The production of novel and interesting entities may seem like a good starting point for a definition of creativity, but I argue that any definition of this kind does not take sufficient regard of the way in which an entity came to be produced. Only certain methods of production can lead to attributions of creativity. I suggest that creativity can be thought of as a two stage process of generation and evaluation, and go on to stress the role of evaluation. Two examples from the arts are provided to support this position. With a sketch theory in place, I turn to the notion of artificial creativity. I look at some philosophical worries that arise when attributing to machines abilities which are usually considered to be exclusively human, or at least biological. Some current computer systems are mentioned, and I indicate why I think some are better proto-models of creativity than others. Finally, I speculate on how we might move further towards artificial creativity.
Download compressed postscript file