Round Hill Sues Suno, Anthropic Over AI Training Copyright Claims

Independent music publisher Round Hill has filed copyright infringement lawsuits against AI companies Suno and Anthropic, alleging that the defendants used hundreds of copyrighted songs without permission to train their artificial intelligence systems. The publisher, which controls rights to tracks including the Goo Goo Dolls’ “Iris” and Bonnie Tyler’s “Total Eclipse of the Heart,” claims potential damages could exceed $1 billion, arguing there is nothing fair about building multibillion-dollar AI businesses on the backs of uncompensated rights holders.

The suits, filed by prominent music attorney Richard Busch on behalf of Round Hill, include a list of 500 songs that the defendants are alleged to have infringed. Round Hill stated in the filings that it intends to amend the complaint to include potentially ten thousand or more of its musical compositions, with total damages rising accordingly. The publisher directly challenged the AI companies’ arguments about technological progress and societal value, asserting that no justification exists for such progress to come at the expense of copyright holders.

Round Hill further argued that the expediency arguments put forward by Suno and Anthropic collapse when weighed against the significant cash valuations those companies have attained while exploiting what the publisher calls illicit copies of copyrighted works. The core of the legal dispute centers on whether the use of copyrighted material as training data for AI models constitutes infringement or falls under fair use protections. This distinction has become a central battleground as AI companies ingest vast amounts of publicly available content to develop their systems.

The music industry is watching closely, as these cases could set precedents affecting how AI companies access and use copyrighted creative works. Suno already faces separate lawsuits from Universal Music Group and Sony Music Group, reflecting a broader campaign by major rights holders to assert control over how their intellectual property is used in AI development. The outcome will likely hinge on whether courts view AI training as analogous to human learning, which has been largely permissible, or as unauthorized reproduction and distribution of copyrighted material.

The suits arrive at a moment of heightened tension between creative industries and technology companies over the value of data and the boundaries of fair use. For investors and analysts, these cases represent a material risk to AI business models that depend on unfettered access to copyrighted content. A ruling against Suno and Anthropic could force AI companies to negotiate licenses, pay retroactive damages, and fundamentally alter the economics of training large models. Conversely, a finding of fair use would validate the current approach and likely accelerate the pace of AI development across creative domains.

Round Hill’s explicit threat to expand its claims to ten thousand works signals an aggressive legal strategy designed to maximize leverage. The publisher is not merely seeking compensation for past harm but is attempting to establish a framework where rights holders receive payment for the use of their works in training datasets. Whether the courts will accept that framework or treat training as a transformative, non-infringing use remains the central unanswered question with billions of dollars in implications.