Theoretical analysis reveals divergent conclusions on machine creativity, indicating that agency separates human creative processes from computational outputs.
This paper discusses “The computability of creativity” in the light of alternative definitions of creativity. An applied approach favours definitions in terms of externally assessed properties of ideas, typically, of their novelty and value, with no process requirements. A pure science approach emphasises internally assessed idea properties, coupled with process requirements, particularly for intentionality and agency. On the pure science “process” view, creativity requires agency, which computer programs (computable processes) do not possess. It is proposed then that the answer to the question of the computability of creativity is “No” on the pure science process approach—as creativity in this approach necessarily involves agency, which is outside the formal, abstract computational realm. On the applied approach, with creativity defined in terms of externally assessed properties of ideas, it is proposed that some computed responses are externally creative (e.g. GPT system answers to Alternative Uses tasks).
No takes yet. Share an insight, caveat, or question.
Ken Gilhooly (2025) studied this question.
Synapse has enriched 4 closely related papers on similar clinical questions. Consider them for comparative context: