Tuesday, June 4, 2013

Life, Love and the Appreciation of Anthropomorphized Robots in Film

Brandon S. Gellis

EDPX 5000: Grad Seminar - Sonic Science Fiction  |  Professor: Trace Reddell   |  Spring 2013

Life, Love and the Appreciation of Anthropomorphized Robots in Film

What conditions and creative anecdotes exist to encourage viewers to empathize with robots in films? This research project will explore how viewers anthropomorphize and validate robot relationships, the emotions of on-screen robots and their love for one another, as seen in Wall-E [i], I, Robot [ii] and Forbidden Planet [iii]. As an exploration into the human-like sonic and visual appearance of robots, the on-screen dynamic relationship between robots and humans, robots and robots, and on-screen robots and off-screen viewers must be examined.

Filmmakers and animators use visual effects, character scores, sound effects and sonic narratives to encourage viewer empathy and anthropomorphization of human-like robots. This project seeks to analyze three primary questions: why do humans anthropomorphize robots in films, how does the degree of human-likeness affect empathy [toward robots] (Reik et al., 2009: 1)[iv], and how does that affect human viewer experience?

Subtle creative cues lead viewers to believe that on-screen robots are able to possess human-like emotions for one another, their human creators or friendly counterparts, and other robots. Human-like appearances (look/design, gestures, eye and facial expressions) and character scores or motifs (tone, voice, breathing, music and sound effects) are often created to encourage audiences to identify with specific attributes of a robot character, to evoke emotion, and aid in plot development and story telling. A 2009 paper, How Anthropomorphism Affects Empathy Toward Robots, argues:
“We found strong support for our hypothesis that people are more empathetic toward human-like robots and less empathetic toward mechanical-looking robots. This result is compatible with Simulation Theory, which states that people mentally `simulate' the situation of other agents in order to understand their mental and emotive state, and that the more similar the other agent is to the empathizer the stronger the empathy expressed. ... as the degree of anthropomorphization increases people neurologically view robots as being more like themselves.” (Reik et al., 2009: 1)[v]
Crafted very intentionally, filmmakers work to create likable, realistically (human-like) emotional visual and sonic cues and narratives to allow viewers to empathize with robots and their experiences. Even in instances where robots are villains, viewers are often caught off guard by the ease of association with such characters, which is often owed to the craftsmanship of the characters’ appearance and score.

Historically, two major films represent and initial convergence of human - CGI / animated character relationships in film. Who Framed Roger Rabbit (1988)[vi] and Cool World (1992)[vii], portray the hybridization of human and animated characters in live-action films. Each presents human - animated sexual relationships and present animated, sexualized female characters. Along side seductive score and character narratives, animators visualize sexuality as a means of provocating and encouraging viewer acceptance of this hybrid film experience.

Due to the maturation of the viewer and the acceptance of on-screen human - robot relationships, Wall-E and I, Robot were not forced to objectify the human - robot dynamic through the same sexually tinted lens. In fact, Who Framed Roger Rabbit and Cool World iconically laid necessary groundwork for the success of films as Artificial Intelligence: A.I., Terminator 2, Transformers, Wall-E and I, Robot.

A great indicator of viewer appreciation and anthropomorphization of human-like robots is a robot characters ability to express freewill. Human-like robots exemplify freewill when they begin to think for themselves and oppose their initial programming – generally speaking this occurs when someone or something he or she “cares” for is threatened. In both Wall-E and I, Robot, the lead robot characters freely accepted roles as leaders and heroes to other robots.
The notion of freewill is typified repeatedly in Wall-E, a 2008 Pixar blockbuster, set in 2805. Wall-E explores the deep issues of human neglect of Earth and the robotic-led efforts to return to Earth to restore it. Ultimately, Wall-E and his female robot counterpart, EVE, share convincingly humanistic attributes of love and affection. Both Wall-E and EVE have individual sound scores personifying and narrating their individual personalities and roles in this film. More so than any other robot, Wall-E expresses a plethora of human-like gestures, visual and sonic expressions, emotion and clarity, indicative of freewill. This film uses carefully crafted sonic and visual cues to anthropomorphize and narrate the individual characters’, emotions and the relationship of between Wall-E and EVE.

Examining “Wall-E and other Disney films, Animation Comes to Life: Anthropomorphism & Wall-E” [viii] discusses how the lack of dialogue – reminiscent of early silent films – allows viewers to empathize with the main characters. The ways in which these characters were animated and scored, harnessing “human enough” displays of emotion, enables viewers to identify with Wall-E as a hero and as more mortal than a utilitarian or mechanical robot. Wall-E permeates emotion toward EVE and when he looses that emotional connection, a kiss re-sparks his devotion and affection for EVE. Symbolically, Wall-E becomes more human as he becomes the link to return space humans to their earthly humanity.

