You are on page 1of 6

Artif Life Robotics (1998) 2:151-156


9 ISAROB 1998

AI~/I I( "1,1'.

Christa S o m m e r e r 9 Laurent M i g n o n n e a u

The application of artificial life to interactive computer installations

Received: February 13, 1998

Abstract This p a p e r reports on the creation of interactive

c o m p u t e r installations that combine artificial life and real

life by means of h u m a n - c o m p u t e r interactions. These
installations have focused on real-time interactions and
evolutionary image processes. Accordingly, visitors to the
installations b e c o m e essential parts of the systems by transmitting their individual behaviors, emotions, and personalities to the image processes of the work. Images in these
installations are no longer static, pre-fixed, and predictable,
but b e c o m e "living systems" themselves, r e p r e s e n t i n g
minute changes in the viewers' interactions with the evolutionary image processes. Natural evolution has brought
about a vast variety of forms and structures in nature. This
research considers how artificial evolution can function as a
tool of the visual creation process; design should no longer
be done by a designer or artist, but should e m e r g e through
the evolutionary image process itself.

In the 1950s, artists like John Cage, R o b e r t

Rauschenberg, N a m June Paik, a n d others also started to
include the audience in the creation process, thus widening
the concept of the art work to a further dimension. " H a p p e n ings," "Performances," and " V i d e o A r t " were part of
a new m o v e m e n t called Fluxus. 3 This considered art to be a
communication process b e t w e e n the artist, the art work, and
the audience. Artists such as Steina and W o o d y Vasulka 4
investigated the new technical and conceptual possibilities of
V i d e o A r t in the 1960s. O t h e r d e v e l o p m e n t s since then
further explored the idea of "art as a process," and new art
forms such as " L a n d Art," "Installation A r t , " and "Conceptual A r t " were pioneered. With the a p p e a r a n c e of new
c o m p u t e r technologies a n o t h e r dimension was introduced to
the artistic creation process: time and virtual space. 5

Artifical life and art

Artificial life has b e c o m e a topic that not only interests the
scientific community, but has influenced the art world as
well. ~ Traditionally, art was considered to be the sublime
creation of the artist, who in I m m a n u e l K a n t ' s terms, 2 was
the genius through whose inherited "ingenium" nature decided the laws of art.
H o w e v e r , in the 1920s and 1930s, artists such as K u r t
Schwitters, M a n Ray, M a r c e l D u c h a m p , and others introduced the idea of "art as a process" in a new art form called
Dadaism. R a n d o m processes and automatisms were exploited for the first time to create such then controversial
works as Schwitters' " U r s o n a t e " and Marcel D u c h a m p s '
famous " R e a d y - m a d e s . "
C. Sommerer (~Z) - L. Mignonneau
ATR Advanced Telecommunications Research Laboratories, MIC
Media Integration and Communications Research, 2-2 Amity
Hikaridai, Seika-cho, Soraku-gun, Kyoto 619-02, Japan
Tel. +81-77495-1426: Fax +81-77495-1408

The principles of artificial life and the e m e r g e n c e of advanced c o m p u t e r technologies in the early 1990s allowed
artists for the first time to study the visual creation process
itself. 6 One of the first artists to i m p l e m e n t artificial life
principles in the visual creation process was Karl S i m s ] In
his " G e n e t i c Images" he allowed visitors to choose images
that would then develop by m e a n s of genetic cross-over and
mutation. T h e resulting images r e p r e s e n t e d a mixture of
h u m a n selection and preferences as well as artificial genetics. O t h e r artists and designers have dealt with artificial life
in a m o r e commercial and game-like fashion, creating, for
example, CD R o m s like "Sim Life," etc.
In 1993-1994 we introduced o u r interactive c o m p u t e r
installation " A - V o l v e , " one of the first systems where visitors could actually create artificial creatures, interact with
them, and watch t h e m evolve. ~ In collaboration with T o m
Ray, 9 a biologist and the c r e a t o r of "Tierra," we d e v e l o p e d
this interactive system by i m p l e m e n t i n g the principles and
m e t h o d s of artificial life in the artistic creation process.
O n e year later, in 1995, we d e v e l o p e d " P h o t o t r o p y , "
a n o t h e r interactive installation that allows visitors to inter-


20 vertex
points (x,y)

Fig. 1 Side view

Fig. 2 Section



act with virtual insects 1~by nurturing them and reproducing I

I vertex points I vertex points I C,
length 9
k [Tex [
In 1996, we became interested in the building blocks of
Y20parame tersY 20 parametersY20 parametcl~/,m0~,para
visual creation, and investigated how simple structures can 20 parameters
create complex looking shapes and forms by using genetic
manipulations. We developed " G E N M A - Genetic Manipulator," an interactive installation that allows visitors to
create, manipulate, and explore the design and form of
artificial insects. In 1997, we developed this concept further
I (R)(G) (B~(R) (G) (B)(A) I
and created an on-line application of the above principles
and implemented them in the interactive environment "Life
color ~brightnes)~ r
Spacies," an Internet Web page and text editor that allows
the creation of genetic code for artificial creatures through Fig. 3 Genetic string
e-mail text messages.
These four interactive computer installations will now be
In order to get a three-dimensional form, we need not
described in more detail.
only a side view but also depth information. This time the
visitor draws a form that represents the section through the
creature's body along the z-axis (Fig. 2). The same process of
acquiring vertex points is applied to the creature's section.
Interactive computer installations
We acquire 20 vertex point for x and 20 vertex points for y, all
of which are added to the creature's genetic code (Fig. 3).
Figure 3 shows the genetic string of the creature with its
In the interactive real-time environment A-Volve, visitors 90 parameters. Each creature in the pool has such a genetic
interact with virtual creatures in the space of a water-filled string, and the creatures are haploid and asexual.
glass pool. 11 These virtual creatures are products of evolutionary rules as well as being influenced by human creation Fitness and speed
and decision. By designing any kind of shape and profile
with their finger on a touch screen, visitors "bear" virtual The m o v e m e n t and behavior of the virtual creature is deterthree-dimensional creatures that are automatically "alive" mined by its form, which is based on how the viewer designs
and swim in the real water of the pool. Algorithms calculate it on the touch screen. Behavior in space is, so to speak, an
a creature's form and m o v e m e n t in space. A-Volve pro- expression of form. Form is an expression of adaptation to
vides a novel system where movement and fitness is linked the environment. Form and m o v e m e n t are closely connected, so the creature's capacity to move will decide its
to the design and shape of a creature.
fitness in the pool.
The distance through which a creature pushes itself forCreation
ward by one muscle contraction is its speed, and this speed
Our special touch-screen editor allows the visitor to draw is equivalent to the creature's fitness. The greater the disany form (Fig. 1). As the visitor draws a side view of the tance a creature can move with one muscle contraction, the
creature, the outline of this drawing will be mirrored faster it is, and hence the fitter it is. 12
around the z-axis. Our software now subdivides this drawing into 20 points or vertices'that provide the length and size Predator - prey behavior depending on energy levels
for each parameter point. We then add this size and length
information to the genetic code of the creature's genetic The fittest creature survives the longest and is able to mate
and reproduce. The creatures compete by trying to get as
string (see Fig. 3).

Birth - death

A n A - V o l v e creature can be born in two different ways:

1. by visitor on the touch screen;
2. by mating and genetic exchange of two p a r e n t creatures.
T h e y can die in three different ways:
1. starving - they can not get e n o u g h energy by killing other
2. natural d e a t h - their m a x i m u m lifetime is reached;
3. being killed - a prey is killed by a predator.

Fig. 4 Cross-over and mutations. Black sections, mutations chosen at


much energy as possible. Thus p r e d a t o r creatures will hunt

for p r e y creatures and try to kill them. A creature is not
born as a p r e d a t o r or a prey, but will decide w h e t h e r to
attack or flee d e p e n d i n g on the other creature's fitness and
energy at the current m o m e n t .
A creature can initially be a p r e d a t o r but later b e c o m e
a prey if a fitter creature with m o r e energy enters the pool.
The creatures therefore have constantly to change their
strategies a b o u t w h o m to attack and w h o m to flee from. A
creatures energy level E-1 when the creature is b o r n on
enters the pool. A s the creature's energy level (E) also
decreases with each m o v e m e n t it makes, it can b e c o m e
hungry as soon as the critical level of E < 1 is reached. T h e n
creature b e c o m e s a potential predator. If it is fit e n o u g h to
catch an other creature, it will look for a suitable prey,
attack it and a d d up the prey's energy to its own energy.
Through his vision system the p r e d a t o r always chooses the
prey that is n e a r e s t to it and which will provide the most
energy with the smallest effort of movement. D e t a i l e d description on the relationship between fitness, energy, lifetime, p r e d a t o r - p r e y behavior and genetic exchange is given
in the literature. 12
Mating and genetic exchange

