Robots and Art pp 365-377 | Cite as

I Want to Believe—Empathy and Catharsis in Robotic Art

Part of the Cognitive Science and Technology book series (CSAT)


Since the early 90s, we have been creating interactive installation and performance projects using robotics, audiovisuals, and processes inspired by Artificial Life. The goal of these projects is to induce empathy from the viewers towards characters that are nothing else than simple articulated metal structures. Our objective is to conceive and realize large-scale robotic environments that aim to question, reformulate and subvert the notions of behavior, projection and empathy that generally characterize interactions between humans and machines.

Robotics as Artistic Medium

Robotic Art is an emerging discipline where scientific research, artistic creation and philosophical investigation are intimately interrelated. Of the few artists actively involved in this field, each one of them has in some way or another developed new technologies, techniques and methodologies of production that enable the creation of innovative works of art integrating robots, machines and automatons. Moreover, these works are raising fundamental philosophical and sociological questions about the relationships between human beings and machines, between the real and the artificial, and between the living and the non-living.

From Karel Capek to Nam June Paik to Survival Research Labs, artists have been exploring the concepts of robots and robotics for a few decades now, sometimes on their own, but often in collaboration with engineers and scientists. In 1997, Eduardo Kac coined the term “Robotic Art” to describe artistic projects based on or developed around robotic technologies. In Foundation and Development of Robotic Art [1], he stated “As artists continue to push the very limits of art, traditionally defined by discrete and inert handmade objects, they introduce robotics as a new medium at the same time as they challenge our understanding of robots”. In the last 20 years, artists like Mark Pauline, Christian Ristow, Eric Paulos, Chico MacMurtrie, Ken Rinaldo, Simon Penny, Stelarc, Guy Ben-Ary, Robotlab and Jim Whiting, just to name a few, distinguished themselves by their impressive artistic application of robotics. Well-known Canadian artists like Max Dean, Norman White, Reva Stone, Istvan Kantor, Louis-Philippe Demers, Rafael Lozano-Hemmer, Janet Cardiff and David Rokeby also used robotics and behavioral systems in many of their works.

Since its early stages, our artistic work has been strongly influenced by scientific advances in the fields of Artificial Life and Robotics. We are particularly interested in creating original artistic projects by appropriating various engineering and scientific concepts and techniques such as cellular automatons, genetic algorithms, adaptive behaviors and reinforcement learning processes in order to subvert them from their intended purpose.

Robotic Art is not a single homogeneous discipline; rather it is a mixture of multiple technological areas involving mechanics, electronics, programming, as well as multimedia. In the same manner, our research program does not focus on one single problem or one field of study, it encompasses a wide variety of research projects that all have one thing in common: producing a work of art as a final outcome. This is why we simultaneously conduct research and develop projects that address machine perception and motion on the one hand, and machine aesthetics in both robots’ visual aspect and behaviors on the other.

An Aesthetics of Artificial Behaviors

Our aim is to artistically investigate how a machine can eventually turn into a sentient creature. We believe that behavior is a keyword in bio-inspired automaton design and actualization. A certain level of realism may be achieved by the illusions induced by actions and reactions of the machines and animats: the success of this dynamic form of computer-mediated communication may be measured by the effectiveness of the simulacrum. An effective simulation of the living is the result of different parameters acting to trigger impressions and empathy (visual appearance, sound emission or physical movement, for example), but behavior may be seen as the most convincing one as it generates a strong impression of autonomy and self-consciousness.

As we have been able to experience throughout the years, uncertainty and variability also play an important role in the behavioral relation with the viewer. Animated metal parts in a robot or dots on a computer screen can be seen as being alive if they move and react in a non-repetitive and unforeseeable way, giving a strong impression of self-decision and autonomy. One may wonder if Artificial Life creatures have to be figurative representations (anthropomorphic, zoomorphic or bio-inspired) to be convincing. From what has been observed in the various encounters with the public, as long as they manifest autonomous behaviors in the interaction process, effective agents could bear any abstract visual form.

The success of our work depends on two main interrelated factors: the make-believe imbedded in the robot artifact and the viewer’s desire to believe (evoking Eco’s intentio auctoris and intentio lectoris [2]). It functions through cathartic projection by triggering sensations, feelings and emotions in the viewer’s eyes. What happens next is a matter of pure subjective interpretation from the viewer’s part. Machines are a perfect reflection of our mind and we can certainly learn more about ourselves by interacting with them.

