CLSep 8, 2025

No Encore: Unlearning as Opt-Out in Music Generation

arXiv:2509.06277v22 citationsh-index: 1
Originality Synthesis-oriented
AI Analysis

This addresses ethical and legal concerns in creative industries, but it is incremental as it explores existing unlearning methods on a new domain.

The paper tackles the problem of preventing inadvertent usage of copyrighted content in AI music generation by applying machine unlearning techniques to a pre-trained Text-to-Music model, providing preliminary results and insights into the challenges of this approach.

AI music generation is rapidly emerging in the creative industries, enabling intuitive music generation from textual descriptions. However, these systems pose risks in exploitation of copyrighted creations, raising ethical and legal concerns. In this paper, we present preliminary results on the first application of machine unlearning techniques from an ongoing research to prevent inadvertent usage of creative content. Particularly, we explore existing methods in machine unlearning to a pre-trained Text-to-Music (TTM) baseline and analyze their efficacy in unlearning pre-trained datasets without harming model performance. Through our experiments, we provide insights into the challenges of applying unlearning in music generation, offering a foundational analysis for future works on the application of unlearning for music generative models.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes