A group of major music publishers led by Concord Music Group and Universal Music Group has filed a lawsuit against AI company Anthropic, accusing it of illegally downloading more than 20,000 copyrighted works including sheet music, lyrics, and musical compositions to train its AI models.
In a statement released Wednesday, the publishers said damages could exceed $3 billion, making the case one of the largest non-class action copyright lawsuits in U.S. history. The lawsuit alleges that Anthropic did not merely use copyrighted material for training, but obtained the works through unauthorised downloading, a distinction that courts have increasingly treated as legally significant.
The case builds on the earlier Bartz v. Anthropic lawsuit, in which fiction and nonfiction authors accused the company of using their books to train products like Claude. In that case, U.S. District Judge William Alsup ruled that training AI models on copyrighted content could qualify as fair use, but drew a clear legal line around how the content is acquired, stating that piracy is not protected under copyright law.
Emmanuel Umahi
That ruling resulted in a settlement worth roughly $1.5 billion, with affected writers reportedly receiving about $3,000 per work for approximately 500,000 copyrighted titles. While the payout was substantial, critics argued it did little to reshape industry behaviour, particularly for well-funded AI companies like Anthropic, which is currently valued at about $183 billion.
According to the publishers, they originally sued Anthropic over the use of roughly 500 copyrighted works, but through discovery in the Bartz case, uncovered evidence that the company had allegedly downloaded tens of thousands more. After a court rejected their attempt to amend the original lawsuit last October, citing insufficient early investigation into piracy claims, the publishers filed this separate case. The new lawsuit also names Anthropic CEO Dario Amodei and co-founder Benjamin Mann as defendants.
“While Anthropic misleadingly claims to be an AI ‘safety and research’ company, its record of illegal torrenting of copyrighted works makes clear that its multibillion-dollar business empire has in fact been built on piracy,” the lawsuit states.
Anthropic hasn’t released a statement to this.
Emmanuel Umahi
Why This Lawsuit Matters
This case could become a defining moment for how copyright law is applied to generative AI, particularly in the music industry. Unlike books or images, music rights are typically fragmented across publishers, composers, lyricists, and labels, making licensing more complex and potentially more expensive for AI companies seeking training data.
If the publishers succeed, the ruling could force AI companies to fundamentally change how they source training data, pushing them toward licensed datasets and paid partnerships instead of scraping or downloading content at scale. That could significantly increase costs for building large language and music-generation models, potentially slowing development or reshaping which companies can afford to compete.
Emmanuel Umahi
The lawsuit could also strengthen the negotiating power of music publishers, songwriters, and composers in future AI licensing deals, similar to how major record labels have recently secured agreements with companies like Suno and Udio to compensate artists for the use of their catalogs in AI training.
More broadly, the case underscores a growing shift in how courts view AI copyright disputes. While judges may be open to fair use arguments around training models on copyrighted material, they appear increasingly unwilling to tolerate unauthorized acquisition methods. That distinction could become a legal standard across future AI litigation.
For the music industry, the outcome could determine whether generative AI becomes another revenue stream or another large-scale extraction of creative labor without compensation. For AI companies, the ruling could define the legal boundaries of model training for years to come.
Emmanuel Umahi
