Please take one minute to fill out our MIT Docubase user survey x

_The Limits of Virtual Reality: Debugging the Empathy Machine

Case Study of The Machine to be Another

by Ainsley Sutherland

“If you grip the handles, sir, you will… share his suffering, as you know, but that is not all. You will also participate in his… World-view’ is not the correct term. Ideology? No. […] No word will do, and that is the entire point. It cannot be described — it must be experienced.”

The Little Black Box, by Philip K. Dick (Published in Worlds of Tomorrow, 1964)

The Machine to be Another is a performance piece that could almost be drawn from the eerily prescient world of 1960s science fiction. Outfitted with an Oculus Rift, headphones, and seated safely in a chair, participants swap perceptual systems with a performer. When I participated in the experience, I was guided blindfolded into a room. Opening my eyes, I found myself staring in a mirror at someone else’s face. As I slowly lifted my hands, I felt an eerie frisson at the minute delay between thought and action, a feeling that intensified when I ran my fingers over a ring I could feel and knew I wore, but could not see.

Through this experience, the designers and their collaborators hope to build empathy in participants. This impulse is widespread  in virtual reality, an increasingly popular medium with the advent of Oculus Rift Developer Kits: In the video Zero Point, billed as the “first 3D 360 short film created for virtual reality”, the makers suggest that VR has a unique capacity to put a viewer “in someone else’s shoes”. “Perspective: Chapter 1 The Party”, which recently premiered at Sundance utilizes the first-person perspective of VR to show viewers the “two different stories” that emerge when an unconscious woman is raped at a frat party. At Stanford’s Virtual Human Interaction Lab, a key research goal is “empathy at scale”. Excitement about the discovery of so-called “mirror neurons” lends a sense of possible increases in scientific rigor, and experiments in body visualization and “perspective-taking” have been met with positive results in medical studies.1 But can the immersive perspective provided by virtual reality actually provoke increases in empathy? Empathy has a long history as a concept bridging aesthetics, psychology, and medicine. Artwork in this space occupies a growing territory, borrowing from scientific discourse to make an artistic intervention.

Be Another Lab, the creative collective behind Machine, is not necessarily interested in the complete reconstruction of the subjective experience of “the other”, nor only in straightforward social perspective-taking. In this performance piece, they explore how the boundaries of the “self” can be manipulated and perhaps expanded.

This case study describes the Machine to be Another project, examines the role of empathy in this work, reactions to the project– including my own– and discusses the technologies involved in its implementation. Some guiding questions are:

1) how can artistic intervention relate to discoveries in psychology and biology, and

2) what effects are obscured when “empathy” is used as a cover-all description of impact?

Project Description

Unlike self-contained VR projects, where the visual, environmental, and narrative experience is complete and is similar to a game or video, The Machine to be Another can be more accurately described as a protocol (or “embodiment system”). This means that there is no virtual world or 3d models, no prerecorded 360 video. “Embodiment system” is the creative term that describes how the experience is a formalized, structured interaction that has some cues, some technological aids, but is not fully scripted. It is reminiscent of the scripts for “happenings” created by artist Alan Kaprow in the 1950s and 60s: structure and conditions are planned, but the experience is open to improvisation and requires creative audience invention. It resonates with the theatrical tradition of Augusto Boal’s Theater of the Oppressed, in blending the roles of participant and performer. Information about the Machine to be Another is made open and publicly available as hardware lists, GitHub repositories, and interaction protocols. The project is the work of a collective, Be Another Lab, and has resulted in particular, formalized instantiations such as Gender Swap, as well as more free-form experiences where participants can repurpose the system in a variety of ways. For example, a mother asked at one performance to use the system with her daughter, neither of whom had been previously trained to use the system. This became a particular performance, “Being Sarah”. When I participated, I was allowed to take on the role of the performer for the next participant.

In one version of the experience, a participant is led blindfolded into a room. The participant is asked to sit down, and is then fitted with an Oculus headset. In my experience of this, I was asked to open my eyes, and to move slowly. When I did, I was gazing at myself. I looked down, and saw the body of someone else. The impulse to stretch my arms out quickly, touch myself and examine my new limbs more closely was tempered by the knowledge that the illusion was fragile, and the surface held together by agreement between myself and the performer to move slowly and to mirror one another. Be Another Lab suggests that this ongoing “agreement” to cooperate between participant and performer helps to ensure that the mirroring process is one of mutual exploration and consent, rather than an experiment in controlling or invasively manipulating another person’s body. While this is certainly true, it is also true that the performer and the participant are in a structured environment and can easily feel pressured to continue. However, this is a different set of conditions than in other VR perspective-taking experiences, because the liveness of the performer allows for an ongoing negotiation and consent to mirror. Inhabiting an avatar does not have this social aspect.

Project Composition & Development

The Machine to be Another uses live performers and doubled objects to produce a tactile simulation. In some versions of the performance, a participant is handed an apple or other object at the same time as the performer, by two additional performers who remain “unseen”.

The visual simulation is produced by the main performer, who carefully watches the participant and mirrors her body movements. The performer wears a head-mounted stereoscopic camera, which is controlled by the head movement and gaze of the participant. As the participant looks left and right, the camera worn by the performer will track this motion.This frees the performer to watch the participant, and mirror her actions, while the participant can look freely about.

The Machine to be Another is situated between art installation and scientific experiment into empathetic feeling. Yet, rather than a creative performance of “complete” scientific concepts, Be Another Lab falls into a category suggested by researchers Georgina Born and Andrew Barry– the public experiment– where artistic production suggests new avenues for research, discovers connections, and affects the direction of scientific work The Be Another Lab has explained that through their position as artists, they are able to work more swiftly and improvisationally than would be possible in a research setting. However, by suggesting that an artistic intervention might provoke specific changes in attitudes and bias, the project is subject to a  contemporary public expectation of scientific validity, borne out in Q&A sessions. At the 31st Chaos Communication Congress during a panel discussion, the be Another Lab encountered audience questions about experimental rigor and potential directions for quantifying their claims for empathy-building. For the time being, this is a question that the group cannot answer. It also raises the question about what we should expect from art, and whether it is necessary that effects be quantified and repeatable before they deserve our trust or scrutiny. The question indicates some confusion about where and how to distinguish between scientific research and artistic experience.

Prior to the Machine to be Another, members of the collaborative worked on project called The Error Machine, which was was presented at L’Estruch in Sabadell Barcelona during a 2013 residency. Philippe Bertrand explains that they were interested in the “high-low approach of the piece… a  neuroscience-art-hack performed on streets with minimum technology.”

One path forward for the group will be collaborations with academic research groups, an approach that may begin to satisfy the desire for measurable results and A/B testing now familiar in parts of the documentary world aimed toward social change. Another approach is to think of the publicness of the project thus far (all code and instructions from Machine to be Another are available on GitHub), and its status as “embodiment system” rather than performance, object, or installation, as an intervention into a different kind of collective knowledge. Rather than achieving impact through officially sanctioned academic research, the perceived effectiveness and individual experiences of shared perception spread from participant to participant.

Aaron Souppouris reports in an article on The Verge of his experience with the system, that “[the performer’s] voice, which was recounting her thoughts on feminism and self-image, became my stream of consciousness.” However, Coxon, the scientist accompanying him had a less convincing experience, noting that he was expecting to be “transported”, but was nonetheless aware of the rift between the performer’s inner experience and his mediated exposure to her visual and haptic experience. Coxon suggests that being primed for the system in different ways could affect its reception, perhaps reducing what he terms “variance” in the experiment. However, Coxon did experience “proprioceptive transference”, a sensation of disruption, or shock: I felt it as well, when I “moved” a hand that was not mine and yet still felt the apple that was handed to me. These responses are interesting because they suggest that the empathetic response to this art is drawn more from “emotional empathy” than “conceptual empathy”.

Empathy & Presence in VR artwork

The word empathy derives from a german word, Einfühlung, which had been used in artistic contexts to describe the “feeling-into” of an object. This word acquired meaning as a way to characterize embodied viewer response drawing both from fields of perceptual psychology and art history, an interest shared between research communities of art and of science. It rests on the idea that emotional identification is an internal, individual ability that can produce particular social feelings. It was also eventually rejected by dramatist Bertolt Brecht, whose participatory methods inspired Augusto Boal. Brecht believed that emotional identification ran counter to the ability of individuals to effect actual social change, because it lacked the moment of “alienation” that could disrupt existing worldviews and provoke internal change. In film and media studies, descriptions of empathetic feeling typically fall into two categories — emotional and cognitive. The first describes sympathetic response, “proprioceptive transference”, and perhaps also is the domain of “mirror neuron” response. It occurs immediately, without conscious thought. The second type describes a conceptual or imaginative experience, where one person tries to reconstruct how something feels to another person, not just how it would feel to them. To understand the distinction, consider observing someone cut their finger on a piece of paper, and compare a reaction to this to an observation of someone receiving news that a loved one has become ill. We may have sympathetic reactions to both experiences, but for the second experience, we must imagine what it must be like to be in that person’s relationship to understand how the news has affected them. Of course, we will also try engage in this “conceptual empathy” process by thinking of our own loved ones, and how we would feel if something happened to them. Neither approach allows us to directly know the experience of another, but we approximate it in various ways. Which of these reactions is the “empathy” that we hope to quantitatively increase? And, in another test of the term, does either provoke an instability of worldview that Brecht considered necessary for social change within aesthetic experience?

The project of Be Another Lab, and the scientific investigation of “empathy” as a pro-social trait, rests on a connection between social change and individuality. In a historical study of how empathy developed as a concept, Gillian Swanson writes that “the imagining of the inner life of others may actually be based on the projection of the individual’s own feelings into the object, rather than an authentic understanding that derived from an engagement with the particularity of the object.” Put otherwise, my experience of an assemblage of the visual and haptic experience of another is putting me in their place, but not actually in their body. And this is the central critique of VR as a successful medium for “increasing” empathy: that it cannot reproduce internal states, only the physical conditions that might influence that. As Coxon noted in Soupporis’ article, the system was incapable of reproducing “inner experience”.
However, this also rests on a particular kind of understanding of empathy, and one which may not be the actual project of Be Another Lab. Despite the name, which does imply a kind of inhabiting of the “other”, it is also possible to imagine this artwork in dialogue with creating a shared experience of mutual attention, rather than empathetic imagining. Think “listening”, in a kind of perceptual way, rather than “guessing”. In an interview with Maarte Roel, one of the creators of the project, Maarte noted that the focus of the experiment is on the relationship formed between performer and participant, not the illusion of body-swapping. However, scientific rigor and generalizability may lose this nuance: by considering empathy as a trait which can be increased or decreased in quantity, rhetoric around empathy reinforces the definition of empathy as an individual, internal experience. Yet in performance and in action, it occurs between people and is relational. There is a need for scientific understanding of how we form social feeling, and the Machine to be Another is aware of the potential gains from harnessing scientific research in this field. But there is also a need for performative and artistic research outside the context of replicability and standardization. Empathy has been constructed as a social and political project, one which is predicated on an individual, isolated “self”. VR is certainly capable of facilitating new experiences of subjectivity, perception, and sociality, but to call all these phenomenon “empathy” limits the potential of the medium. The Machine to be Another achieves this through performance and real-time improvisation, aspects that are heightened through VR but are not dependent on it. While academic rigor can associate changes in chemical levels, galvanic skin responses, heart rates, and other physical metrics with an experience, it is equally the suspension of normal social conditions, experimentation, and surprise that forge the essential experience of the Machine to be Another. And this is perhaps where the “public experiment” does more than suggest new directions for science: it can also be imagined as a way for artists and the public to reclaim authority over knowledge and experience, to create new definitions for “empathy” that go beyond perspective-taking.


1- Freedberg and Gallese “Automatic empathetic responses constitute a basic level of response to images and to works of art. Underlying such responses is the process of embodied simulation that enables the direct experiential understanding of the inten- tional and emotional contents of images.” pp. 202

like 1


show comments

Join the Discussion

powered by_