Anthropic faces $41 million lawsuit over Elvis Presley songs
Rights holders are suing the Claude AI developer for allegedly using 274 copyrighted compositions to train its models without permission.
The short version
- Rights holders allege Anthropic copied 274 copyrighted compositions, including hits like 'Can't Help Falling in Love'.
- The lawsuit seeks up to $41.1 million in statutory damages for willful copyright infringement.
- Plaintiffs claim the AI can generate unauthorized derivative works that retain recognizable hooks and structures.
- The complaint also includes allegations of violating the Digital Millennium Copyright Act by removing copyright-management info.
Why it matters
This case highlights the ongoing legal battle over whether training AI on copyrighted material constitutes 'fair use.' A ruling could determine how much control music publishers have over the data used to build commercial AI systems.
The legal complaint
The rights holders of Elvis Presley's music have filed a federal lawsuit against Anthropic, the developer of the Claude AI model. The complaint, filed in the U.S. District Court for the Northern District of California, accuses the company of using 274 copyrighted compositions without permission. These works were allegedly used to train the artificial intelligence systems.
Among the specific songs identified in the filing are well-known tracks such as "Don't Be Cruel," "All Shook Up," and "Return to Sender." The plaintiffs, which include ABG EPE IP LLC and various music publishers, are seeking up to $41.1 million in statutory damages. They argue that Anthropic copied protected lyrics from online sources to build its training datasets.
The lawsuit further alleges that Anthropic failed to secure any licenses to reproduce this material. Instead, the company is accused of retaining unauthorized copies within its internal systems. The plaintiffs are seeking a permanent injunction to stop this unauthorized use and are asking for the destruction of infringing copies.
Allegations of derivative works
Beyond the initial training of the models, the lawsuit claims that Anthropic's technology enables the creation of unauthorized derivative works. The complaint suggests that users can prompt the AI to generate rewritten versions of copyrighted songs. These outputs are said to retain recognizable structures, hooks, and lyrical arrangements even when the wording is changed.
The rights holders argue that Anthropic's safeguards are insufficient to prevent these types of outputs. They claim that users can easily bypass restrictions by requesting specific adaptations of existing famous songs. This capability allows the AI to produce material that closely resembles protected compositions without compensation to the original creators.
Attorney Richard Busch, who represents the Presley rights holders, stated that his clients are "deeply protective of these iconic songs." The legal action aims to hold the company accountable for building its business through what the plaintiffs describe as unauthorized copying on an extensive scale.
Copyright and fair use
A central theme in this litigation is the legal definition of fair use in the context of generative AI. Fair use is a doctrine that allows for the unlicensed use of protected works based on factors like the purpose of the use and the effect on the market. AI developers often argue that training models falls under this category because it creates a new type of technology.
Conversely, music publishers maintain that copying entire works to develop commercial systems requires explicit authorization. They argue that the scale of the copying and the commercial nature of the AI models weigh against a fair use defense. This case joins a growing list of disputes involving publishers from Universal Music, Sony, and Warner.
The lawsuit also includes claims under the Digital Millennium Copyright Act. These allegations suggest that Anthropic circumvented technological measures meant to protect copyrighted material. The plaintiffs also claim the company removed or altered copyright-management information, such as identifying details associated with the works.
What comes next
Anthropic will have the opportunity to respond to these allegations as the case proceeds through the federal court system. The outcome will depend on whether the plaintiffs can successfully establish infringement and meet the requirements for enhanced damages. If the court finds the infringement was willful, the damages could reach as high as $150,000 per work.
The plaintiffs are also requesting a full accounting of the training materials used by Anthropic. They want to see exactly what data was ingested into the models to develop Claude. This could potentially force the company to disclose more about its data collection practices and the specific sources it relies on.
This case is part of a broader trend of litigation involving various types of media. Other lawsuits are currently seeking to stop the use of journalism and other protected materials in AI training. The resolution of this Presley-related case may provide a significant precedent for how music rights are handled in the age of generative AI.
What to watch
- Whether the court rules that training AI on copyrighted lyrics constitutes fair use.
- If Anthropic is forced to purge specific copyrighted works from its training datasets.
- How other music publishers respond to this specific litigation strategy.
Try it
Check the official website of the U.S. District Court for the Northern District of California to look for the specific filing details of this case.
Sources and links
Liked it? Say why
Add a line for other readers.
No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.
Listen to the AI Wire headlines
Listen to the AI WireListen to our daily AI shows every morning: on your commute, in the shower, wherever you are
Listen to the daily showsFirst reported by Law Commentary · Read the original. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.
