Composers' Evaluations of an AI Music Tool: Insights for Human-Centred Design

TL;DR

The paper investigates how to advance human-centered design for Generative AI in music composition by conducting a qualitative study with professional composers interacting with a baseline transformer-based jazz-variation generator (Music Overpainting). It highlights trust, transparency, traceability, ethical data use, and controllability as central design concerns, proposing a continuous feedback loop to guide both practical tool design and foundational GenAI research questions. The findings underscore the value of end-user collaboration to align AI-assisted creativity with real-world workflows and suggest extending the approach to other GenAI domains. Overall, the work offers concrete guidance for building transparent, controllable, and ethically aware GenAI tools that partner with composers rather than replace them.

Abstract

We present a study that explores the role of user-centred design in developing Generative AI (GenAI) tools for music composition. Through semi-structured interviews with professional composers, we gathered insights on a novel generative model for creating variations, highlighting concerns around trust, transparency, and ethical design. The findings helped form a feedback loop, guiding improvements to the model that emphasised traceability, transparency and explainability. They also revealed new areas for innovation, including novel features for controllability and research questions on the ethical and practical implementation of GenAI models.