[E-Daily Reporter YUN GI BACK ] The copyright dispute surrounding the artificial intelligence (AI) company Anthropic is spreading beyond the publishing industry into the pop music industry. As major global music companies have filed lawsuits alleging that Anthropic obtained tens of thousands of musical works without permission and used them for AI training, Anthropic’s method of collecting training data is once again under judicial scrutiny.

According to reports by U.S. media outlets on the 29th (local time), music publishing companies—including Sony Music Publishing and Warner/Chapel Music—filed a copyright infringement lawsuit against Anthropic the previous day in the U.S. District Court for the Northern District of California. Not only was Anthropic named as a defendant, but its co-founders, CEO Dario Amodei and Benjamin Mando, were also named as defendants.
The plaintiffs’ issue is not merely the fact that AI was trained on copyrighted works. They allege that Anthropic obtained these works through illegal downloads and data collection from websites.
According to the complaint, the plaintiffs allege that Anthropic downloaded millions of illegally copied books via torrent, some of which contained lyrics and sheet music from commercial songs. They also point out that Anthropic collected large amounts of data from online sites that legally provide lyrics and used it to train its AI model, “Claude.” According to Music Business Worldwide (MBW), a media outlet specializing in the music industry, the lawsuit alleges infringement involving tens of thousands of musical works.
These included well-known hit songs. Mariah Carey’s “All I Want for Christmas Is You” and Taylor Swift’s “Paper Rings” were cited as representative examples.
Allegations were also made that copyright management information (CMI) was removed or altered. The plaintiffs claimed that Anthropic repeatedly copied the works it had obtained and used them to train its AI, causing Claude to generate copyrighted lyrics and other content.
Music publishing companies characterized Anthropic’s actions as “one of the largest and most brazen instances of intellectual property theft in history.” If intentional copyright infringement is found, they are seeking statutory damages of up to $150,000 (approximately 207.07 million won) per work. They also sought up to $25,000 (approximately 34.51 million won) per instance for the deletion or alteration of copyright management information. Given that the scope of the infringement spans tens of thousands of cases, there is a possibility that the total damages could reach billions of dollars, depending on the court’s ruling.
The plaintiffs are not merely seeking monetary damages. They have also demanded that Anthropic destroy all copies of the infringed works in its possession and disclose the details of the data used to train Claude.
What makes this lawsuit even more significant is that it comes on the heels of a major dispute Anthropic has already faced regarding published works. In the “Bartz v. Anthropic” lawsuit filed by authors, the issue of acquiring books through pirate sites came to light, and Anthropic agreed to a settlement worth $1.5 billion (approximately 2.0707 trillion won). In this latest lawsuit, the large-scale acquisition of pirated books—which was revealed at that time—has once again been raised as an issue.
This dispute is particularly noteworthy because it expands the copyright debate surrounding generative AI from the question of “whether training constitutes fair use” to the question of “how the data used for training was obtained in the first place.” Issues such as whether copyrighted content was obtained through illegal channels, whether copyright information was removed, and whether the training results reproduce the original works verbatim have all emerged as simultaneous points of contention.
As the scope of these large-scale lawsuits has expanded from published works to include song lyrics and sheet music, the legal burden on AI companies regarding their data acquisition practices is expected to increase significantly. In the music industry, in particular, Universal Music Publishing Group (UMPG) and Concord are already pursuing lawsuits against Anthropic. With Sony Music Publishing and Warner/Chappell Music joining the fray, the publishing divisions of all three of the world’s largest music companies are now embroiled in legal disputes with Anthropic.
Anthropic has denied these allegations. An Anthropic spokesperson told TechCrunch, “We disagree with the music publishers’ claims and will vigorously defend ourselves in court.”