Human Bias Against AI Composed Music
Overview
The bias against computational creativity is the hypothesis that computationally-generated artifacts are often judged to be less interesting, valuable, and less creative than human-generated ones. Anecdotal evidences include David Cope’s Experience in Musical Intelligence (EMI, 1981), an early style imitation system that was received with controversy. Since then, many researchers, including researchers at the Metacreation Lab, have continued to study the bias of AI composed music and visual art (See Table 1).
Table 1: A review of studies examining AI bias in visual art and music.
Our Study
In our study, we asked 163 participants to rank 10 excerpts of music (selected and ordered randomly from 30 possible compositions) . Five of the excerpts were computer-composed and Five of the excerpts were human-composed. Our ANOVA results did not validate the hypothesis that participants ranking post-revelation will explicit a bias against computer-composed music Additionally, our results did not support the hypothesis that those with higher level of computational literacy would have a lesser bias. In future work we will continue this research on co-creative approaches since unconditioned generation from scratch is not a representative use case of generative models.
Papers and Posters
Zlatkov, D., Ens, J., Pasquier, P. (2023). Searching for Human Bias Against AI-Composed Music. In: Johnson, C., Rodríguez-Fernández, N., Rebelo, S.M. (eds) Artificial Intelligence in Music, Sound, Art and Design. EvoMUSART 2023. Lecture Notes in Computer Science, vol 13988. Springer, Cham. https://doi.org/10.1007/978-3-031-29956-8_20
Pasquier, P., Burnett, A., Thomas, N. G., Maxwell, J. B., Eigenfeldt, A., & Loughin, T. (2016, June). Investigating listener bias against musical metacreativity. In Proceedings of the Seventh International Conference on Computational Creativity (pp. 42-51). (pdf)