Elvis catalogue dispute brings another AI copyright case
Authentic Brands Group, the licensing organization that owns rights to Elvis Presley’s music and likeness, has sued Anthropic, alleging the artificial intelligence company used 274 Presley songs to train its models without permission.
The lawsuit was filed on Friday and centers on songs including Can’t Help Falling in Love, All Shook Up and A Little Less Conversation. The plaintiffs are seeking statutory damages of $150,000 for each allegedly infringed track, an amount the report places at roughly $41 million if applied across all 274 songs.
The allegations have not been adjudicated. Still, the case adds a recognizable artist catalogue to the growing legal fight over how AI developers obtain, use and reproduce copyrighted creative material.
What the complaint alleges
According to the lawsuit, Anthropic allegedly used Presley works as part of training for its Claude AI model. The complaint also alleges that Claude has generated infringing copyrighted song output and contends that measures intended to limit such output have been ineffective.
Those claims sit at the heart of a wider unresolved question for generative AI: whether the use of copyrighted material in model development requires licensing or permission, and what legal responsibility follows when a model can produce material that resembles protected works.
Authentic Brands Group’s complaint uses unusually broad language, alleging that Anthropic’s business model rests on copyright infringement at unprecedented scale. That is the plaintiffs’ characterization, not a court finding. Anthropic will have an opportunity to respond, and the legal process will determine whether the asserted use and outputs infringe the claimed rights.
Why this case stands out
AI developers have already faced copyright claims from the Recording Industry Association of America and major music companies, including Sony, Universal and Warner Records. This case is notable because it is described as one of the first major legal challenges brought directly by an artist’s estate or rights holder against an AI developer.
That matters beyond the Presley catalogue. Music rights are often administered through a mix of owners and licensing arrangements, and a ruling or settlement in a high-profile case could influence how artists’ estates, labels and technology companies approach licensing talks in the future.
The dispute also brings together two issues that are often discussed separately: the material used to train a model and the material a model may produce after deployment. Copyright plaintiffs have increasingly focused on both. Training claims raise questions about copying in the development process, while output claims ask whether a system can reproduce or enable reproduction of protected expression.
The money at stake
The damages request illustrates why individual-catalogue cases can carry substantial financial stakes. At $150,000 per track, the figure cited by the plaintiffs adds up quickly even before considering litigation costs, possible injunction requests or any future licensing negotiations.
But the most consequential outcome may not be the final dollar amount. A court decision could clarify how legal rights apply to music used in AI development, or it could leave important questions unresolved while the parties reach a settlement. Either path would be watched closely by rights holders and AI companies operating in a market where training data remains a central source of legal and commercial tension.
For artists’ estates and licensing organizations, the case is also an effort to preserve control over works whose value extends far beyond streaming or traditional synchronization licenses. Songs associated with an artist as culturally prominent as Presley are not only recordings or compositions; they are enduring commercial and identity assets. The complaint’s inclusion of Presley’s likeness rights holder reinforces that broader rights-management context, even though the reported allegations focus on songs.
AI companies face an expanding legal map
Generative AI systems are built by processing enormous amounts of material, and disputes over the origin of that material have emerged across text, images, code and music. The Presley case is another signal that rights holders are prepared to test those questions through litigation rather than wait for industry-wide norms to emerge.
It is too early to say whether this lawsuit will establish a precedent. The reported case has only been filed, and its specific facts, Anthropic’s response and the court’s treatment of the claims will all matter. A single case may not settle the law for the entire industry, particularly where different kinds of copyrighted work and different technical practices are involved.
Even so, the complaint increases pressure for clearer answers. AI developers need predictable rules for acquiring and using content, while creators and rights holders seek a workable way to control or be compensated for use of their catalogues. The conflict is no longer confined to abstract debate over data. It is now being framed around specific songs, a major cultural legacy and a detailed demand for damages.
Key points
- Authentic Brands Group alleges Anthropic used 274 Elvis Presley songs to train AI models without permission.
- The plaintiffs seek $150,000 per allegedly infringed track, or about $41 million in total.
- The suit is among the first major AI copyright challenges brought directly by an artist rights holder.
- The allegations remain unproven and will be tested through the legal process.
This article is based on reporting by The Guardian. Read the original article.
Originally published on theguardian.com
