Studying Creativity in AI

Authors

Abstract

Computing research and industry has been working for decades in making computer applications robust and deterministic. Millions were invested in improving software predictability and even its immunity to environment deviations and noise.

It may thus seem a waste of time, and resources, to investigate ways of giving computers the ability to exhibit some kind of creative behaviour. Yet, novelty is a key issue in domains like scientific discovery, theorem proving, technical design and arts, to name a few. Surprising answers are welcome in important classes of applications where huge spaces of possibilities are to be explored in divergent ways.

To some degree, there are other historical factors that may contribute to feel the research in computational creativity as a surprising thing in itself. The views of creativity as some kind of divine inspiration, or originated in an innate intuition talent, i. e. , the inspirational and romantic views of creativity, as Margaret Boden calls them [Boden90], are still very common. Even when people accept views of creativity as a general ability of human intelligence, which is supported by a consolidated record of research on creativity, with routes on Guilford's work on the Structure of Intellect [Guilford 67], even then, people keeps sceptical when considering the possibility of computers behaving in a creative way.

Creativity is by no means a simple theme of research. Nevertheless, it is a fundamental trait of intelligence and one of the most remarkable characteristics of the human mind. It seems thus inconceivable to assume the research of intelligence without considering such an outstanding aspect.

In recent years, computational creativity attracted a growing number of AI researchers who have been working towards the study and proposal of abstract explanation theories, adequate computational models, and applications.

Subject

Creative Systems

Journal

KI - Zeitschrift Kunstliche Intelligenz, Vol. 2004, #1, pp. 45-46, Fachbereich Kunstliche Intelligenz der Gesellschaft fur Informatik e.V., February 2004