Javascript must be enabled for the correct page display

The Impact of authorship information on perceived creativity of AI-produced music: The moderating role of musical sophistication

Seredjuk, Maurice (2025) The Impact of authorship information on perceived creativity of AI-produced music: The moderating role of musical sophistication. Bachelor thesis, Psychology.

[img]
Preview
Text
BachelorThesis-s5168481.docx.pdf

Download (358kB) | Preview

Abstract

The study examined whether authorship information influences the perceived creativity of AI-composed music and whether the listener's degree of musical sophistication moderates this effect. Previous research has frequently shown that people evaluate AI-labeled art more negatively, a tendency referred to as the “AI-composer bias”. Eighty participants listened to the same AI-composed pop track but were randomly assigned to one of three authorship conditions: AI-composed, a collaboration condition between a human and AI, and a fully human-composed condition. Participants then rated how creative they perceived the song. We measured perceived creativity by using adapted items from the Creative Product Semantic Scale and assessed the level of musical sophistication by using a short version of the Goldsmiths Musical Sophistication Index. Contrary to our expectations, we found no significant main effect of authorship information on perceived creativity ratings, nor was there a significant interaction between authorship and musical sophistication. However, we found a significant main effect of musical sophistication. Participants higher in musical sophistication rated the music as less creative, regardless of the label. Additionally, we found a conditional effect: Participants with low musical sophistication rated the collaborative condition as significantly less creative than the human one. These findings suggest that the listeners’ background influences creativity judgements more than authorship information alone and challenge the assumption of a universal “AI-composer bias”, implying a possible shift in how audiences engage with AI-labeled music beyond previously established biases. Keywords: AI-Composed Music, AI-Composer Bias, Perceived Creativity, Authorship Information

Item Type: Thesis (Bachelor)
Supervisor name: Meerholz, E.W.
Degree programme: Psychology
Differentiation route: None [Bachelor Psychology]
Date Deposited: 09 Jul 2025 08:51
Last Modified: 09 Jul 2025 08:51
URI: http://gmwpublic.studenttheses.ub.rug.nl/id/eprint/5380

Actions (login required)

View Item View Item