If two strong creatures with E > 1 meet, they can create an

offspring. This new creature carries the genetic code of its
parents. M u t a t i o n and cross-over provide a nature-like
r e p r o d u c t i o n mechanism.
Figure 4 shows an example of how the cross-over o p e r a tion is p e r f o r m e d between parents. Some m u t a t i o n is applied at r a n d o m to the genetic code of the child c r e a t u r e J 2
The newly b o r n offspring will now also live in the pool,
interacting with visitors and other creatures.

A s the genetic code of the offspring is t r a n s p o r t e d from

generation to generation, a n d the o p e r a t i o n of the system
is b a s e d upon selection for fitter creatures, the system is
able to evolve over time t o w a r d having fitter creatures.
A l t h o u g h evolution could t a k e place by itself without any
influence from outside, the system is designed in a way that
allows the interaction and creation of forms by the visitor to
have a significant influence on the evolutionary process. W e
can consider the visitor as a kind of external selection
T h e three main internal p a r a m e t e r s - fitness, energy, and
lifetime - regulate interaction, reproduction, and evolution
a m o n g the creatures. The external p a r a m e t e r s are the visitors' drawings on the touch screen as well as their interactions with the creatures.
C r e a t u r e - visitor i n t e r a c t i o n

The creatures also interact with the visitors by reacting to

their hand m o v e m e n t s in the water (Fig. 5). If a visitor tries
to catch a creature, it will try to flee or it will stay still if it
gets caught. Thus, the visitor is able to influence evolution
by, for example, protecting p r e y against predators.
N o n e of the creatures is p r e d e t e r m i n e d ; they are all b o r n
exclusively in real time through the interaction of the visitors and the interaction of the creatures. Thus a large
variety of forms is possible, representing h u m a n and evolutionary rules. By closely connecting the real natural space of
the water to the unreal virtual living space of the creatures,
A - V o l v e minimizes the b o r d e r s b e t w e e n "real" and "unreal," creating a further step, after "Interactive Plant G r o w ing", m4 in the search for natural interfaces and real-time
A - V o l v e is installed on a long-term basis at the N H K
building, N T T Nagoya. M o r e information a b o u t A - V o l v e is
available at: and in the
literatureJ 2
" P h o t o t r o p y " is a biological t e r m describing the force that
m a k e s organisms such as bacteria or plants follow light in
o r d e r to get nutrition and hence to surviveJ ~

Phototropy is also an interactive computer installation
where visitors interact with virtual insects using a normal
flashlight. By shining the light onto parts of a 3-m x 4-m
screen, the viewer awakes virtual insects that are born in
cocoon-like growth forms. The insects soon start to fly and
follow the actual physical beam of the viewer's flashlight,
seeking their energy source, light. If the insects reach
enough light, they can live longer and reproduce, if they do
not reach enough light, they will soon die. Since all insects
want to gain as much energy as possible, a big swarm of
insects will follow all changes and movements of the
visitor's flashlight beam. The visitor has to be careful: shining too much light on the insects will burn them. However,
if they are carefully guided, they will mate more frequently
and increase their population.
The genetic code of a single insect is transferred from
parent to child, thus ensuring that new generations and
individuals of virtual insects are constantly being born, all
following and fighting for light. The visitor in Phototropy
thus supports, develops, and enhances the lives of artificially
living insect populations.
G E N M A - Genetic Manipulator
G E N M A ~5is a machine that enables visitors to manipulate
artificial nature on a micro scale: abstract, amoeboid, artificial, three-dimensional forms and shapes. The principles of
artificial life and genetic programming are implemented in
the creatures' structure, allowing the visitor to manipulate
their virtual genes in real time.
Looking into a mirrored glass box, the visitor sees the
creatures as stereo projections. The visitor puts one or both
hands into the glass box and tries to grab the creatures that
are virtually floating in the space of the box. The genetic
code of each creature is schematically displayed on a touch
screen. By moving a finger on the touch screen, the visitor
can manipulate the genetic code of the creature, and thus
modify its appearance in the glass box in real time.
By selecting, merging, and recombining different parts of
the genetic string, the user can engage in more stimulating
experiments and learn how to create complex forms out of
seemingly simple original structures (Fig. 6). G E N M A allows the visitor to explore the tools of genetic manipulation
by taking parts of the genetic strings, cutting, pasting, or
multiplying them, and adding mutations and variations.
On a visual level, G E N M A further explores the concept
of "natural design" or "auto design," i.e. a design that is not
pre-fixed and controlled by the artists, but which represents
the degree of interest and interaction of each single visitor.
Each visitor creates the desired forms, aided by artificial
genetics, mutation, and manipulation. To push it further,
the visitors thus become "creators" or "artists" themselves,
using the power and possibilities of such tools.
Life Spacies - from text to form on the Web
In 1997, we were asked to develop a new interactive installation called "Life Spacies ''16 for the permanent collection

