Relazioni Emergenti: Experiments with the Art of Emergence

Abstract

Progress in the scientific understanding and simulation of natural evolutionary mechanisms may be creating the basis for a new stage in evolution: the coming of artificial beings and artificial societies. Culture itself, aesthetics and intelligence are coming to be seen as the emergent, self-organizing qualities of a collectivity, evolved over time through both genetic and linguistic evolution. This paper sketches the development of hybrid digital worlds, in which artificial beings are able to evolve their own cultures, languages and aesthetics. Finally, the authors discuss their interactive audio-visual art installation Relazioni Emergenti, based on artificial-life environments. In this work, digital beings can interact, reproduce and evolve through the mechanisms of genetic mutation. People can interact with these artificial beings, creating hybrid ecosystems.

Patterns of Convergence

Recent trends in information and communication technologies reveal an impressive convergence among different research disciplines: artificial intelligence (AI), virtual reality (VR) and artificial life (A-Life). More and more projects involving environments where real people can interact with artificial entities are being realized. The concept of "virtual reality" is now merging with the wider concept of "virtual world," as the conjunction of a virtual environment with artificially evolving creatures. The more advanced of these efforts toward the development of intelligent beings acting as autonomous agents in a virtual environment raise basic questions relating to the current definition of such terms as life, intelligence and evolution.

The human-machine interface is changing into a hybrid ecosystem (Fig. 1). The virtual environment, once a static set of rules, has become dynamic, co-evolving with the action of the living agents. Interaction and dynamics in such worlds cannot be based on an action-reaction list of correspondences but must be based on an open-ended evolution in which genetics and self-organization allow the survival of the overall system.

Impressive progress has recently been made in the field of reproduction of autonomous behavior and more particularly in the field of A-Life. The historical dream of Von Neumann [1] of self-reproducing machines today seems nearer at hand than ever: recent advances have produced simulated machines capable of developing increasing complexity with each successive self-reproduction cycle (see the "Langton loop" [2]). The first attempts at creating self-reproducing machines have recently shown interesting results [3]. The well-known Tamagotchi, a very-low-technology game, had enormous success due merely to its clever emulation of some simple life functions. Sony's AIBO robot toy [4] is probably the first commercially distributed example in what may become a long series of toys that can learn to play and develop autonomous behavior. Current experimentation on developing communicative language in robots [5] will probably give additional answers to the question of whether artificial beings could develop "culture." An interesting further development has occurred in the application of artificial societies in order to develop strategies for controlling complex industrial processes [6].

The above-mentioned basic questions raised by all these ideas, stimuli and advances are being answered today by two camps of A-Life thinkers: the weak A-Lifers (who see A-Life entities as just a simulation of real life) and the strong A-Lifers (who consider these entities to be realizations of a new and real life form). One of the main contributions to this discussion comes from Maturana and Varela's concept of autopoiesis [7], that is, a network of production processes in which the function of each component is to participate in the production, or transformation, of other components of the network. In short, living systems are characterized by the production/transformation of "themselves." In A-Life-based systems, it is often extremely difficult to assess this property, and most of the time A-Life

Fig. 1.

Volute, picture generated by the interactive installation Re-lazioni Emergenti. (All images ©Mauro Annunziato and Piero Pierucci) Starting from one or more ancestors, a number of offspring are generated with each iteration. Individuals, represented by filaments, struggle for food, space and energy. Each individual has its own character specified in terms of genetic code. After thousands...