Anthropic faces lawsuit over using Elvis Presley songs to train AI

The legal challenge against Anthropic

Anthropic is currently facing a copyright infringement lawsuit in the U.S. District Court for the Northern District of California. The legal action was filed on the 23rd by several organizations, including the Authentic Brands Group. This group is responsible for managing 274 of the hit songs belonging to the legendary singer Elvis Presley.

The plaintiffs argue that the AI company built its entire business model on a massive scale of copyright infringement. They claim that Anthropic used these works without obtaining any permission from the rightful copyright holders. Furthermore, the complaint alleges that Anthropic kept all the profits generated from these unauthorized uses for itself.

According to the reports by 매일경제, the lawsuit specifically targets the way the company trained its next-generation Claude models. The plaintiffs believe that by using these songs, Anthropic has created a system that bypasses the traditional protections of intellectual property. They are seeking a permanent prohibition on these activities and the destruction of all unauthorized copies.

How the AI models were trained

The core of the complaint involves how Anthropic allegedly stored Presley's music within its own private library. The plaintiffs claim these songs were used as primary data to train the Claude AI models. This training process allows the AI to learn the patterns, styles, and structures of the original music works.

Because of this training, the plaintiffs argue that Claude users can now generate new songs that are derived from the original hits. These new creations can mimic the arrangements and choruses of Presley's music without the creators receiving any separate payments. The lawsuit highlights that the AI can produce content that feels similar to the originals when prompted by users.

Specific examples of the music covered in the lawsuit include famous hits like "All Shook Up" and "Can't Help Falling in Love." The plaintiffs claim that when users ask the AI to create music for modern listeners, the results often mirror these specific works. This ability to replicate a specific artist's style is a central point of contention in the legal filing.

Financial stakes and previous disputes

The organizations filing the lawsuit are seeking significant financial compensation for the use of the 274 works. They are requesting damages of up to $150,000 for each individual work involved in the complaint. This brings the total amount of damages sought to $41.1 million, which is about 54.5 billion won.

This is not the first time Anthropic has faced legal hurdles regarding the data it uses for its AI models. Last year, the company reached a settlement of $1.5 billion with a group of authors who accused it of using pirated books. That settlement addressed the unauthorized use of literary works to train the company's technology.

The music industry continues to see a wave of similar litigation against AI developers. In January of this year, Universal Music and Concord Music Group sued Anthropic over the alleged infringement of copyrighted song lyrics. Additionally, Sony Music Publishing and Warner Chappell Music are pursuing their own legal actions regarding these issues.

Anthropic's response and next steps

Anthropic has officially responded to the allegations made by the Authentic Brands Group and the other plaintiffs. The company stated that it disagrees with the claims brought against it in the lawsuit. They have indicated that they will actively contest the legal proceedings as they move forward in the California court.

The outcome of this case will likely depend on how the court views the distinction between 'training' on data and 'infringing' on a copyright. If the court finds that the AI's output is too similar to the original works, it could lead to stricter regulations on how companies source their training data. It could also force a change in how royalty payments are distributed to music owners.

For now, the legal battle remains in the early stages as the court reviews the evidence provided by the plaintiffs. The case will be closely watched by the music industry and the tech sector alike. It serves as a major test for the boundaries of fair use in the era of generative artificial intelligence.