Anthropic is facing another major copyright battle, this time from two of the world’s biggest music publishers.
Sony Music Publishing and Warner Chappell Music have sued Anthropic in federal court, accusing the Claude maker of illegally acquiring and using copyrighted song lyrics and sheet music to train its AI models.
The lawsuit, filed on August 28, 2026, in the U.S. District Court for the Northern District of California, alleges that Anthropic obtained copyrighted musical works through torrenting, scraping and other unauthorized methods. The complaint also names Anthropic CEO Dario Amodei and co-founder Benjamin Mann as individual defendants.
The publishers are seeking statutory damages of up to $150,000 for each work that is found to have been willfully infringed, along with additional damages related to alleged removal of copyright-management information.
That distinction matters: $150,000 is the statutory maximum being sought for a qualifying work, not a $150,000 penalty that Anthropic has already been ordered to pay for every song.
Why Are Sony and Warner Suing Anthropic?
At the center of the lawsuit is a basic but increasingly important question in the AI industry:
Can an AI company use copyrighted music to train a commercial model without permission or a license?
Sony Music Publishing and Warner Chappell say Anthropic went much further than simply using copyrighted material for AI training.
According to the complaint, Anthropic allegedly acquired thousands of copyrighted musical compositions through unauthorized sources, including pirated material, and incorporated those works into the development of Claude.
The publishers claim Claude can also reproduce copyrighted lyrics when users ask for them and can generate new lyrics that compete with human-created musical works.
Anthropic disputes the allegations.
An Anthropic spokesperson said the company disagrees with the publishers’ claims and intends to defend itself robustly in court.
How Much Money Is Sony and Warner Seeking?
The headline figure is $150,000 per work.
Under U.S. copyright law, statutory damages can reach $150,000 for a work that is found to have been willfully infringed.
Sony and Warner are seeking that maximum amount for qualifying infringements.
The complaint identifies thousands of allegedly infringed musical works, while reporting on the filing describes the potential scope as thousands if not tens of thousands of compositions.
If a court ultimately found widespread willful infringement and awarded the maximum statutory damages across a very large number of works, the financial exposure could reach into the billions.
But that is a potential damages calculation, not the amount Anthropic currently owes.
The publishers are also seeking up to $25,000 for each alleged removal of copyright-management information, adding another potential layer of damages.
Which Songs Are Mentioned in the Lawsuit?
The complaint names a number of well-known compositions.
Reported examples include:
- “Ain’t No Mountain High Enough”
- “All I Want for Christmas Is You”
- “Eye of the Tiger”
- “Hallelujah”
- “I Am the Walrus”
- “Paper Rings”
- “Cruel Summer”
- “Uptown Funk”
The artists and songwriters connected to these works include major names such as The Beatles, Taylor Swift, Mariah Carey, Marvin Gaye, Leonard Cohen and others.
The publishers’ complaint is therefore not about a small collection of obscure works.
It targets a huge catalog of commercially important musical compositions.
Did Anthropic Actually Pirate Song Lyrics?
That is one of the central allegations—and it has not been established as a final court finding in this case.
Sony and Warner allege that Anthropic obtained copyrighted material through methods including torrent downloads and scraping.
The lawsuit also points to earlier litigation involving Anthropic’s acquisition of copyrighted books.
Reuters reported that the music publishers allege Anthropic obtained lyrics and sheet music through torrent downloads and used them to train Claude.
Anthropic, however, has rejected the publishers’ characterization and says it will defend itself.
Because this is an active lawsuit, allegations in the complaint should not be presented as proven facts.
Why Are Lyrics So Important to the Case?
Song lyrics create a particularly interesting copyright problem for AI.
Unlike many factual texts, lyrics are highly creative works that receive strong copyright protection.
The publishers argue that Claude was capable of reproducing copyrighted lyrics nearly verbatim when prompted.
They also allege that the model could generate new lyrics that compete with the market for human-created music.
That creates two separate issues:
Training:
Did Anthropic have the right to acquire and use copyrighted lyrics and musical compositions to train Claude?
Output:
Can Claude reproduce protected lyrics or create outputs that unlawfully substitute for copyrighted works?
The lawsuit puts both questions directly into the spotlight.
The Lawsuit Also Targets AI-Generated Music and Lyrics
Sony and Warner’s complaint goes beyond the argument that Claude memorized existing songs.
The publishers allege that Anthropic’s models can use knowledge obtained from copyrighted music to produce large amounts of supposedly new lyrics.
Their argument is that this could create a market substitute for human songwriters.
In other words, the publishers aren’t only worried about Claude repeating a Taylor Swift lyric when someone asks for it.
They are also concerned about AI systems learning from enormous catalogs of human-created music and then generating new commercial content using what they learned.
That could become one of the most important legal questions in the AI-music industry.
This Isn’t Anthropic’s First Copyright Fight
Anthropic is already deeply involved in copyright litigation.
In 2023, Universal Music Group, Concord and ABKCO sued Anthropic over the alleged use of copyrighted song lyrics in Claude.