Robot Ontology and Perception

Recent advancements in Artificial Life and robotic technology encourage a new kind of art form that combines artificial morphogenesis, immersive environments, interactivity and reactivity with cognitive machines (robotics, automation and animatronics) to achieve aesthetic results. We often use the expression “theatrical machines” to describe physical and autonomous robotic agents integrating some kind of multimedia objects in their ontology (sound, light, video, etc.) as mean of expression. Application examples of this new practice include emulation of realistic creatures and lifelike systems, conceptual exploration in the aesthetics of artificial perceptions, behaviors and interactions, embodiment of machine mechanisms, etc.

Our research projects are principally based on the notion of perception: the viewer’s perception of the robot and the robot’s perception of the environment, as well as itself. Perception guides the effect created on the viewer, as our work is steered by the fundamental assertion that it is possible to create an impression of life simply through human-machine reactive behaviors of abstract robotic structures.

We can integrate both notions of sentience and embodiment in the larger concept of ontology. An ontology describes how the world in which the agent lives is constructed, how the agent perceives this world and how the agent may act upon its world. Our work is based on the merging of aesthetic, philosophic and scientific questions related to machine ontology, its awareness, perception and potential sentience. Our research projects also investigate the notion of the artificial construction of the “self” as one of the leading themes of our creative work.

Early Artistic Work

We started to develop Robotic Art projects in 1992, with the initial intent to animate sound and light in space in response to the viewers’ presence. Espace Vectoriel, a collaboration with Louis-Philippe Demers, was an interactive mechatronic piece where eight robotic tubes project sound and light beams in a dual choreographed and behavioral manner (see Fig. 1). Each tube contained a speaker and a light source and was mounted on a pan-tilt mechanism. Viewers were detected using an array of ultrasound devices. This installation was then presented in many international events dedicated to New Media and Electronic Arts and eventually followed by other projects of the same kind. For example, The Frenchman Lake (1995) also used the same concept of replicating a basic robotic audiovisual unit multiple times, in order to create a more complex overall environment.
Fig. 1

Espace Vectoriel (1993) Photo B. Vorn and LP Demers. Each robotic tube is projecting sound and light

Among these earlier works, La Cour des Miracles (1997) has certainly been a milestone in our trajectory. With this project, we moved away from simple duplication and produced multiple different types (or “species”) of robotic creatures, each one exhibiting specific behaviors in response to the visitors. Based on the conceptual framework of a “misery of the machines” and somehow strongly inspired by Victor Hugo’s Les Misérables [3], these machines were designed to express such notions as “pain” and “affliction”, as if they had their own difficulties in life. For example, the Crawling Machine was creeping laboriously on the floor. Slow and vulnerable, it tried to run desperately away from the viewers approaching. The Harassing Machine called upon the viewers passing by while moving its articulated arms towards them. At the extremity of these limbs, small tentacles agitated by compressed air tried to tease the intruders with importunate touches. The Convulsive Machine was a thin metal structure shaking with frequent but irregular spasms, especially when viewers come too close. The Heretic Machine was locked up in a cage, and when curious viewers came close by, it rushed violently towards them, grabbing the metal grid with its claws and shaking furiously its cage [4].

Le Procès (1999) was a live multimedia performance staging a world populated exclusively by robotic actors (see Fig. 2). It was presented for the first time as part of Zulu Time, a theatre play by Robert Lepage. Because it was our first robotic performance, this project was a logical following to our perceptually subversive démarche of creating machinic automata and cybernetic organisms showing metaphoric behaviors, as well as inventing surrealistic immersive environments where viewers are both visitors and intruders. Le Procès showed in a symbolic way the trial of machines by men, as well as the trial of men by machines. It acted like a reflexive tribunal where identities intermixed, where judges, jurors, victims and accused, took flesh in metal creatures born from our own conception of the world, of what is good and what is bad, of what is alive and what is not. As in Kafka’s famous novel [5], of which crime are we accused? Who’s judging? What will be the verdict?
Fig. 2

Le Procès (1999) Photo B. Vorn and LP Demers. Le Procès at EMAF 2002 (Osnabrück)

