Songwriting and AI
(2024)
author(s): Celine Lyng
published in: University of Agder, Faculty of Fine Arts
In this thesis, I explore and research songwriters' experience of using artificial intelligence as a songwriting tool. To do so, three participants have explored the use of two different artificial intelligence tools in their songwriting session, in collaboration with me. One of the tools is used to inspire lyrical writing and one is to create a starting point for chord progressions. To gather data, I have had three collaborative songwriting sessions, where the artists and I use the tools whilst writing a song for their artist project. As well as the sessions, there was held six qualitative interviews, two per artist. The first one was for documenting their prior experience and expectations. The second interview was held after they had received a produced demo. This was used to ask about their experiences writing songs with artificial intelligence. We also talked about the song related to ownership. The thesis questions whether artificial intelligence will make one more productive, too productive, or even if the tools can be perceived as distracting.
In addition to using Artificial Intelligence in three different songwriting sessions, I have researched the ethical dilemmas with using AI, such as copyright infringements, and discussing options for opting out of having one’s work be a part of the training of an AI.
My personal journey highly inspires the thesis through using artificial intelligence for my musical project. By using the same process as the three participants are taking part in, my debut EP, Sugarcoated, was created.
This thesis is approved by SIKT (816719).