While “Wall-E’s stature and mobile arms are similar to those of the robots in Silent Running [1972 [ix]], and his binocular shaped eyes resemble those of Johnny Five from Short Circuit [1986]”[x] (Balkind, 2010: 18), he was clearly developed to embrace robot-like characters while harnessing human-like sentimentality. Wall-E and EVE both exemplify human-like characteristics through their demeanors, gestures, eyes, personalities and character scores.

Riding on the immense success of Wall-E and The Incredibles (2004) [xi], computer-animated characters have become an everyday influence in film and popular culture, and if developed correctly positively affect viewer experiences. “Anthropomorphism influences perception of computer-animated characters’ actions”[xii] discusses how the appearance of these characters affects viewer responses to them. Lead author, Thierry Chaminade discusses the Signal Detection Theory, in which all rationalization and decisions are made with a level of uncertainty. Chaminade summarizes the experiment suggesting that by “Using a biological motion classification task, we found that the tendency to perceive a simple running motion as natural is modulated by the appearance of the character used to render the motion.” (Chaminade, 2007: 216)

The truly evocative aspect of this essay is the fine-line of how computer-animated characters based closely on real-human forms can sometimes seem both “biological” and “artificial,” and therefore may be perceived by viewers as too realistic. Ultimately, Chaminade posits that the specific graphical characteristics affect viewers’ perception and therefore empathy toward an animation, “The response bias c reflects the tendency towards answering biological [opposed to “artificial”], and indicates to what extent the subjects considered the perceived motion as natural. As the same motion sets are used to animate all characters, differences in sensitivity and response bias can only be explained taking into account the characteristics of the graphical character used to render the movement.” (Chaminade, 2007: 209)

As discussed in the Uncanny Valley[1], robots that take on “too realistic” of forms make viewers feel threatened and uncomfortable. In these instances, viewers have pleaded that robots should appear to be robots and less of a computer-animated version of a real human. “Stylized” computer-animated characters, like The Incredibles, and human-like robots, like Sonny (I, Robot, 2004)[xiii], serve as easier examples for viewers to empathize with and embrace.

2004’s science fiction, action film I, Robot, set in 2035, pervasively examines the notion of robot - human relationships. Specially designed and programmed, robots are coded with the “Three Laws of Robotics” (I, Robot, 2004):
1.     A robot may not injure a human being or, through inaction, allow a human being to come to harm.
2.     A robot must obey the orders given to it by human beings, except where such orders would conflict with the First Law.
3.     A robot must protect its own existence as long as such protection does not conflict with the First or Second Laws.
It is only after the death of Dr. Alfred Lanning, chief roboticist and creator of the NS-5 robot species, that Homicide Det. Spooner has proof to launch an investigation into the destructive intent of the robot race. 

Not only do viewers learn that NS-5 robots have the ability for emotional thought (love, hate, lust, jealousy, etc.) as seen in Sonny, Dr. Lanning’s specially designed personal robot and friend, but it is seen that a robot is responsible for the death of Dr. Lanning. As Sonny pleads with human characters to acknowledge his ability to harness human emotions, the viewer is left with no other task but to sympathize with and anthropomorphize robots in this film and recognize the subtle illumination of Sonny’s freewill.

Illustrated in a discussion between Dr. Calvin, a U.S.R. roboticist, and Det. Spooner (I, Robot, 2004: 18:05) [xiv]:
Dr. Calvin:
“…I specialize in hardware-to-wetware interfaces, in an effort to advance U.S.R.’s
robotic anthropomorphization program.”  
Det. Spooner:
“So, what exactly do you do around here?”
Dr. Calvin:
“I make the robots seem more human.”
This discussion exemplifies the recursive dynamic set into play; human viewers are watching a film, in which one is expected to empathize with and anthropomorphize Sonny, a robot character, while on-screen Dr. Calvin is discussing the intentional creation of human-like robots for the express purposes of human comfortability with the robot race.
As seen in this conversation between Dr. Calvin and Det. Spooner, the desire to make human-like robots appear as realistic and compelling as possible is an essential driver and indicator for the film industry as well. Mastering Computer Generated Imagery (CGI) to be realistic enough for modern film audiences has been a 25-year task. “Final Frontiers: Computer-Generated Imagery and the Science Fiction Film”[xv] hypothesizes that “This convergence is located within this re-conception of the body both on-screen and off, as the traditional sf [science fiction] cyborg has escaped the confines of the representational space and entered the real world of film production, where actor and computer technology are increasingly being merged into a new form of digital/human hybrid.” (Abbot, 2006: 90)
As discussed, the reason the T-1000 (Terminator 2: Judgment Day, 1991)[xvi] looked so real was in part because this was a mix of human and CGI character, and in part because the editing software used allowed for human and CGI characteristics (nose, face, eyes, etc.) to automatically align in editing. The reason CGI worked so well in I, Robot was due to the lack of bone and flesh features of Sonny. Viewers allowed themselves to identify with Sonny’s humanistic disposition because they were not made uncomfortable by “his” form.