During the same period, we developed a series of Max software functions [6] called LifeTools and explored cellular logic by building monumental audiovisual cellular automatons. In projects like the Evil/Live (1997, 2002, 2004) series, Conway’s Game of Life [7] was used to generate patterns of light and sound in a large-scale aluminum matrix of halogen light bulbs. In the different versions, viewers were either consciously (by using video game-style controllers) or involuntarily (by using discrete sensors hidden in the environment) modifying the evolution of the light patterns on the grid. This series of audiovisual installations aimed to create a paradoxical context confronting the single-plane world of a cellular automaton to the 3-dimensional immersive environment surrounding the viewer. By using fast stroboscopic changes in light and quadraphonic sound effects, it produced a clear illusion of physical volume.

In a similar way, in the Stèle 01 (2002) installation, a cellular automaton was used to control an array of 128 small pivoting mirrors on top of which an anthropomorphic robot was standing, vaguely mimicking a statue towering above a mortuary stele (see Fig. 3). Inspired by monuments and tombstones from the Père Lachaise cemetery in Paris, this piece was designed to evoke the intimate dichotomy between the real and the virtual, life and death, movement and inertia.
Fig. 3

Stèle 01 (2002) Photo B. Vorn. The Stèle robot facing an array of rotating mirrors

Recent Work

The Hysterical Machines robotic installation (2006) was very much inspired by similar ideas as La Cour des Miracles. It was conceived on a principle of deconstruction, suggesting dysfunctional, absurd and deviant behaviors through a functional machine. It operated on a dual-level process expressing the paradoxical nature of Artificial Life. The first prototype of the Hysterical Machine (it was then renamed Prehysterical Machine) appeared in 2002, but later on we built ten more machines inspired by this prototype that became part of a larger environment. More recently, we have also created the Mega Hysterical Machine (2010), a supersized version of the original robot (eight times the size of the Hysterical Machine in volume and weight). Until now, this huge robot has only been exhibited on wide theater stages in places such as the Théâtre National de Toulouse, the Théâtre des Salins (Martigues) and the Théâtre de l’Avant-Seine (Colombes).

Each Hysterical Machine has a spherical body and eight arms made of aluminum tubing (see Fig. 4). It has a sensing system, a motor system and a control system that functions as an autonomous nervous system (entirely reactive). These machines are suspended from the ceiling and their arms are actuated by pneumatic valves and cylinders. Ultrasound sensors allow the robots to detect the presence of viewers in the nearby environment. They react to the viewers according to the amount of stimuli they receive (how close are the viewers, how many viewers walk by). Programmed with sets of very simple internal rules, the perceived emergent behaviors of these machines engender a multiplicity of interpretations based on single dynamic pattern of events.
Fig. 4

Hysterical Machines (2006) Photo B. Vorn. One of the hysterical robots equipped with small lasers

Built in continuity with our investigations in the aesthetics of artificial behaviors, Red Light (2005) was another interactive robotic environment conceptually similar to Hysterical Machines and La Cour des Miracles. In this case, the project evoked a certain “deviance of the machines” as it would exist in the hottest areas of a fictive world populated exclusively by these specific cybernetic creatures. This installation also explored techniques and technologies related to parallel mechanics and pneumatics with the construction of homemade pneumatic muscles. A parallel mechanism is a mechanical system that is connected to its base by two or more independent kinematic chains (assemblage of links and joints). A pneumatic muscle (also called McKibben actuator) is a flexible air piston made of inflatable material such as silicone or latex that contracts when activated. In Red Light, six suspended machines reacted to the presence of viewers by generating sound and light and by moving their body in a very organic and unusual way (see Fig. 5). Each robot unit was an assembly of four freely moving segments joined together by twelve McKibben actuators. Each one possessed a small network of pyroelectric sensors that allowed detection of moving visitors and triggered the various effectors part of the robot.
Fig. 5

Red Light (2005) Photo B. Vorn. Red Light being tested in the Hexagram black box

At that time, we had been working with different types of parallel mechanisms (for example, the two center-stage robots mounted on Stewart platforms in Le Procès) and pneumatic muscles (like the suspended robot tentacles in Red Light) and it appeared that they were able to provide unusual types of physical motion that could produce a more organic feel to our machines. Since then, we have explored various designs and build several experimental prototypes of machines that make use of these technologies to create lifelike artificial creatures.

