Towards Responsible AI Music: an Investigation of Trustworthy Features for Creative Systems

Copyright and training data transparency are critical concerns: The paper identifies copyright infringement from vast training datasets as a key ethical issue in generative music AI, highlighting the need for clear disclosure about what data these systems learn from and how artists' work is being used.

Seven dimensions of trustworthiness should guide music AI design: Rather than addressing isolated problems, the research applies the European Commission's framework for trustworthy AI across seven macro requirements, providing a comprehensive approach that considers transparency, explainability, fairness, and other factors together in music generation systems.

Human creativity displacement requires proactive mitigation: The paper identifies potential displacement of human musicians as a major concern, suggesting that responsible AI music systems need built-in safeguards and design choices that preserve meaningful roles for human artists rather than simply replacing them.

Interdisciplinary collaboration is essential, not optional: The research emphasizes that designing responsible music AI requires collaboration across AI experts, musicians, ethicists, legal specialists, and other stakeholders—no single discipline can adequately address the complex challenges alone.

Current music AI systems lack sufficient explainability and fairness mechanisms: The paper points out that existing generative music systems often operate as "black boxes" without clear explanations of how they work or mechanisms to ensure fair treatment of different artists and musical traditions, making accountability difficult.