Sony, UMG Sue Suno Again Over AI Model Trained on User Data

Universal Music Group and Sony Music Entertainment have filed a second copyright lawsuit against AI music platform Suno, targeting the company's recently released v6 model despite Suno's claims that it was trained exclusively on licensed content.

The complaint, filed in federal court on Friday, alleges that Suno trained its new model on "users' interactions" with earlier versions of its software—systems the labels had already accused of mass copyright infringement in their first lawsuit filed in 2024.

The technical argument

The labels' legal theory centers on what they describe as "knowledge distillation and related model-transfer techniques." In this process, a new "student" model learns to reproduce the behavior of a predecessor "teacher" model. Because Suno's earlier models were allegedly trained on unlicensed recordings, the labels argue, the v6 model "inherits the value" of those copyrighted works even if it never directly ingested them.

UMG and Sony point to a September interview Suno CEO Mikey Shulman gave to Billboard, in which he acknowledged the company trained v6 on user interactions with its service. The labels interpret this as confirmation that outputs and preference signals from the allegedly infringing earlier models formed part of v6's training data.

Suno has not yet responded to requests for comment on the new lawsuit.

The licensing paradox

The timing of the suit underscores the fractured state of AI music licensing. Suno launched v6 earlier this month as its first model trained on licensed music, announcing partnerships with Warner Music Group, BMG, and Believe. Warner, which joined the original 2024 lawsuit alongside Sony and UMG, has since settled with Suno and become a partner.

At the v6 launch, Shulman described the licensing deals as "a blueprint for how AI and the music industry can strengthen one another." Yet two of the three major labels remain in active litigation, now escalating their claims to target the very model Suno positioned as its first fully licensed offering.

The market saturation claim

Beyond direct infringement, the labels argue Suno contributes to market oversaturation on streaming platforms with waves of AI-generated tracks that "directly competes with, cheapens, and ultimately drowns out the genuine sound recordings on which the service is built."

This framing extends the legal battle beyond technical questions of training data to broader concerns about AI's impact on the recorded music ecosystem.

Why it matters

The lawsuit reveals a fundamental disagreement about whether AI models can ever truly escape their training lineage. If courts accept the labels' theory that knowledge distillation allows copyright infringement to persist across model generations, it could complicate the path forward for AI companies attempting to transition to fully licensed systems. The case also highlights how licensing deals with some rightsholders don't necessarily resolve disputes with others—a reality that may slow industry-wide AI adoption in music.

The Hollywood Reporter first reported the filing of the second lawsuit.