In 2007, pursuing our experiments with parallel mechanical systems, but with a totally different approach, we started to work on the Grace State Machines project. The name of this project was inspired by a virtual “state of grace” that could be expressed by automatons and other finite state machines. This piece was a stage performance involving solely a human performer and a group of machines (see Fig. 6). Both were linked via a custom-made wireless motion capture system (based on fiber optics) and a set of specialized interfaces. By monitoring the human body movements and internal states and transposing this information to the robots’ body, we aimed to establish a dynamic and symbiotic relationship between the actors. They all eventually blended into a single organism, where flesh, bones, wires and tubes became a whole individual body.
Fig. 6

Grace State Machines (2007) Photo B. Vorn. Emma Howes interacting with one of the GSM robots

In this performance project, four robotic machines were built as abstract shapes and composed of stacked Stewart platforms (actuated sections similar to flight simulator platforms) and capable of producing very complex movements. These machines sometimes reacted to the performer’s body movements, sometimes moving on their own, inducing a response from the performer. With this project, we wanted to question the notions of physical perception, body expression and personal identity, and address kinesthesis not only as an internal proprioceptive mechanism but also as a potential exterior phenomenon actualized through the robotic extension of the body.

Also very different from our previous works, Partie de chasse (2010) was an interactive installation project that aimed to turn an industrial robot arm into a reactive organism. For this project, we used a Fanuc M16iB industrial robot. An aluminum moose head was installed at the tip of the robot arm and moved towards the viewers nearby (see Fig. 7). In order to detect the presence and location of the viewers in the surrounding space, we used the ManyEars microphone array system [8] and an elaborate set of sensors. (For obvious security reasons, viewers were kept at some distance from the robot.) When a viewer talked, the microphone array detected the position of the sound source in the room and the robot moose head moved in its direction. The robot moose was also able to react to certain vocal commands, but it was up to the visitors to find out what these were.
Fig. 7

Partie de chasse (2010) Photo B. Vorn. The aluminum moose head on the Fanuc robotic manipulator

The particularity of this project resided in bypassing the normal programming paradigm of this type of robot in order to have it execute real-time commands instead of a predefined sequence of actions. Many artists have used industrial robots in the past but they have always used them as simple automatons, in a similar way they are normally used in car factories. Few have ever tried to turn them into autonomous reactive creatures. With this project, we wanted to build a sensitive and responsive machine, which was conceptually based on adaptive and evolutive behaviors.

In our latest piece, DSM-VI (2012), the installation staged creatures expressing symptoms of “abnormal” psychological behaviors and stuck with some serious “mental health” problems, such as neurosis, psychosis, personality disorders, paranoia, schizophrenia, depression, delirium, and other forms of behavior and mental disorders. The project title was inspired by the famous reference manual published by the American Psychiatric Association, the DSM-IV [9].

The robotic creatures are the sole characters and actors of this singular interactive allegory. They were built in order to evoke dysfunctional behaviors that make believe in the disease that they internally bear. These machines are abstract structures made of aluminum, plastic and silicone, with no deliberate intent of visually representing anything (see Fig. 8). Above all, they are just machines and it is mainly their behaviors that give them an organic and living aspect.
Fig. 8

DSM-VI (2012) Photo B. Vorn. Early prototypes of DSM-VI robots

In the center of the DSM-VI installation, eight Psychotic Machines stand on their legs, lie down on their side or on their back (see Fig. 9). These machines have a pneumatics-actuated pair of aluminum legs, speakers, lights, sensors, and DMX pan-tilt LED spotlights. They react to the approaching viewers, but they also react to each other. They look like they are going to jump or run away, but they are helplessly fixed there, sometimes very calm, sometimes completely agitated, like a herd of untamed but chained animals.
Fig. 9

DSM-VI (2012) Photo B. Vorn. Opening of the BIAN 2012 exhibition in Montreal

In the surrounding space of the installation, three independent robots revolve on their base. They seem like they live in their own world, not so much connected to the environment. We call them the Autistic Machines. They are free-spinning turrets, on which a pneumatic robotic arm actuates something that looks vaguely like a human face. This facial impression is caused by the visual combination of a speaker and two pan-tilt robotic cameras. Using these cameras, the robots can look around in the environment using a face-tracking software. But instead of following the viewers like we would expect, they tend to avoid them. Also, due to the face-tracking software limitations, the robots sometimes see faces in the environment where there are not and suddenly fall in trance looking at the wall or the ceiling.

