Institute of Art Design + Technology
Dún Laoghaire

Mark Butler 

BA [Hons] Creative Music Production

Hey everyone! I am a 21 year old sound engineer, who also works in festival management and conducts research on AI. I have a strong passion for all aspects of the music creation process, from recording to mastering. Over the course of my time at IADT, I've developed skills in creativity, teamwork, project management and research. I co-founded a production company, Kabinc Productions Ltd, in 2023 with a fellow student, where we work with a variety of musicians. Along with working for Festival Republic on the management team for the Irish festival Electric Picnic and concerts at Marlay Park, I am both a web developer and researcher for tech start-up AI:OK.

Project Description

This research project was an examination of digital music streaming platforms’ current relationship with artificial intelligence-composed works. The thesis explores the idea of marking AI compositions on streaming platforms as a means to highlight these songs for consumers. This concept can be likened to the fair trade logo other industries use. AI has become quite prevalent in the music industry during the last 12 months, with generative composition software becoming sophisticated at a rapid pace. An ethical solution, such as a marking system, is needed to put musicians ahead of AI. Laws alone cannot provide this because they are ununified globally and slow to change to implement the necessary protections. Placing an emphasis on this area should be justified to protect the career path of musicianship in the future.

Project Objectives

This project had several objectives, but the main one was to find an answer to the following research question: Can integrating an ethical identifier that highlights AI compositions on digital music streaming platforms positively impact the music ecosystem?
Along with this research question, I also wanted to highlight both the high quality and inconsistency of generative AI composition software. AI already dominates certain genres of music, such as meditation and lo-fi study, and it is becoming increasingly harder to distinguish human and AI compositions. However, much of the work produced by AI at present is inconsistent in terms of quality.
To achieve these objectives, interviews took place with key figures dealing with both AI and music, and two listening experiments were conducted with the public.

Project Outcomes

Overall, the project had several interesting outcomes. It was deemed that implementing an AI identifier to solely highlight AI compositions on digital streaming platforms was too complex a challenge. AI can be involved in too many different parts in the production of a song and would be too hard to identify. Instead, a badge that sets standards around the use of AI and ethics is a more appropriate option. This badge, in theory, could be obtained by companies, artists, streaming platforms, etc., if they adhere to the ethical standards outlined by the badge. This can act as a mark of trust for consumers regarding AI.
In terms of AI-generated music, experiment A demonstrated the gap between distinguishing music composed by AI and humans is diminishing. Only just over a quarter of participants correctly identified the AI composition. In experiment B, the AI composition was sub-par in quality and was easily identified by participants with two thirds choosing the correct composition. This highlights the inconsistent nature of AI generative software.

Thesis Title

AI and Music Digital Streaming Platforms: The Effectiveness of Implementing an Ethical Identifier to Highlight AI Compositions