Ars Electronica 1990
Festival-Program 1990
Back to:
Festival 1979-2007


ARTificial Intelligence & ARTificial ART

'Florian Brody Florian Brody / 'Mario Veitl Mario Veitl

In showing the interconnections between art and artificial intelligence we want to analyse the connections of these two areas and how they can learn from each other. We don't want to give a summary of recent or possible projects; instead, we will focus on new tendencies in order to animate the discussion on such topics.

Communication is defined as a primary concept in ART and AI and suffers from similar limitations in both areas. The model by Shannon and Weaver, which is seen as the classical model of communication, has similar borders in both areas.
The fundamental problem of communication is that of producing at one point either exactly or approximately a message selected at another point. Frequently the messages have meaning; that is they refer to or are correlated according to some system with certain physical or conceptual entities. These semantic aspects of communication are irrelevant to the engineering problem. The significant aspect is that the actual message is one selected from a set of possible messages. The system must be designed to operate for each possible selection, not just the one which will actually be chosen, since this is unknown at the time of design.

Shannon C. L, Weaver W.; The mathematical theory of communication, Urbana 1949.
This model defines a source of information which transmits information via a channel (with noise) to a receiver and the destination. This technically oriented concept influenced all fields of research in communication for many years. ART and AI were influenced as much as psychology, biology and other areas. Terms in our everyday vocabulary such as message, channel, or information source show that very clearly. The success of this theory lies in its apparently simple applicability to all sorts of technical problems.

In ART, problems of differentiation between "message", "medium" and meaning" are immediately obvious. Transcending these borders opens up new areas for creative and innovative work, and leads to an emphasis of the notion of "process" not only in computer art. Cybernetic works by Nicolas Schöffer, Jacques Polieri and Piotr Kowalski can be seen in this way.

Likewise, in Artificial Intelligence, defined by Barr and Feigenbaum in the following way, there is no possibility of ignoring the problem of communication:
Artificial Intelligence is the part of computer science concerned with designing intelligent computer systems, that is, systems that exhibit the characteristics we associate with intelligence in human behavior.

Avron Barr and Edward A. Feigenbaum: The Handbook of Artificial Intelligence, 3 vols, Los Altos CA: Wiliam Kaufman, 1981–1982
In working with AI-Systems, researchers have observed that an enormous amount of background knowledge is necessary for a program to solve everday problems. A further example can be found in the field of pattern recognition, where the excellence of our visual system, like the recognition of objects under different light conditions, has not been matched by any artificial system. Learning systems, which are able to structure their problems and solve them, or neural networks, which are not only successful in pattern recognition, are two examples of new methods in Al. A theory that shows us the problems under a new perspective was proposed by Maturana and Varela in 1979. Here, the starting point is the cognitive system itself. These autopoetic systems are operationally closed and in this way autonomous, they operate in a self-referential way in an environment to which they are structurally coupled. Through this operational closure, the systems are determined in structure and state, but this is not similar to a deterministic system in the classical sense. The term "autopoesis" describes the principle of self-generation and plays a central role. Autopoesis describes the organisation of organisms, their variety and their process structure. How can we discuss communication in this way?

Ashby makes us aware that information is not an intrinsic quality of a message.
Communication thus necessarily demands a set of messages. Not only is this so, but the information carried by a particular message depends on a set it comes from. The information conveyed is not an intrinsic property of the individual message.

Ashby, W. R.; An introduction into cybernetics, London 1956
The influence of a disturbance to an operationally closed system depends on the system's internal state. To establish an efficient way of communication, the human being constructs a consensual space and operates in it. In daily life this process is not usually observed, but when trying to explore new consensual spaces in art, immediately becomes apparent.
Interaction is the basis of communication. Without interaction there is no communication. Wolfram K. Köck therefore proposes using the term "communication" to designate a class of special intentional interactions between organisms.

The introduction of "interactive art" looks very dubious, especially since for every perception, an interaction is necessary. It is absolutely impossible to make non-interactive art. What is interactive art? The engagement with a reactive system could only be quantitatively different, for an interaction takes place in any case.

The computer makes it possible to look for new ways of expression; one example can be the creation of ARTificial ART. In the spirit of Maturana, creativity is the result of interactions of cognitive systems outside their domain of structured coupling. Here might arise a very fruitful common domain, which should not only result in better systems, but includes also the exploration of the social impact, an extremely interesting point.
Der Computer ist nicht nur ein Werkzeug, um bestimmte Bilder, Töne oder skulpturähnliche Installationen zu verwirklichen.

Stephen Wilson; Kunstforum Bd. 103, 1989
Artefacts can be the basis of interaction with human beings and thus resemble an interface – a term from computer technology that was introduced with teletype-machines and describes the way the computer presents itself to the user. The box or shape of the CPU can only tell us something about the hardware basis of the device. The dialogue interaction with a line-orientated or menu-driven system fools the user into believing that there is a real dialogue going on. Walker states in a paper from Autodesk:
I believe that conversation is the wrong model for dealing with a computer – a model which misleads inexperienced users and invites even experienced soft ware designers to build hard-to-use systems.

Walker John; Through the Looking Glass, an internal Autodesk paper.
Autodesk inc. Sausalito CA 1988
The concept of conversation based on words, which are exchanged with a partner is obviously inadequate. The search for a new-dubious-realism, an accomodation to given facts, is taking place on the wrong level.
When you're interacting with a computer, you are not conversing with another person. You are exploring another world.

Walker John; Through the Looking Glass, an internal Autodesk paper.
Autodesk inc. Sausalito CA 1988

The meaning of this statement becomes clear when we look at cyberspace systems. The glass of the tv-tube was still a border, but now the user finds himself in the program. Weizenbaum points out the pointlessness of trying to understand the functionality of a computer by opening the device.

With the development of cyberspace, applications related to ART and AI enter a new virtual working area, which cuts off the desire to imitate the real world. The possibility of direct experience is initiated by audio-visual input, but only the use of force-feedback systems will make it a real experience, because the human being needs, at least in the initial stage, a "remapping" to the real world.

The possibility of connecting humans to a machine through detectors is discussed in a work by Gibson and 0. Wiener writes about some consequences of bioadapters.

A further interconnection between ART and Al is proposed by Marcel Duchamp in his concept of "bachelor machines". These kinds of machines have been well known for a long time, but Duchamp gave them their special name.

Carrouges writes:
Eine Junggesellenmaschine ist ein phantastisches Vorstellungsbild, das Liebe in einen Todesmechanismus umwandelt. (…) Eine Junggesellen-maschine ist zunächst eine unwahrscheinliche Maschine. (…) ihre" Hautstruktur beruht auf mathematischer Logik. Die Junggesellenmaschine besteht immer aus zwei Bildbereichen, dem sexuellen und dem mechanischen, beide unterteilen sich wieder in einen männlichen und einen weiblichen Bereich.

Carrouges Michel; Les Machines Celibataires. Arcanes Paris 1954
The significance of a machine which is able to mimic the human in every way is pointed out very clearly here in the field of art.
If ART reaches its borders with static representations, it comes up with the concept of movement (cubismus), and in this way ART loses dependency on the object. In AI also, static representations are not powerful enough. AI and ART systems can include memory and intelligence and thus come into social coupling in the sense of Maturana. Work on epistemological questions must not be seen as isolated from this consideration.

Both AI & ART have a claim concerning representation and explanation of the world. Although the methods are different, we can observe some important similarities, which point to more than a simple coexistence of a common interdisciplinary research area. The examination of communication has shown us that the electronic technologies not only pose new questions, they also provide the possibility of bringing together different areas.