Muhammad Ibadurrohman
The escalating legal standoff between the titans of the recording industry and the vanguard of artificial intelligence reached a new intensity on Friday, September 18. In a filing that expands the scope of an ongoing copyright battle, Universal Music Group (UMG) and Sony Music Entertainment (SME) have targeted the generative AI company Suno with a fresh lawsuit. This action incorporates over 61,000 specific musical compositions into the litigation, signaling that the major labels are far from satisfied with the recent pivot in Suno’s business strategy.
The Conflict Over Model v6 and Licensing
The core of this new filing centers on Suno’s recently released AI model, v6. This release was accompanied by a marketing campaign highlighting partnerships with several industry stakeholders, most notably Warner Music Group. Suno had framed these partnerships as a template for a harmonious future, suggesting that the divide between generative AI developers and the creative industries could be bridged through licensing agreements.
However, UMG and Sony have explicitly rejected this narrative. By filing this new suit, the labels argue that the release of v6 does not cure the alleged underlying illegality of Suno’s training practices. The labels contend that the "poisoned tree" metaphor remains apt because the new model, while potentially utilizing some licensed data, continues to rely on the "synthetic data" and user interactions derived from previous, allegedly unauthorized training cycles.
A Chronology of the Legal Entanglement
The friction between the music industry and AI developers is the defining legal struggle of the 2020s. The timeline of this specific conflict illustrates a rapidly shifting landscape:
- Mid-2024: UMG, Sony, and Warner Music Group collectively filed an initial copyright infringement lawsuit against Suno and Udio, alleging that the companies scraped vast catalogs of protected recordings without permission to train their models.
- November 2024: Warner Music Group reached a landmark settlement with Suno, establishing a licensing framework that allowed for the legal use of its catalog. This created a notable fracture in the united front initially presented by the three major labels.
- September 2025 (Approximation): Suno launched its v6 model, promoting it as a collaborative effort with the music industry.
- September 18, 2026: UMG and Sony filed the new, expansive lawsuit, explicitly aiming to address the thousands of songs they claim remain infringed upon and to challenge the legitimacy of the v6 training methodology.
Legal Arguments: The "Fair Use" vs. Market Harm Debate
At the heart of the dispute lies the contentious legal doctrine of "fair use." Suno and other AI laboratories argue that the process of training models on existing works constitutes a transformative use that does not infringe on the original copyright. They posit that the models learn the "patterns" and "structures" of music rather than creating derivative copies.
Conversely, the plaintiffs argue that the scale of the ingestion—millions of copyrighted works—is not a transformative endeavor but an industrial-scale substitution. A pivotal point in the new lawsuit is the argument regarding market harm. Historically, AI firms have claimed that there is no market for "training data," and therefore, their unauthorized use of music cannot harm a market that does not exist.
UMG and Sony argue that Suno’s recent licensing deals have effectively destroyed this defense. By signing agreements with Warner Music Group, Suno has acknowledged that there is a legitimate, paid market for the right to use music in AI training. The labels argue that Suno cannot simultaneously claim it is not harming a market while actively participating in that same market as a licensee.
Data and Economic Implications
The economic stakes are profound. The music industry is currently undergoing a structural transformation as AI-generated content becomes more accessible. According to various industry estimates, the volume of AI-generated tracks uploaded to streaming platforms has increased by triple-digit percentages year-over-year.
Legal experts observing the case note that the "synthetic data" argument raised by the labels—the claim that v6 is trained on the outputs of models that were themselves trained on stolen data—could set a significant legal precedent. If the court finds that "laundering" data through iterative training cycles does not cleanse it of copyright infringement, it could force a massive restructuring of how AI models are developed and verified.
Corporate Responses and Statements
In a statement provided to the media, a Suno spokesperson maintained that the company’s operations remain lawful and that the lawsuit is fundamentally misguided. "v6 was trained on content licensed from our partners, interactions including creations and preference signals from our community, and the accumulated learnings from our team," the spokesperson stated. The company further emphasized its commitment to a future where AI serves as a tool for, rather than a replacement of, human creativity.
The plaintiffs, however, remained resolute in their filing. Attorneys for the labels wrote: "Suno’s wholesale theft threatens the entire music ecosystem and the numerous people it employs. It also degrades the rights of artists to control their works, determine whether future uses of their works align with their aesthetic and personal values, and decide the products or services with which they wish to be associated."
The "Shulman" Factor: CEO Rhetoric as Evidence
The lawsuit makes extensive use of public statements made by Suno CEO Mikey Shulman. By citing interview transcripts and podcast appearances—including Shulman’s comment that many people do not actually "enjoy" the labor-intensive process of making music—the labels are attempting to establish a pattern of intent.
The plaintiffs argue that this rhetoric reveals a disregard for the creative process that copyright law is specifically designed to protect. By framing human labor as "friction" to be engineered away, the labels contend that Suno’s business model is inherently adversarial to the interests of professional musicians, songwriters, and producers.
Broader Industry Impact
The outcome of this case will likely have implications far beyond the music industry. It serves as a bellwether for the entire creative economy, including publishing, visual arts, and film. If major labels can successfully argue that "synthetic data" training is a violation of copyright, it would impose a significant burden on the AI sector to document the provenance of every byte of data used in model development.
Furthermore, the "saturating the market" claim introduces a new dimension to the litigation. The labels are not just arguing that their individual copyrights are being infringed; they are arguing that the collective presence of AI-generated music is degrading the value of the human-made music market. This "market dilution" theory is a complex legal concept that, if upheld, could fundamentally alter how intellectual property is valued in the digital age.
As the case moves forward, the legal community will be watching to see if the court permits the inclusion of the 61,000 new songs within the context of the existing litigation or if the procedural requirements for separate filings will force an even more fragmented and prolonged legal battle. For now, the divide between the major record labels and the AI music sector remains as wide as ever, with the fundamental question of ownership in the age of automation still awaiting a definitive judicial answer.
The struggle is, at its core, a clash between two visions of the future: one where technological progress is integrated into the existing legal and royalty structures of the music business, and one where the democratization of music creation through AI is permitted to disrupt, and potentially replace, the established economic model of the recording industry. With the filing of this latest lawsuit, it is clear that the industry’s largest players are prepared to litigate the matter to its furthest conclusion, regardless of how many versions of the underlying technology are released.
