Gordon A. Gow (U Alberta Arts) has posted “SONGPRINT: A Voluntary Labelling Framework for AI-assisted Music Creation” on SSRN. Here is the abstract:
As generative artificial intelligence (AI) tools increasingly permeate music creation, questions of transparency, authorship, and creative practice take on renewed urgency. This paper introduces SONGPRINT, a multi-dimensional labelling framework designed to assist musicians in reflecting on and disclosing AI’s role in the compositional process. Prompted in part by the recent Velvet Sundown controversy—a high-profile case of undisclosed AI-generated music—the project advances SONGPRINT as a conversation starter: a modest prototype intended to encourage critical engagement with evolving norms of attribution, labour, and listening in an AI-assisted musical landscape.
This paper is informed by my practice as a part-time songwriter and musician. As a product of the analog generation, I have witnessed firsthand the evolution of music production, from magnetic tape and discrete transistor-based audio components to the advent of digital audio workstations and integrated software-based signal processing, and now to the emergence of generative music platforms. Over the past year, I have conducted a series of experiments using platforms like Suno to explore how AI can serve as a creative collaborator in shaping melodies, interpreting lyrics, and producing polished musical outputs. These experiences inspired the development of SONGPRINT as both a reflective and practical framework for discussing the dialectic between generative AI and human creativity in music production.