Hiking the Uncanny Valley (Valerie, Natalie, David, Michael)

Hiking the Uncanny Valley

Our topic covers androids and its overlapping relationship with the uncanny valley. Through a written evaluation, we intend to gather data to determine the psychological explanations behind the reactions regarding the uncanny valley and how it relates to artificial intelligence. The “uncanny valley” is defined as a level of realism in artificial robots where humans begin to experience a feeling of unease when observing the robot, similar to the reactions of a wax figure. The differentiating factors between wax sculpture and a humanlike android is the addition of artificial intelligence and the uncanny quality of, not just the appearance, but of the movement as well.  In our experiments we hope to untangle aesthetics and intelligence in hopes of understanding them more thoroughly. Does an artificially intelligent robot pose the same threat if it does not resemble a human? Similarly, does a robot that only looks human, with no intelligence, create the same eerie feeling? 

In 1919 Sigmund Freud offered the first major attempt to explain our uncanny response toward life-like human models. Freud published an unknown concept in psychology- the “uncanny”. It was not until 1970 when Masahiro Mori, a Japanese roboticist, coined the term “Uncanny Valley” in a Japanese periodical called Energy. His hypothesis was that as AI’s become more humanlike, they would get to a point where certain facial features and gestures make them more and more eerie. Mori created a chart describing how our degree of identification and empathy with inanimate objects increases as their appearance is similar to our own. Mori’s idea was insignificant at the time. Now it’s apparent that his concept has become widespread today due to the rapid advancement in technology.

The “uncanny valley” has become a required concern for designers, psychologists, and philosophers alike. The uncanny valley has become more applicable today, and it’s incorporated in a variety of different areas that involve aesthetics. For example, the uncanny valley can become problematic for designers who work in CG animation for movies and videogames. Prosthetic hands receive the highest eeriness rankings compared to human or robotic hands, but the key observation is that the uncanny valley can be a response to any combination of expectations from the senses. For instance, some of our test subjects reacted to listening to an AI’s voice, whereas some reacted more to seeing an image of an AI.  If more reactions are generated through the sound and not the images does that mean the AI trumps the aesthetics in eeriness? Or vice versa with the images? What we found through our data was that the combination of images and sound that generated the most uncomfortable feeling in the test subjects. This shows the complex relationship that AI and the uncanny have, with the physical human likeness being only one property of the valley. We also saw in our data as well that the images take precedence over sound for eeriness. We believe this is due to a more heavy reliance on the sight sense for most humans. This is an important topic today because these technologies are being produced, whether people agree or not, and how we understand these new creations will directly impact the roles they play on our lives.

Our question surrounding the ideas of the uncanny valley and AI is; why create robots that look like humans in the first place?  In the mid 2000’s Karl MacDorman set out to discover a more in depth analysis on the uncanny valley and the different properties that contribute to it. He set up an experiment to try and better understand the workings of common feelings around the uncanny valley. He showed 56 participants 14 different videos that had a combination of humanoid, robot, and human movement. He asked each subject to rate each video on the two separate scales, strange versus familiar and mechanical versus humanlike. The results he found were somewhat surprising; there was no definitive correlation between the two concepts and the uncanny valley. MacDorman felt this was due to other factors that informed the emotions of the test subjects, which shows us that through other technological advancements, like speech and movement, the traditional uncanny valley, which mostly relies on sight, is obscured.            

We believe that one of the reasons that people feel disturbed is their potential to excel beyond human capabilities. It seems as though without human likeness, people can be more able to see that the robot is “dead” and not apart of our biosphere, but by giving it human qualities those lines become blurred. This aggressively challenges our notion of what appears living, as well as what should be included in our biosphere. Is this “more than human” attitude the actual driving force behind an uncanny response? Where we are fearful of creating something that could potentially bridge the gap between popular notions of living and non-living matter?

Works Cited

Buchanan, Bruce G."Creativity at the Metalevel: AAAI-2000 Presidential Address." Creativity at the Metalevel: AAAI-2000Presidential Address 22.3 (2001): 13-57. Ebscohost. Web. 26 Apr. 2015. The purpose of this article Bruce Buchanan explains that the key to building more creative programs is to give them the ability to reflect on and modify their own frameworks and criteria.

"CogSci-2005Workshop: Toward Social Mechanisms of Android Science." CogSci-2005 Workshop: Toward SocialMechanisms of Android Science. Trans. Karl F. MacDorman and Takashi Minato. 2005. Web. 28 Apr. 2015. This website was particularly helpful in deciphering the logistics of Masahiro Mori's Uncanny Valley Chart.

Erden, Yasemin J."Could a Created Being Ever Be Creative? Some Philosophical Remarks onCreativity and AI Development." Could a Created Being Ever Be Creative? Some Philosophical Remarks on Creativity and AI Development. Academia.edu, 24 June 2010. Web. 27 Apr. 2015. This article was used in the beginning stages of our research, when we were interested in the creativity side of Artificial Intelligence. I appreciated her stance that replication is not the best way forward for AI development in matters of creativity.

"JULES -Hanson Robotics Inc." HansonRobotics Inc RSS. Web. 25 Apr. 2015. This website was used to collect data about the Jules robot that was used in our presentation slideshow.

