You Can Call Me AI: An Experimental Study on How Emotional Response is Affected by Perceiving Music to be Composed by Artificial Intelligence

Abstract: This thesis examines how emotional response is affected by perceiving music to be composed by artificial intelligence (AI). Through a theoretical framework, based on previous research in music and emotion, the aim is to contribute to the research on AI music by focusing on cinematic music and how emotional strength is affected by the perceived composer, AI, or human. We created an experiment based on three musical pieces in the cinematic genre. Two of the pieces were composed by AI and one was composed by humans. Participants were asked to guess which piece(s) were composed by AI and which piece(s) were composed by Humans. They were also asked to report the emotional strength of the "basic" and "aesthetic emotions" they experienced. The empirical results seem to indicate that there is no difference in emotional strength between the perception of AI music (AIM) and human music (HM). In addition, gender, musical interest, technological interest, and attitude appear to not influence the relationship between perception and emotional strength in any direction. We also found that people, in general, were unable to distinguish AIM from HM. Our first implication is that companies that use music to influence their consumers can freely explore AIM in their emotional communication. Our second implication is these companies can thus benefit from economies of scale, lower labor costs, and new innovative business models, without the risk of interfering with emotional strength. The avenues for future research are diverse and much remains to be discovered in this technological shift in music.