where a human specialist will further process your financial request.


Now, another scenario - you sit in the same bank waiting room, but now a human-like robot (or android) comes to you, invites you to the room and gets there with you. The android further proceeds with processing your financial request under the supervision of a human in the same room. While your request is being processed, you can’t help but notice that while the android in front of you completes the task and looks like a human, it still acts not really human-like. The movements are not fluid due to mechanical joints; the face, while human-like, does not have natural facial expressions; the android doesn't blink, or it blinks in a very strange manner, in equal intervals, with no regard to external factors like blinding sunshine from the nearest window directly to their eyes. When an android asks you a question, it is formulated quite sophisticatedly; when you respond, there is a neutral expression on its face with no indication it heard you. It may be programmed to tell you a joke - but when it says it or laughs - you do not have the urge to smile in return. The smile of an android is another story - the face convulses in an attempt to imitate a human’s smile - mechanisms move behind the layer of a human-like silicone skin designed to eventually form one of a myriad of possible human smiles - friendly, shy, understanding smiles - but does it really feel like that? Or does the mechanical smile or laughter make you uncomfortable?


If we step away from the rare possibility of encountering an android in an average bank, do you feel comfortable when a human voice answers your call to the bank, and when you engage in the conversation, you realise eventually that you are talking not to a human? Do you feel frustrated when they cannot understand a number you say or your query? Or do you feel even more uncomfortable when you realise after a smooth, turn-by-turn conversation that that was actually not a human, but you thought it was? Or do you feel fascinated by the advances in technology or scared? When you are at an online retailer's website, and a chatbot appears to help you make purchase decisions and answer your questions, do you prefer the avatar to be more cartoonish and less hyperrealistic-animate? Does it, even unconsciously, make you reluctant to use the chatbot and close it for good?


If you feel uneasy and scared, or at least a little uncomfortable, this reaction is actually researched and introduced in a work by robotics professor Masahiro Mori in 1970 [1]. In his work, he discussed and hypothesised about the levels of people's emotional responses when exposed to human-like, humanoid robots.



Uncanny valley graph
Uncanny valley graph


In short, he gives an example of a prosthetic hand. When there is one not resembling a real hand, clearly made of plastic or wood or medical metal, it can be considered less human-like, and when people see it, they remain at a positive level of affinity. However, when they see a person with a very realistic prosthetic hand, one of skin-like color with simulated wrinkles, veins, and fingernails, they consider it a real hand. However, when they see it moving unnaturally or go for a handshake, and the texture and temperature of a hand are unexpected, they might get an eerie sensation when something they consider to be human is actually artificial. When this happens, people lose their sense of affinity, and the hand becomes uncanny. This a phenomenon called “uncanny valley”.


As you see on a graph, industrial robots are the ones that do not really resemble humans or humanoids; hence, they do not evoke the eerie sensation; they are just expected to complete their tasks. The same goes for toy robots; even though they might resemble humanoids, they are still considered artificial. And further is the example of the prosthetic hand I mentioned earlier, referring to the drop in the level of affinity. This way, it is considered an “uncanny valley”. Further is the point representing Bunkaru Puppet, “a traditional Japanese form of musical puppet theater dating from the 17th century”, which was used in an original work of Masahiro Mori in 1970 to exemplify the human-like but more artificially looking object, a puppet. And, on the top is the average human.


Moreover, movement amplifies the look and influences the characteristics of the human-like artificial object. It means that velocity, acceleration, and deceleration must approximate human movement. Do you remember I mentioned in the example at the beginning of the video that the android in the bank, while processing your request, would smile unnaturally and make you uncomfortable? Well, there is a similar example in the work of Professor Mori; it is about how, once in Japan, in 1970, at the World Exposition in Osaka, one robot had 29 pairs of artificial muscles in the face to make it smile in a human-like fashion. While originally designed at speed to make it realistic, the decrease in speed in half led to the face smiling not happily, as it was intended to be, but creepily. Hence, there is a thin line between the acceptable human-like movement of an artificial object and, in case of its disruption in speed, easy tumbling down into the uncanny valley.


For more details, you can check the website with the translated work of Professor Mori, I am attaching the link.



Uncanny valley graph Movement
Uncanny valley graph Movement


Okay, this is about the physical embodiment of a human-like, moving artificial object.


But what about the example of the artificial voice assistant that answered your call to the bank? Or the image of the avatar used as a “face” for the chatbot?


Well, there is a trend in research that touches upon the types and effects of using such systems in customer communications. Different types of approaches were analysed, and the results are presented in the papers, available through open access.


For example, one paper states [2] that enhancing the human likeness of a chatbot agent significantly increased users’ feeling of eeriness; the feeling of eeriness negatively influenced users’ trust in the chatbot agent; and trust, determined by the feeling of eeriness, significantly affected users’ purchase intention and willingness to reuse the chatbot.


Even though you can proceed with the search of papers on “uncanny valley”, my advice is also to check the term (or keyword) anthropomorphization, which is defined as “giving human characteristics to artificial objects” [3] and it can “strengthen brand-consumer relationships” [4] if the brand utilizes this approach [5]. This term also refers to the influence of human-like characteristics of artificial objects on customer’s perceptions. Additionally, there is also the term “uncanny valley of mind” that refers to the growth of AI and its capabilities to simulate and sense customers’ emotions.


We may dive deeper into this topic in one of the following videos.


That’s it for today! I hope it was useful.


Thank you for staying with me. It was Megi. Good luck, and see you in future videos!




[1] All the information was originally published at this source M. Mori, "The Uncanny Valley," Energy, vol. 7, no. 4, pp. 33–35, 1970 (in Japanese); translated: https://web.ics.purdue.edu/~drkelly/MoriTheUncannyValley1970.pdf


[2] Stephen Wonchul Song & Mincheol Shin (2022) Uncanny Valley Effects on Chatbot Trust, Purchase Intention, and Adoption Intention in the Context of E-Commerce: The Moderating Role of Avatar Familiarity, International Journal of Human–Computer Interaction, DOI: 10.1080/10447318.2022.2121038


[3] Beltramini E. 2019. Evil and roboethics in management studies. Artificial Intelligence and Society 34 (2): 921–929.


[4] Heine K., Atwal G., Crener-Ricard S., Phan M. 2018. Personality-driven luxury brand management. Journal of Brand Management 25 (5): 474–487.


[5] Gogua M. M., Smirnova M. M. 2020. Revisiting personalization through customer experience journey. Vestnik of Saint Petersburg University. Management 19 (4): 430–460. https://doi.org/10.21638/11701/spbu08.2020.402

Comentários


bottom of page