of the NTI'-ICC InterCommunication Museum in Tokyo,

Life Spacies is an evolutionary communication and interaction environment that allows remotely located visitors to
interact with each other in a shared virtual environment.
Visitors can integrate themselves into a three-dimensional
complex virtual world of artificial life organisms that
react to their body movement, motion, and gestures. These
artificial beings also communicate with each other as well as
with part of an artificial universe, where real and artificial
life are closely interrelated through interaction and exchange.
Through the Life Spacies Web page (Fig. 7), people all
over the world can contribute to the system by simply sending an e-mail message to the Life Spacies Web site (http:// to create their own artificial
These creatures will immediately start to live in the
Life Spacies environment at the ICC Museum and interact with the visitors on-site. When a visitor virtually
touches a creature, it will produce a clone of itself, when two
visitors each touch a creature these two creatures produce
an offspring, a genetic mix of the two parent creatures.
The artificial species can be created in two different
ways. First, through incoming international e-mail messages. A text-to-form editor creates the genetic code for
each creature: one message is one creature; complex text
messages create complex creatures; different levels of complexity within the text represent different species. Second,
by the creatures themselves: reproduction initialized by the
visitors helps the creatures to propagate their genotype in
the system so they can form groups of different species.
After sending an e-mail message to the Life Spacies Web
page, the sender soon receives a curriculum vitae for his
creature, as well as an image of how it looks. When the
creature dies, a report is sent to its creator, telling him or
her how long the creature lived and how many offspring and
clones it produced.
Life Spacies is again based on the idea of evolutionary
design, which is not predetermined by the artist but depends solely on the interaction of the visitors and the evolutionary process. Only the messages mailed from people all
over the world and the reproduction and evolution of the
creatures themselves will determine how the creatures will
look and how they will behave. Thus one cannot really
predict how the work will evolve and what kind of creatures
will emerge. It will depend exclusively on how many people
send messages, how complex these messages are, and how
the creatures reproduce among themselves and through the
selection of the visitors in the museum.
In Life Spacies, interaction, interrelation, and exchange
happens on human-human, human-creature, creaturecreature, human-environment, creature-environment, and
life-artificial life levels.
As the interaction rules are nondeterministic and
multilayered, an open system was created where each entity, whether real life or artificial life, whether actually
present (at the ICC Museum) or virtually present (the
users on the Net, or the creatures as code), is regarded

Fig. 5 Visitors interacting with
artificial creatures in "A-Volve"

Fig. 6 GENMA - Genetic Manipulator

as an equally i m p o r t a n t c o m p o n e n t of a complex life-like


Natural interfaces
W e have b e e n most interested in, and have p i o n e e r e d , the
invention of N a t u r a l Interfaces as they t r a n s p o r t the concepts of life, variation, and personality. 6 Using living plants
as an interface, for example, not only provides an interesting new connection b e t w e e n computers and a living entity,
but also poses the question of what a plant is, how we
perceive it, and how we interact with it. Natural Interfaces
allow us to project our personalities into the virtual space.
T h e y also diminish the anxiety one has when entering the
virtual space. W e have so far investigated the use of plants,

Fig. 7 Life Spacies - an evolutionary communication and interaction



water, unencumbered 3D full-body detection ("3D Video

Key"IT), and light as interfaces.

