Anthropic Settles Copyright Dispute with Music Publishers Over AI Lyric Use

Anthropic has made a deal to settle parts of a copyright infringement lawsuit ​​brought against the maker of the Claude AI model for allegedly distributing protected song lyrics.

Share:

Alade-Ọrọ̀ Crow

An image showing a slightly off-kilter grid of happy-looking robot faces with speech bubbles containing music notes.

Anthropic has recently reached a settlement regarding a copyright infringement lawsuit concerning the Claude AI model, which was accused of distributing protected song lyrics. by US District Judge Eumi Lee, mandating that Anthropic implement existing guardrails in future AI model training and establish a process for music publishers to intervene in instances of suspected copyright infringement.

In October 2023, several music publishers, such as Universal Music Group and ABKCO, filed a copyright infringement lawsuit against Anthropic, alleging that the company trained its AI system using lyrics from over 500 protected songs. According to the complaint, when Claude was prompted for lyrics to popular songs, such as Beyoncé’s “Halo,” it provided responses containing significant portions of those lyrics.

While music publishers acknowledged that platforms like Genius distribute lyrics online, they pointed out that those sites pay licensing fees for the use of protected works. The lawsuit contends that Anthropic “intentionally removed or altered copyright management information” from the songs when collecting data from these sources to train its AI models.

As part of the agreement signed by Anthropic, the company has committed to maintaining the guardrails currently in place to prevent its AI models from infringing on copyrighted content. These measures will also be applied to any future AI systems developed. Anthropic and music publishers will collaborate in “good faith” to address any situations where these guardrails may fall short, with the court prepared to resolve any disputes.

In a statement to The Hollywood Reporter, Anthropic asserted, “Claude isn’t designed to be used for copyright infringement, and we have numerous processes in place to prevent such infringement. Our decision to enter into this stipulation aligns with those priorities. We look forward to demonstrating that, under existing copyright law, training generative AI models with potentially copyrighted material is a classic example of fair use.”

The music publishers involved in the original complaint are seeking a preliminary injunction to prevent Anthropic from training future models using their protected song lyrics, with a court ruling expected in the coming months.

Latest in