AI Training with Taylor Swift’s Songs... This Company Sued Again

Sony and Warner Music Sue Anthropic
"Used Thousands of Songs Without Authorization for AI Training"

Anthropic, an artificial intelligence (AI) company, has become embroiled in a massive copyright infringement lawsuit brought by the popular music industry, following a similar case from the publishing sector. The allegation is that Anthropic used thousands of copyrighted songs without authorization to train its AI models.

On August 30, Yonhap News Agency reported that “the day before, Sony Music and Warner Chappell Music, a subsidiary of Warner Music, filed a lawsuit against Anthropic in the U.S. District Court for the Northern District of California, claiming that Anthropic had illegally used tens of thousands of song lyrics and sheet music to train its AI.” According to the report, not only the Anthropic corporation itself, but also Anthropic CEO Dario Amodei and co-founder and head of technology Benjamin Mann, were named as defendants.

Anthropic Artificial Intelligence (AI) service 'Claude' logo. Photo by Reuters Yonhap News

In their complaint, Sony Music and Warner Chappell Music asserted that Anthropic obtained pop music scores and lyrics through pirated books downloaded from illegal websites, scanned copies of physical books, and data scraped in large quantities from internet sites in violation of their terms of service, and then used these materials to train its AI models. They stated that Anthropic used more than several thousand songs without authorization for training purposes, citing Mariah Carey’s iconic Christmas carol ‘All I Want for Christmas Is You,’ Taylor Swift’s ‘Paper Rings,’ Leonard Cohen’s ‘Hallelujah,’ and Bon Jovi’s ‘Livin’ on a Prayer’ as key examples.

Additionally, the two companies alleged that Anthropic systematically altered and deleted copyright management information (CMI) in order to conceal the copyright infringement. As a result, when users of the Claude AI model inquired about certain song lyrics, the model output verbatim reproductions of the lyrics without any information about the original creators.

The plaintiffs characterized Anthropic’s actions as “the largest and most blatant misappropriation of intellectual property in history.” They demanded damages of up to 150,000 dollars (about 200 million won) per unlawfully copied song. They also sought the permanent destruction of Anthropic’s entire database of illegally reproduced lyrics and all related training files.

This is not the first time Anthropic has faced litigation over allegations of using pirated materials for AI training. Previously, in September of last year, Anthropic agreed to pay 1.5 billion dollars (about 2 trillion won) to authors in a lawsuit regarding the use of pirated books in its AI models. That case was considered the largest copyright-related settlement in U.S. history.

Regarding the current lawsuit, Anthropic told the IT news outlet TechCrunch, “We do not agree with their claims, and we plan to vigorously defend ourselves in court.”

This content was produced with the assistance of AI translation services.