From Lawsuits to Licensing
Science and Technology Law Review · 01.11.2026
DOI: 10.52214/60cxsf41 ↗Kurz gefasst
Fachartikel in der Zeitschrift Science and Technology Law Review, online erschienen am 01.11.2026. Verfasst von Yussof Khalilian. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Author: Yussof Khalilian '28 | Editor: Austin Goetz AI-generated music is rapidly expanding, raising questions about the copyrighted works used to train the models behind it. [1] AI music companies defend their training practices as fair use, arguing that learning from existing recordings resembles a musician learning by listening. Record labels and artists dispute that analogy, contending that these practices involve unauthorized copying of their work. [2][3] This disagreement has fueled high-profile litigation over the boundary between technological innovation and copyright infringement. Among the most closely watched disputes were two lawsuits brought in June 2024 by major labels, including Universal Music Group and Warner Music Group, against Suno and Udio. [2][3] The labels’ 2024 complaints against Suno and Udio alleged extensive unauthorized copying of their sound recordings. They also offered examples of generated music said to resemble protected recordings and argued that competing AI music would threaten their markets. The AI music companies acknowledged using music available on the open internet but responded that existing music enables new expression. The competing framings were the heart of the dispute, and a court ruling would have decided which one prevailed. Scholars remain divided over the application of fair use to AI training. Mark Lemley and Bryan Casey argue that training should generally qualify when it enables a system to learn unprotected ideas, facts, and patterns rather than reproduce protected expression. On this view, copying for learning serves a different purpose from copying to distribute or replicate a work’s creative content. Jacqueline C. Charlesworth, by contrast, argues that generative AI training exploits protected expression to produce new expressive content, undermining the claim that the training serves a transformative purpose. The labels’ claims, however, led to settlements rather than judicial resolution of the debate. Universal’s October 2025 agreement with Udio combined a compensatory settlement with licenses for recorded music and publishing. Warner’s subsequent agreement with Suno likewise paired settlement with a licensing partnership, turning former litigation opponents into commercial collaborators. These agreements may satisfy both sides of the litigation, but one group was left out of the deals: the artists themselves. The record labels’ shift from suing AI companies to partnering with them leaves the separate question of whether and how creators share in the financial benefits unresolved. Copyright in a sound recording is distinct from copyright in the underlying musical composition, and different parties may control those rights. Musicians may also be entitled to compensation under contracts or collective bargaining agreements, even when they do not own the copyright. A label’s authority to license a recording for AI use therefore does not, by itself, determine what it owes the people who helped create the recording. The June 2026 lawsuit by the American Federation of Musicians (AFM) seeks to ensure that musicians benefit from the licensing agreements between record labels and AI music companies. [4] The union alleges that Universal- and Warner-related entities licensed recordings for AI training and music generation without providing the notice and compensation required by the Sound Recording Labor Agreement. [4] According to the AFM, these activities constitute a “new use” triggering obligations to compensate the covered musicians who performed on them. [4] It is true that the announced agreements promise artists some benefits. Warner’s Suno announcement includes revenue opportunities and control over specified uses of artists’ and songwriters’ names, images, and likenesses. Those commitments deserve acknowledgement. Their existence, however, does not establish whether every contributor receives compensation, whether consent covers training as well as generated outputs, or whether musicians can verify and enforce their entitlements. The AFM’s lawsuit may provide a contractual backstop for musicians whose interests are not fully addressed by these agreements. For that reason, as we await the results of the litigation, there is a great deal on the line for artists and musicians who are the foundation of the music industry. Sources: [1] Vaughn Gendron, A New Frontier: The Music Industry’s Struggle Against Generative AI, 33 U. MIA Bus. L. Rev. 161, 163 (2024). [2] Complaint ¶ 12, UMG Recordings, Inc. v. Suno, Inc., No. 24-cv-11611 (D. Mass. June 24, 2024), ECF No. 1. [3] Complaint ¶ 12, UMG Recordings, Inc. v. Uncharted Labs, Inc., No. 24-cv-04777 (S.D.N.Y. June 24, 2024), ECF No. 1. [4] Complaint, Am. Fed’n of Musicians of the U.S. & Can. v. Warner Music Grp. Corp., No. 26-cv-04760 (S.D.N.Y. June 5, 2026), ECF No. 1.
Zitieren
APA
Khalilian, Y. (2026). From Lawsuits to Licensing. Science and Technology Law Review. https://doi.org/10.52214/60cxsf41
Vancouver
Khalilian Y. From Lawsuits to Licensing. Science and Technology Law Review. 2026. doi:10.52214/60cxsf41
BibTeX
@article{Khalilian2026,
author = {Khalilian, Yussof},
title = {From Lawsuits to Licensing},
journal = {Science and Technology Law Review},
year = {2026},
doi = {10.52214/60cxsf41}
}
EUVIMED