Do you want your robot in human form?
I am not sure if it was a joke, but Elon Musk did it again and created hype for tech fans with the Tesla Bot presentation. So far, we only know it will be a humanoid robot, it will be used in routine tasks and some of its basic physical abilities are also shared. Although advancing in technology and consumer products is exciting, I have doubts about how Tesla pictured the future of robots. I will try to explain why I have doubts in very simple two concepts:
- People don’t need robots in human form.
- People don’t want robots in human form.
Why don’t we need human form?
Robots have been used for decades. Recently, commercial applications of robots have been increased such as drones, robotic vacuums, programmable robotic toys and so on. And it is just the tip of the iceberg. The use of robots in manufacturing and warehouses is massive compared to consumer and commercial robots. Of course, robots will be part of our daily life more in the future.
The robot is the general term for machines that operate by themselves or by remote control. This definition makes both UAV drones and Wall-E a type of robot. Social robot is a sub-category of robots where we can find few human-like robots. As they interact with humans and other robots they are categorized as social robots. They are mostly used for companion, personal assistance, telepresence, healthcare and robotics research.
If we look into the robots in the market — as they are in the market they are more than vaporware and have a potential for business success — they have an abstract appearance rather than a human form. This is because they do not need arms, legs, and head to perform their duties in stores, hospitals, and hotels. Actually, bipedal motion is mechanically more challenging to operate a robot than wheels. In the presentation, Tesla Bots are predicted to do repetitive and boring tasks. Musk added, “think about there is no shortage in labor”. But why do we want to replace industrial robots that are designed to perform their specific duties with robots in human form? For carrying an object from one place to another, warehouse robots and drones are already doing a great job. So, having hands and arms might not be good as existing solutions. Human form does not provide any advantage on performing the duty, moreover, it can decrease the performance of a robot.
Few robots have a head and face in different abstraction levels. Most of these robots are companion robots or personal assistants. The face is used to build some empathy with users. Even facial expression is a function of these robots, it is used in an abstract way rather than a real human face. Why we prefer abstract robots over humanoids is a part of the next chapter.
Why don’t we want human form?
The Uncanny Valley concept is presented in an academic article in 1970 by a Japanese robotic professor Masahiro Mori. The concept was not built on an empirical study but it was a product of Mori’s tacit knowledge and experience in the field. When human-likeness of an object (or a robot) increases, our sympathy for that entity increases as well. However, there is a point where the entity is too realistic and we feel very strange about it. This point creates the uncanny valley. While the curve goes up until the real human, it falls in too realistic and creates the valley. Mori noted that we will feel closer to a toy robot than an industrial robot because of its human-likeness, but we feel uncanny about a very realistic prosthetic hand and our sympathy increases again for a real human.
This work has not been translated into English for many years and when it does it attracted many human-robot interaction researchers. Various empirical studies confirmed this hypothesis and few studies supported the opposite view. In other words, the Uncanny Valley view is backed by many tests with users (although the idea is still open for an academic debate at different levels).
One of the main motivations behind making humanoid robots is, as humans, what we are looking for is similarities with humans in the social robots. But we feel familiar with all kinds of similarities with humans and it does not have to be in appearance. We tend to understand our surroundings with human characteristics, as this is what we know best. We see faces on clouds, we give personalities to animals, cars and even countries.
Attributing human emotions and intentions to non-humans is called anthropomorphism (anthropos = human, morphe = form) and widely examined in human-robot interaction field. Several studies on this topic found that we perceive robots that are doing expressive body movements more smart, friendly and familiar. If we think about the uncanny valley, transforming human movements into robots would be a better choice than copying human appearance to build familiarity and sympathy.
The Uncanny Valley is not the only reason why we don’t want humanoid robots. Sci-fiction movies and books influenced technology as we know today. There are people who are good at imagining things and there are people who are good at building these imagined things. The people who are good at building feel excited when they see something they can build. That is the reason why some companies tried to build flying skateboards from the iconic movie Back to the Future. Many science people were probably affected by Jules Verne’s ideas about submarines and rockets. If we go back to why we fear humanoid robots, we watched too many movies where humanoid robots were terrifying such as Terminator, Matrix, Ex Machina, and so on.
In the presentation, Musk made a joke about their robot can go at the speed of 5 mph and you can easily run away from it. He added, “hopefully that does not ever happen but you never know”. I am not blaming Musk for referring a really bad scenario for their product. Because as humans, that is what we think about humanoid robots: They are a potential threat. This is the first thing people think when they see a humanoid robot.
What about Tesla Bot?
Musk said we are the biggest robotics company in the world by mentioning the autonomous cars that they are producing. The problem here is not the ability of production and how robust your software can work but how this technology will blend into daily life. Honda announced Asimo in 2000 and it was pretty much amazing for a humanoid robot at that time. After developing it for 18 years, they decided to shut down the project. Was technology not enough at that time or was the society not ready? I don’t think what happened between Apple Newton and iPhone will be same for Honda Asimo and Tesla Bot. The technology was not mature enough to create a seamless user experience when they announced Apple Newton, so the multi-touch screen of iPhone brings success to the pocket-size computers after 14 years. Asimo, Tesla or any humanoid robot does not have a commercial application area. So, the problem is not building a robot that has better computer vision, mechanics, ability to walk in unbalanced surfaces and natural language processing. The main issue is which problems that humanoid robots are solving better, where do we need them and does it have to be in a humanoid form to perform its duties?
Tesla is an important company with many promising products. However, they have been building the hype before building the real product. Does Tesla’s humanoid robot make headlines? Yes. Does it increase company value? Probably. But when I evaluated the idea with user-centered perspective, it is easy to realize the flaws of the product and how it can fail in the market.
Today, even the biggest companies don’t put the design first in their product development process. Rather they prefer to be more technology or sales oriented. However, designers represent the users most on the table and guide the process in a human-centric way. I believe, for a successful product, the key is understanding the user. People who are good at imagining try to understand what the user might need and want, not only today but also tomorrow.
Further Reading
Bartneck, C., & Forlizzi, J. (2004). A design-centred framework for social human-robot interaction. In RO-MAN 2004. 13th IEEE international workshop on robot and human interactive communication (pp. 591–594).
Duffy, B. R. (2003). Anthropomorphism and the social robot. Robotics and autonomous systems, 42(3–4), 177–190.
Fink, J. (2012). Anthropomorphism and human likeness in the design of robots and human-robot interaction. In International Conference on Social Robotics (pp. 199–208).
Fong, T., Nourbakhsh, I., & Dautenhahn, K. (2003). A survey of socially interactive robots. Robotics and autonomous systems, 42(3–4), 143–166.
Leite, I., Martinho, C., & Paiva, A. (2013). Social robots for long-term interaction: A survey. International Journal of Social Robotics, 5(2), 291–308.
Mori, M., MacDorman, K. F., & Kageki, N. (2012). The uncanny valley [from the field]. IEEE Robotics & Automation Magazine, 19(2), 98–100.
Schmitz, M. (2010). Concepts for life-like interactive objects. In Proceedings of the fifth international conference on Tangible, embedded, and embodied interaction (pp. 157–164).
Woods, S. (2006). Exploring the design space of robots: Children’s perspectives. Interacting with Computers, 18(6), 1390–1418.