A 2011 paper, “Humans, animals, and robots: A phenomenological approach to human-robot relations”[xvii] suggests that the way in which humans respond to and anthropomorphize pets, may be analogous to the way in which developers create and filmmakers portray human-like robots. If on-screen human-like robots are non-threatening and evoke emotions and sentiment in viewers, like pets often do, viewers will often positively empathize with these characters. Author, Mark Coeckelbergh writes, “Consider, for instance, people’s emotional responses to WALL-E, a little robot which appears as having emotions (for example feeling lonely). The fact that WALL-E only exists on the screen does not prevent people from seeing the robot as an other. … As indicated above, relations are dynamic. But the conditions for the development and persistence of the alterity relation will depend on appearance. Thus, this approach is applicable to both ‘real’ and virtual robots that we might relate to. When we encounter them in our labs, homes, games, virtual worlds, or films, what matters for our relation to them is how they appear to us as subjects-in-context. I have argued that this is analogous to what happens in human-animal relations.” (Coeckelbergh, 2011: 201)

From a more historical standpoint, social conditioning has made it so that when many humans see robots, they immediately think of human-robot relationships as portrayed in early science fiction films. Therefore, viewers often imagine robots as eager to serve humans and emulate human interactions. Films like I, Robot break through this condition and present robots and humans as “partners.” The mirror of this is that in real life, some humans interact with robots as though they are human. Again, the Uncanny Valley presents a problem for humans as robots, and images of robots, which too closely resemble humans, begin to make viewers uncomfortable and are therefore less relatable.

Outdated, but prevalent, the image of robot subservience to man is witnessed in the first science fiction film set completely in space. Forbidden Planet (1956) illustrates an adventure of human space travel and the encountering of self-proclaimed recluse, Dr. Edward Morbius, his trusty robot, Robby, and daughter, Alta. A robot-like machine, the creation of "tinkering," (Robby) provides a seemingly utilitarian purpose as an aide to Dr. Morbius. Robby, created with the inability to harm humans even if directed, however, will on command place “himself” in precarious situations. Though Robby looks like a typical robot-like machine, his personal sound effects present to viewers a significant divergence from traditional "can-like" robotic sound effects. With his appearance and the lack of emotional verbal communication, it is challenging for viewers to empathize with Robby.

Opposite of human-like robots, robots like Robby, move in more mechanical and less organic ways, distinguishing him from examples like Sonny. “Exposure to robots in movies affects our ability to empathize with them”[xviii] discusses how in humans (and monkeys) the Mirror Neuron System (MNS)[2] reacts the same when a person performs an action; the same as when a person watches someone or something else perform the same action. “They have found that MNS plays a key role in creating empathy. It allows people to transform the actions that they see others performing into internal representations.” (Staff Writer, 2010)

This article and study speak specifically to the appearance of robots and the ways in which they move. “The hand movements in each case were either smooth (human-like) or jerky and mechanical (robot-like). According to the researchers, MNS seems to respond only when robots move like robots and [virtual] humans move like humans, not otherwise.”[xix] (Staff Writer, 2010) Study observers and movie viewers alike recognize behaviors of human-like robots to mirror that of actual humans. In Robby’s case, recognized as robot-like, viewers are willing and able to identify with him, but clearly as he is; an engineered robot built to serve a utilitarian purpose. Allowing for no confusion or viewer discomfort, all of Robby’s characteristics are represented mechanically, from his voice, sound effects, appearance, movements, and lack of emotion. At best, viewers see a gender-neutral, non-threatening, subservient robot creation. Further, unlike the other robot examples discussed, Robby does not express any attributes associated with freewill.
Freewill is one of the largest contributing factors to the difference in how viewers recognize and anthropomorphize on-screen robots. Wall-E, EVE, Sonny and Robby were all designed with specific goals and objectives in mind – to better the human species. Regardless of appearance, sonic score, gestures or personality traits, each has also been formulaically created to affect viewers in intentional but different ways. All except Robby, are able to harness freewill and therefore act emotionally or human-like. More over, Wall-E, EVE and Sonny all exhibit hero-like behaviors, further differentiating them from other on-screen robots, and making them easier for viewers of all ages to empathize with.
As it’s been illustrated, filmmakers and animators use visual effects, character scores, sound effects and sonic narratives to encourage viewers to empathize with and ultimately anthropomorphize human-like on-screen robots. Human viewers anthropomorphize human-like robots in films to be more comfortable with non-threatening robots as helpful and supportive to human needs (e.g. the human - robot relationship), understand human-like relationships between robots, and appreciate robot affection for humans. So long as human-like robots do not represent specific humans (e.g. The Polar Express, 2004[xx]) or are created too closely resembling humans (The Uncanny Valley[xxi]), viewers are able to distinguish between humans and robots, and can appreciate the dynamic place, which robot characters hold within the film industry.