Current and Future Work

In collaboration with Louis-Philippe Demers (Nanyang Technological University, Singapore), we are currently developing the Inferno project. Inferno is a robotic performance inspired by the representation of the different levels of hell as they are described in Dante’s Inferno [10] or Haw Par Villa’s Ten Courts of Hell (which is based on a Chinese Buddhist representation) [11]. In this piece, the “circles of hell” concept is mainly a scenographic framework, a general working theme under which the different parts of the performance are regrouped.

The specificity of this performance project resides in the fact that the different machines involved in the show are installed on the viewers’ own body. The public then becomes an active part of the performance. Depending of the kind of mechanism that they are wearing, the viewers are free to move or in a partial or entire submission position, forced by the machines to act/react in certain ways. Like if they were inverted exoskeletons, some mechanical structures coerce the viewers in performing certain movements; others induce a physical reaction from them. With this work, our aim is to question the “cyborgification” of the modern man, as well as how technology imposes its rules upon us.

At the same time as Inferno is being completed, we are also developing the Copacabana Machine Sex performance project. It can be described as a mini Music Hall show where kitsch and machine aesthetics blend together in a single theatrical delirium. More conventional in its form, it involves a succession of different musical numbers where machines perform on stage as actors, musicians and dancers. The performance will be created in a way that it can either be configured for a traditional à litalienne stage or with viewers standing all around a central ground-level presentation.

Loosely inspired by Chico MacMurtrie’s Robotic Opera (1992), where a small group of humanoid robots performed various percussive musical pieces [12], the Copacabana project wants to present music-making machines as well as acting and dancing robots. Our goal is not to replicate a real nightclub, but to conceive a metaphorical extravaganza in response to the very deep question: “What would happen if machines would be on the stage of a cabaret?”



Special thanks to Concordia University (Montréal, Canada) for its support; the Canada Council for the Arts; the Conseil des arts et des lettres du Québec; the Fonds de Recherche du Québec Société et Culture (FRQSC); Martin Peach, who has been a dedicated technician for many years; as well as the numerous graduate and undergraduate students who have been working as research assistants on many of these projects.


  1. 1.
    Kac E (1997) Foundation and development of robotic art. Art J 56(3):60–67CrossRefGoogle Scholar
  2. 2.
    Eco U (1991) The limits of interpretation. Indiana U Press, BloomingtonMATHGoogle Scholar
  3. 3.
    Hugo V (1862) Les Misérables. A Lacroix, Verboeckhoven & Cie, ParisGoogle Scholar
  4. 4.
    Ismert L (1998) La Cour des Miracles. Musée d’art contemporain de MontréalGoogle Scholar
  5. 5.
    Kafka F (1925) Der process. Verlag die Schmiede, BerlinGoogle Scholar
  6. 6.
    Max (or MaxMSP) is a graphic object-oriented programming environment dedicated to multimedia control. Accessed 3 Aug 2014
  7. 7.
    Adamatzky A (ed) (2010) Game of life cellular automata. Springer, New YorkMATHGoogle Scholar
  8. 8.
    ManyEars is a sound localization software developed by IntRoLab at University of Sherbrooke. It implements real-time microphone array processing to perform sound source localization, tracking and separation. It was designed for mobile robot audition in dynamic environmentsGoogle Scholar
  9. 9.
    The DSM (2013) (Diagnostic and Statistical Manual of Mental Disorders) is now in its 5th editionGoogle Scholar
  10. 10.
    Alighieri D (circa 1308–1321) Divina Commedia. English edition: Alighieri D (2013) The divine comedy. Arcturus Publishing Limited, LondonGoogle Scholar
  11. 11.
    Turbeville T, Brandel J (1998) Tiger balm gardens: a Chinese Billionaire’s fantasy environments. Aw Boon Haw Foundation, Hong KongGoogle Scholar
  12. 12.
    MacMurtrie C (2014) Accessed 3 Aug 2014

Copyright information

© Springer Science+Business Media Singapore 2016

Authors and Affiliations

  1. 1.Concordia UniversityMontrealCanada

Personalised recommendations