Additional music publishers have brought further cases in 2026, including BMG and Round Hill Music.
The Sony-Warner case therefore adds another major group of rights holders to Anthropic’s growing legal challenges.
But there is an important difference between this case and Anthropic’s earlier author litigation.
This lawsuit is specifically focused on musical compositions, lyrics and related copyright interests.
Anthropic Previously Settled a $1.5 Billion Copyright Case
The new lawsuit also comes after Anthropic agreed to a $1.5 billion settlement in litigation involving authors whose copyrighted books were allegedly used in AI training.
That earlier case became a major reference point in the debate over AI training data.
Sony and Warner argue that the previous settlement did not provide enough deterrence.
The publishers’ complaint characterizes Anthropic’s continued conduct as evidence that the earlier settlement was insufficient to change the company’s behavior.
Anthropic has pushed back against that interpretation and has argued that AI training can qualify as fair use.
Why the Earlier Copyright Ruling Matters
The legal fight isn’t simply about whether copyrighted works appeared somewhere in an AI company’s training process.
The source and method of acquisition can also matter.
In Anthropic’s earlier book litigation, the company faced allegations concerning millions of pirated books.
TechCrunch reported that the earlier court proceedings distinguished between using copyrighted works to train AI and the legality of acquiring those works through piracy.
That distinction could become important again in the music case.
If a court concludes that certain training uses are transformative or otherwise protected, the analysis may still be different if copyrighted works were unlawfully acquired in the first place.
What Does Anthropic Say?
Anthropic has rejected Sony and Warner’s allegations.
The company said it disagrees with the publishers’ claims and intends to defend itself robustly.
Anthropic has also pointed to fair-use arguments from its previous copyright litigation, maintaining that training AI models on copyrighted material can constitute transformative use.
The company has also characterized the new case as another lawsuit brought by lawyers involved in existing litigation against Anthropic.
The ultimate legal question will therefore depend on how the court evaluates both the training process and the alleged acquisition of copyrighted works.
Why This Case Could Be Huge for AI
The stakes extend far beyond Anthropic.
AI companies need enormous amounts of data to train increasingly capable models.
Music publishers, record labels, authors, news organizations and other rights holders argue that companies should not be allowed to simply absorb copyrighted creative work into commercial AI systems without appropriate permission or compensation.
AI companies, meanwhile, have argued that training models on existing information can be transformative and may be protected under copyright law.
The courts are now being asked to determine where that boundary lies.
Why Music Companies Are Taking a Hard Line
Music has already become one of the most aggressively contested areas of generative AI.
AI systems can now generate songs, imitate musical characteristics and create lyrics within seconds.
That creates obvious opportunities for new forms of creativity—but also raises concerns about the economic value of human-created music.
If AI models can learn from millions of songs and then produce commercially useful material, rights holders want clear rules about how that training is allowed to happen.
Sony and Warner’s lawsuit is therefore part of a much larger battle over who owns the data used to build AI.
What Happens Next?
The case will now move through the U.S. federal court system.
Sony and Warner are seeking a jury trial, damages and restrictions on Anthropic’s alleged use of their copyrighted works.
Anthropic is expected to challenge the allegations and defend its training practices.
The court will eventually have to address several complicated questions, including:
- How Anthropic obtained the copyrighted works
- Whether those works were used in Claude’s training
- Whether the alleged uses qualify as fair use
- Whether Claude can reproduce protected lyrics
- Whether AI-generated outputs compete with copyrighted music
- Whether alleged removal of copyright-management information occurred
- What damages, if any, should be awarded
Those questions could have consequences for virtually every major generative-AI company.
The Bigger AI Copyright Battle
The Sony-Warner lawsuit shows that the copyright fight around AI is moving into a new phase.
The early debate focused heavily on books, articles and visual artwork.
Music is now becoming an equally important battleground.
The issue is no longer simply whether AI can create something that sounds like a human song.
It is increasingly about whether the underlying AI model was built using creative works that the developers had a legal right to acquire and use.
That distinction could shape the business model of generative AI for years.
Read More:- Claude Just Proved Fermat’s Last Theorem — A Math Milestone Explained
Final Takeaway
Sony Music Publishing and Warner Chappell Music have sued Anthropic over alleged unauthorized use of copyrighted songs and lyrics in Claude’s development, seeking statutory damages of up to $150,000 per work for qualifying willful infringement.
The lawsuit names Anthropic as well as CEO Dario Amodei and co-founder Benjamin Mann, and alleges that copyrighted musical works were obtained through torrenting, scraping and other unauthorized methods.
Anthropic denies the publishers’ claims and says it will defend itself.
The most important point is that $150,000 per song is a maximum statutory damages figure being sought—not a judgment already imposed on Anthropic.
Whatever happens in court, the case could become a landmark fight over one of the biggest unanswered questions in generative AI:
How much copyrighted human creativity can AI companies legally use to build commercial models?