[1] The Uncanny Valley (Bukimi no tani), originated by M. Mori (1970) is the principle that if a robot too closely represents true human form, that viewers are repulsed by or fearful of its existence, rather than being able to identify with it.
[2] The mirror neuron system (MNS) is a collection of neurons in various parts of the brain, including the premotor cortex and the primary motor cortex, which allow humans to recognize similar and differing actions.

[i] Wall-E. Motion Picture. Directed by Andrew Stanton. 2008; Emeryville, CA: Pixar Animation Studios.
[ii] I, Robot. Motion Picture. Directed by Alex Proyas. 2004; Hollywood, CA: 20th Century Fox.
[iii] Forbidden Planet. Motion Picture. Directed by Fred M. Wilcox. 1956: Hollywood, CA: Metro-Goldwyn-Mayer.
[iv] Riek, Laurel D., et al. "How anthropomorphism affects empathy toward robots." Proceedings of the 4th ACM/IEEE international conference on Human robot interaction. ACM, (2009).
[v] Riek, Laurel D., et al. "How anthropomorphism affects…” (2009).
[vi] Who Framed Roger Rabbit. Motion Picture. Directed by Robert Zemeckis. 1988; Los Angeles, CA: Touchstone Pictures.
[vii] Cool World. Motion Picture. Directed by Ralph Bakshi. 1992; Culver City, CA: Paramount Pictures.
[viii] Balkind, Nicola. “Animation Comes to Life: Anthropomorphism & Wall-E,” Film International. http://filmint.nu/wp-content/uploads/2010/06/FINT_web_june_Balkind.pdf (2010).
[ix] Silent Running. Motion Picture. Directed by Douglas Trumbull. 1972; Toluca Lake, CA: Universal Pictures.
[x] Short Circuit. Motion Picture. Directed by John Badham. 1986; Los Angeles, CA: TriStar Pictures.
[xi] The Incredibles. Motion Picture. Directed by Brad Bird. 2004; Burbank, CA: Walt Disney Pictures.
[xii] Chaminade, Thierry, Jessica Hodgins, and Mitsuo Kawato. "Anthropomorphism influences perception of computer-animated characters’ actions." Social cognitive and affective neuroscience, 2.3 (2007): 206-216.
[xiii] I, Robot, 2004.
[xiv] I, Robot, 2004.
[xv] Abbott, Stacey. “Final Frontiers: Computer-Generated Imagery and the Science Fiction Film”. Science Fiction Studies, (2006) #98 = Volume 33, Part 1.
[xvi] Terminator 2: Judgment Day. Directed by James Cameron. 1991; Los Angeles, CA: TriStar Pictures.
[xvii] Coeckelbergh, Mark. "Humans, animals, and robots: A phenomenological approach to human-robot relations." International Journal of Social Robotics 3.2 (2011): 197-204.
[xviii] Staff Writer, “Exposure to robots in movies affects our ability to empathize with them,” Asian News International [New Delhi], January 23, 2010.
[xix] Staff Writer, “Exposure to robots…” 2010.
[xx] The Polar Express. Motion Picture. Directed by Robert Zemeckis. 2004; Los Angeles, CA: Warner Bros.
[xxi] The Uncanny Valley (Bukimi no tani). Originated by M. Mori, 1970.


Accompanying this research project is an additional creative research component – Space, Robots and the Voices We Hear – that has taken the form of an audio and video re-mastering. This additional element seeks to explore how viewer responses may change if character sound scores are switched or redubbed, between films. The intention has been to re-master the existing robot visual appearance of one film with the character score from another.

Utilizing the three primary films discussed in this project, Wall-E, I, Robot and Forbidden Planet, Robby has been re-mastered by EVE, Sonny has been re-mastered by Robby and EVE has been re-mastered by Sonny. In all film instances, only the robot has been re-mastered, and in the case of Wall-E, only EVE has been re-mastered, to maintain the essence of the original scene. In order to highlight the, often, confusing dynamic of human – robot and robot – robot relationships, splices of audio have been re-mastered over other characters, without dialogue being re-recorded exactly as it originally occurred. As an interesting discovery, Wall-E and EVE, and even Robby’s character scores are extremely limited making it challenging to mimic or reproduce their original theatrical context.