Bias in Favour or Against Computational Creativity : A Survey and Reflection on the Importance of Socio-cultural Context in its Evaluation
Year: 2023
Type: conference-paper
Abstract: This paper surveys 27 published studies exploring bias in the evaluation of computational creativity.These studies look specifically at the involvement of AI, in generating music, images and graphics, poetry, and news articles.While some have found evidence of bias (43%), others find no bias (27%), or show modulation of bias through socio-cultural factors (30%) resulting in a lack of consensus on this issue.We argue for the importance of taking into account socio-cultural context when considering such biases in these creative pursuits.What styles do the artefacts belong to?Who are the participants involved in the study, and what are their relationships to the styles at hand?We discuss the implications of such considerations for future research in computational creativity.We propose some safeguards when conducting a study on bias in the evaluation of computational creativity, and propose directions to study more specifically when, and with whom it can be observed.Task Paper Style/Topic N Reported conclusion Music Dahlig and Schaffrath (1998) German Folk songs 432 Varied Moffat and Kelly (2006) Contemporary, free jazz, Bach 20 No bias Friedman and Taylor (2014) Classical 58 No bias
Authors: Déguernel, Ken, Sturm, Bob
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
Funder: H2020 European Research Council
OpenAlex ID: W4392278874
