Round Hill Music has filed lawsuits against Anthropic and Suno, claiming the AI companies used more than 500 songs from its catalog to train their models without permission. The legal action underscores a widening rift between copyright owners and the fast-growing AI sector, and could change how AI firms handle licensed material.
The claims at the heart of the cases
The lawsuits, filed in separate actions, allege that both Anthropic and Suno reproduced and used Round Hill's songs as training data for their AI systems. Round Hill, which owns or controls the rights to a catalog of over 500 compositions, says the companies did not seek a license or pay for the use. The publisher is seeking damages and an order to stop the alleged infringement.
Neither company has yet filed a formal response, and the details of their defenses are not public. But the core question is straightforward: does training an AI on copyrighted music without permission count as infringement? Round Hill argues it does. The companies may argue that the use is transformative or falls under fair use, though no such arguments have been filed yet.
A familiar battle over training data
The dispute is part of a broader wave of litigation as AI developers scrape the internet for text, images, and audio to teach their models. Publishers, artists, and news outlets have sued over the use of copyrighted works in training sets. Round Hill's cases target two very different types of AI companies: Anthropic, a general-purpose AI firm, and Suno, which focuses on music generation. Both rely on large datasets that often include copyrighted material.
For music publishers, the stakes are high. Songs are the core asset, and if AI models can be trained on them without payment, that could undercut the value of licensing deals. Round Hill's catalog includes songs from a range of writers and artists, though the specific titles are not listed in the initial filings.
What the outcome could mean
The rulings could set a precedent for how AI companies treat music copyrights. If the courts side with Round Hill, AI firms may need to negotiate licenses for any copyrighted music they use in training. That would be a significant shift from the current practice, where many companies argue that using publicly available data is permissible.
On the other hand, if the AI companies prevail, it could open the door to more aggressive data collection without compensation. Either way, the cases are likely to be closely watched by both the music industry and the AI sector.
The lawsuits are still in their early stages. Round Hill is seeking injunctive relief and damages, and the defendants will have to answer the claims in court. The outcome won't come quickly, but the legal arguments could shape how AI and copyright law interact for years.
