When is a computer being creative?

Dr Graeme Ritchie

Tuesday 05 February at 11am

Abstract

Recent years have seen the implementation of computer programs which propose mathematical conjectures, construct jokes, write poems, devise stories, etc. This has given new prominence to the idea of a program being "creative". Discussions of the creativity of computer programs often concentrate on broad and profound philosophical questions such as: "can a computer ever be creative?". We claim that if these very general questions are to be answered empirically, then we first have to formulate a way of answering a more prosaic and small-scale question: "has this behaviour by this computer program demonstrated creativity?" That is, we need explicit, formal criteria which we can apply to the behaviour of a computer program and which will count as evidence of creativity. We propose a formal (and rather simplified) outline of the relevant attributes of a potentially creative program. Based on this, we posit a number of criteria that could be applied to assess the extent to which the program has behaved creatively. Our guiding principle is that the question of what computational mechanisms might lead to creative behaviour is open and
empirical, and hence we should clearly distinguish between judgements about creative achievements and theoretical proposals about potentially creative mechanisms. The intention is to focus, clarify and make more concrete the debate about creative programs.

Short resume

Dr Graeme Ritchie is a Senior Lecturer in the Division of Informatics, University of Edinburgh, Scotland, and from October 2001 to September 2002 he is a Leverhulme Research Fellow (visiting Macquarie University until April 2002). He has worked in computational linguistics since 1973, and has published a wide range of papers, on topics including parsing, morphology, semantic interpretation, question-answering systems, computer creativity, and humour.

Back to HAIL Home Page