Leave your comment on sections Summary to Defining Creativity using this entry.
Artificial Creativity (or computational creativity) is a branch of Artificial Intelligence that deals with the development and exploration of systems that exhibit creative behavior. This includes systems capable of such things as scientific invention, visual artistry, music composition and story generation.
Artificial creativity is a branch of artificial intelligence (AI) that deals with constructing machines with creative abilities. Although tightly intertwined with general AI research, recent years have brought increased focus on creativity as an independent branch of AI due to its complexity and difficulty of implementation.
In the 1956 Dartmouth Proposal for Artificial Intelligence, in a seven-point list of the Artificial Intelligence problem, creativity and randomness were specifically addressed:
7. Randomness and Creativity
A fairly attractive and yet clearly incomplete conjecture is that the difference between creative thinking and unimaginative competent thinking lies in the injection of a some randomness. The randomness must be guided by intuition to be efficient. In other words, the educated guess or the hunch include controlled randomness in otherwise orderly thinking.
In addition, originality, invention and discovery was mentioned on several occasions throughout the proposal.
As with natural intelligence, creativity has remained extremely hard to define. Creativity exhibits itself and effects behavior to a large degree (especially in humans), which makes it very hard to identify its distinguishing features and nature. No empirical definition or authoritative perspective on creativity exists within scientific circles.
A temporary solution has been to define artificial creativity by reference to its natural counterpart: “Artificial Creativity is concerned with producing behavior in machines, which if acted out by a human, would be deemed creative“. This has also been noted in artificial intelligence research in general, where a system is deemed intelligent if it exhibits traits similar to natural intelligence.
One of the ultimate goals of artificial creativity research is to replicate creativity as it appears in humans. A strong emphasis has therefore been on creating systems which mimic our abilities and specifically those of artists, as creativity is easily identified in the arts. This includes abilities such as music composition, painting and storytelling.
Pioneers of the field such as Margaret Boden have proposed definitions centered around human abilities, which in recent years has been garnering support and been formalized at the computational creativity workshops (see links). However, the complexity of human activities has led research to fragment into focus on specialized systems for particular activity, rarely encompassing systems integration to any significant extent.
This divide & conquer approach to human creativity has arguably proven limited for explaining the origins and nature of creativity, and how its multifaceted processes are intertwined with intelligence. It has been proposed that research on creativity in less complex animals might prove more fruitful for understanding it, and perhaps even provide a necessary foundation for advancing general artificial creativity.
A fundamental problem in implementing creative systems
- For further information on this section, see the post Why it’s hard to make machines think original thoughts
Not only do we lack understanding of our own creative mechanisms, but the basics of computer programs seem to oppose the idea of achieving unbound originality.
To properly explain this problem, how programming seems to oppose creativity, we must understand what computer programs are: instructions. A set of steps the computer executes. Typically, when we create computer programs we specify a certain problem and in turn devise a set of instructions that addresses this problem.
Instructions are what defines a programs behavior and outcome. We impose restrictions — a confined set of instructions out of all the possibilities in the world. This is necessary for the system to do anything at all. For example: A goal limits the objectives of a system and thereby helps organize how the system will behave1.
The basics of programming require explicit designing of mechanisms that produce certain outcomes. By giving these explicit instructions the potential of a program acting in novel ways is decreased, since clearly it means that it’s known beforehand how the system will behave. The instructions that define programs (and make them work) are in turn the exact reason it’s hard to produce surprising, novel and interesting ideas.
This problem is being slowly overcome with the use of neural networks, genetic algorithms and other complex systems and complex adaptive systems.
Please note: This is page is a supplementary to Wikipedia’s. It should not be considered stand-alone or complete.
Leave your comment on the section External Links using this entry.
Conferences and Consortiums
Books and Literature
- The Creative Mind (Boden, M. A., Weidenfeld/Abacus & Basic Books, 1990; 2nd edn. Routldge, 2004;)
Labs, Institutes and Companies
- Center for Analysis and Design of Intelligent Agents at Reykjavik University (CADIA)
- Creative Systems Area of CISUC
- Creative Systems Lab at the University of Sussex
- Imagination Engines Inc.
- AAAI Creativity Section
- Artificial Creativity on Wikipedia
- Design Computing on Wikipedia
- List of people working on Generative Art
Related Think Artificial posts
- External Links section added, September 19th, 2007
- Original introduction to this article, September 5th, 2007
Please note: This is a living article. It is regularly updated and should not be considered finalized. Feel free to send along questions or comments you might have on the artificial creativity!