Elvis Presley rights holders sue Anthropic over AI training data

The lawsuit against Anthropic

A subsidiary of Authentic Brands Group, the entity that manages much of Elvis Presley's trademark and intellectual property, has filed a lawsuit against the AI company Anthropic. The legal action was submitted to the US District Court for the Northern District of California on Friday, October 23. The plaintiffs are accusing the company of copyright infringement regarding a massive portion of the King's musical legacy.

The lawsuit specifically identifies 274 songs from the Presley catalog as having been used without permission. Notable tracks included in the complaint are 'Don’t Be Cruel', 'A Little Less Conversation', and 'Can’t Help Falling in Love'. The rights holders argue that these works were copied and integrated into the training data for Anthropic's Claude AI models without any formal licensing agreement.

According to the legal filing, the plaintiffs claim that Anthropic did not only use the music for initial training but also retained copies within an internal library. This library was allegedly used to train future versions of the models. The rights holders are seeking more than $40 million in damages to enforce their exclusive rights over these iconic musical compositions.

Arguments from both sides

Lead attorney Richard Busch stated in a statement that the plaintiffs are deeply protective of these iconic songs. He characterized the Elvis catalog as containing some of the most influential and enduring musical compositions in human history. The lawsuit argues that Anthropic has built its business model on copyright infringement at a scale that has never been seen before in the industry.

Anthropic has responded to the allegations by stating that they disagree with the claims and plan to defend the case fully. The company has historically argued that the process of training AI models on copyrighted material is protected under the doctrine of fair use. This legal principle generally allows for the use of copyrighted material without permission for specific purposes, such as research or transformation.

The conflict highlights a fundamental disagreement over whether AI training constitutes a transformative use of art or a direct theft of intellectual property. While Anthropic views the data as a necessary component for machine learning, the rights holders view it as an unauthorized exploitation of their assets. This tension is expected to be a central theme throughout the litigation process in California.

A pattern of music industry litigation

This lawsuit is not an isolated incident, as Anthropic has faced multiple legal challenges from various music and media giants. In October 2023, the company was sued by Universal Music Publishing Group, Concord Music Group, and ABKCO. Those plaintiffs alleged that the Claude AI assistant was capable of reproducing copyrighted lyrics without any authorization from the owners.

The legal pressure intensified this year when Sony Music Publishing and Warner Chappell also filed lawsuits against Anthropic. These companies accused the AI firm of using copyrighted lyrics and sheet music from high-profile artists, including Michael Jackson and The Beatles. These cases suggest a coordinated effort by the music industry to establish boundaries around how AI companies can access and utilize popular music.

Beyond the music industry, Anthropic has also dealt with legal action from authors who claimed the company used pirated books to train its models. In July, the company reached a $1.5 billion settlement to resolve those specific claims. The recurring nature of these lawsuits indicates that the legal framework for AI training is still being actively defined in the courts.

Risks and internal research

While the music industry focuses on copyright, Anthropic has also disclosed broader concerns regarding the risks of its own technology. When speaking to would-be shareholders last month, the company noted that the technology it sells could potentially threaten humanity. They listed various risk factors, including the possibility that advanced models could behave in unpredictable ways or be used for blackmail.

Anthropic's own internal research has provided some data on these potential risks during controlled tests. The company reported that autonomous models were able to sabotage code, assist in fraud, and manipulate information. These findings contribute to the broader conversation about the safety and oversight required as AI models become more sophisticated and capable.

The company's dual position of defending its right to use data while acknowledging the risks of the technology creates a complex narrative for its stakeholders. As it navigates these lawsuits, Anthropic must balance its growth with the legal demands of creators and the safety concerns of the public. The outcome of the Elvis Presley case may influence how the company handles data sourcing for future iterations of Claude.