ready pointed out, "all essential things appear in spite of

MulUlayer interaction and artificial life


Interaction is interesting to the visitors if it is not linear or

predictable, but like a journey. The more one engages in
interaction, the more one can learn about it and explore it.
Nonlinear interaction and multilayered interaction should
be easy to understand at the very beginning and contentrich, so one is able continuously to discover different levels
of experience. This is where artificial life can provide a new
form of creation process that is not predesigned and predictable, but which varies according to the parameters employed in the system.
In A-Volve, for example, artificial evolution and selection mechanisms create a system of virtual creatures that
are semi-autonomous and provide a variety and complexity
of forms and interaction not created by the artist or designer, but by the interaction between the visitor and the
artificial creatures themselves.

1. Kusahara M, Sommerer C, Mignonneau L (1996) Art as living

system. Syst Control Inf 40(8):16-23
2. Kant I (1996) Kritik der aesthetischen Urteilskraft. In: Kritik der
Urteilskraft, Suhrkamp Taschenbuch Wissenschaft B180, 181,
A178, 179, p 241
3. Goodman C (1995) The electronic frontier: from video to virtual
reality. In: Info Art '95, Kwangju Biennale Foundation, pp 23-42
4. Vasulka S, Vasulka W (1996) Machine media. San Francisco Museum of Modern Art
5. Stiles K (1996) Art and technology. In: Theories and documents
of contemporary art. University of California Press, Berkeley, pp
6. Sommerer C, Mignonneau L (1998) Art as a living system. In:
Sommerer C, Mignonneau L (eds) Art at Science. Springer,
Vienna, New York, pp 148-161
7. Sims K (1991) Artificial evolution for computer graphics. Comput
Graphics 25:319-328
8. Sommerer C, Mignonneau L (1994) A-Volve: a real-time interactive environment. ACM Siggraph Visual Proceedings, pp
9. Ray T (1991) An approach to the synthesis of life. In: Langton
C (ed) Artificial life II. Addison Wesley, Reading, pp 371408
10. Sommerer C, Mignonneau L (1995) Phototropy. In: Mattei M (ed)
Oltre il villaggio globale - Beyond the global village, Electra edn,
Milan, pp 134 ff
11. Sommerer C, Mignonneau L (1997) A-Volve - an evolutionary
artificial life environment. In: Langton C, Shimohara K (eds) Artificial life V. MIT Press, Cambridge, pp 167-175
12. Sommerer C, Mignonneaul (1997) Interacting with Artificial
Life: A-Volue. In: Complexity Journal, John Wiley & Sons 2(6):
13. Sommerer C, Mignonneau L (1993) Interactive Plant Growing.
ACM Siggraph Visual Proceedings, pp 164-165
14. Sommerer C, Mignonneau L (1993) Interactive Plant Growing. In:
Weibel P (ed) Ars Electronica 93 - Genetic art Artificial life, pp
15. Sommerer C, Mignonneau L (1996) G E N M A Genetic Manipulator. In: Ars Electronica '96 - Memesis: the future of evolution.
Springer, Vienna, New York, pp 294-295
16. Sommerer C, Mignonneau L (1997) Life Spacies - an evolutionary
communication and interaction environment. In: ICC Concept
Book, NTT-ICC, pp 96-101
17. Sommerer C, Mignonneau L (1995) Trans Plant. In: Moriyama T
(ed) Imagination. Tokyo Metropolitan Museum of Photography,
chap 2 ff
18. Bachelard G (1988) Der neue wissenschaftliche Geist. Suhrkamp
edn, p 12

Interactivity and artificial life teach us to rethink our definition of art and broaden our view, because they allows us to
integrate personality, variety, the processes of nature, and
new reflections on art and life itself. The artist who creates
such installations only provides the framework: the visitors
themselves must then create the artwork through their interaction with each other, with the system, and with the
image processes of the work. As the images in the installations are no longer static, pre-fixed, and predictable, they
come to resemble living processes themselves, reflecting the
influence of the viewers' interactions with them and
the internal principles of variation, mutation, and evolution.
The image processes are no longer reproducible, but continuously changing and evolving. The artwork could therefore be metaphorically considered to be a living system
itself, 6 representing the relationship and interaction between life and artificial life. As Friedrich Nietsche has al-

everything". TM