Katayama, Lisa."How Robotics Master Masahiro Mori Dreamed Up the 'Uncanny Valley' | WIRED." Wired.com. Conde Nast Digital, 29 Nov. 2011. Web. 27 Apr. 2015. This is an interview with Masahiro Mori discussing the origins of the uncanny valley to Wired Magazine.

Kotler, Steven."The Uncanniest Valley: What Happens When Robots Know Us Better Than WeKnow Ourselves?" Forbes. Forbes Magazine, 20 July 2014. Web. 27 Apr. 2015. This article discusses the biological reasons we are so disgusted with human-like robots. Steven Kotler explains, "To protect us from biologically costly mates, the brain’s pattern recognition has a hair-trigger mechanism for recognizing signs of low-fertility and ill-health."

Kuang, Cliff."Artificial Intelligence Is The New Uncanny Valley." Co.Design. Fast Company, 25 Apr. 2013. Web. 27 Apr. 2015. This article discusses our fascination with the aesthetic uncanny valley, and how we’ve overlooked a different, deeper uncanny valley. Cliff Kuang reiterates that it’s not the way robots look that will bother us, but how these systems act.

"Living Foreveras Robot? Prototype Lets Humans Upload Their Mind into Mechanized 'heads'"RT USA. 6 Feb. 2015. Web. 26 Apr. 2015. Bina-48 is a new concept that could potentially be the future of recording history. Bina-48 basically takes a human being's memories, thoughts and feelings into a computer to create a Mind Clone. The robot version is designed to carry on a conversation, with scientists hoping that these mind clones could give human owners a sort of artificial afterlife.

Macdorman, Karl F."Subjective Ratings of Robot Video Clips for Human Likeness, Familiarity,and Eeriness: An Exploration of the Uncanny Valley." Subjective Ratings of Robot Video Clips for Human Likeness, Familiarity, and Eeriness: An Exploration of the Uncanny Valley. Web. 26 Apr. 2015. This is just one example of an experiment done to test the uncanny in robots. This test was similar to us in that they used a variety of robots that were on different levels of the uncanny valley. he results indicate that the perceived human likeness of a robot is not the only factor determining the perceived familiarity, strangeness, or eeriness of the robot. They go on to discuss findings that suggests that other factors could be manipulated to vary the familiarity, strangeness, or eeriness of a robot independently of its human likeness.

Mitchell, Wade J.,Kevin A. Szerszen, Amy Shirong Lu, Paul W. Schermerhorn, Matthias Scheutz, andKarl F. MacDorman. "A Mismatch in the Human Realism of Face and VoiceProduces an Uncanny Valley." I-Perception. Pion, 1 Mar. 2011. Web. 26 Apr. 2015. A Mismatch in the Human Realism of Face and Voice Produces an Uncanny Valley is an experiment conducted to determine if a cross-modal mismatch in human realism causes uncertainty about whether an entity is animate or inanimate, thereby eliciting feelings of eeriness: a robot with a human voice, or a human being with a synthetic voice, will be perceived as eerier than a robot with a synthetic voice or a human being with a human voice.

"Nao: AFriendly and Interactive Robot." Aldebaran Robotics. Web. 20 Apr. 2015. This website was used to collect data about the Nao Robot that was used in our presentation slideshow.

"SeedMagazine: Into the Uncanny Valley." Into the Uncanny Valley § SEEDMAGAZINE.COM. 27 Apr. 2015. Web. 27 Apr. 2015. In this article, we pulled information regarding Sigmund Freud's contribution to the idea of the Uncanny Valley. The article explains the connection Freud makes between death and how it correlates to the Uncanny Valley.

Shwarz, Rob."10 Creepy Examples of the Uncanny Valley." Stranger Dimensions. 25 Nov. 2013. Web. 24 Apr. 2015. With this website, we used some of the robots in the article for our testing. For example, Jules and Tara the Android.

Talbot, Margaret."Pixel Perfect - The New Yorker." The New Yorker. 28 Apr. 2014. Web. 22 Apr. 2015. The New Yorker goes into depth in the science of cloning actors, for example, Tom Hanks- who was cloned for the animated film, The Polar Express. This is one medium that the uncanny dwells in, not just robotics.

"33rd Square| IBM Says Computers Will Have All Five Senses In Five Years." 33rd Square | IBM Says Computers Will Have All Five Senses In Five Years. 12 Dec. 2012. Web. 27 Apr. 2015. We're interested in the future of AI and where it could lead, in particular the senses that we aren't able to touch on. This article discusses IBM's goals with cognitive computing and how to get a computer to behave, think and interact the way humans do. Through these releases, the company points out where their researchers are focused, and what the possible outcomes of their innovations and inventions might be.

Tracey, Janey."Turing Test Alternative Lovelace 2.0 Tests AI for Original Thought."Turing Test Alternative Lovelace 2.0 Tests AI for Original Thought. Outer Places, 19 Nov. 2014. Web. 27 Apr. 2015. In our original topic research, we planned to base our project on artificial intelligence creativity. Here is one study done that rivals Alan Turing's Test, called LoveLace, 2.0. The Lovelace test aims to determine the defining characteristics of human thought in order to judge whether an artificial intelligence shares these characteristics.