Anthropic copyright lawsuit: Sony Music Publishing and Warner Chappell sue over alleged mass piracy
Anthropic copyright lawsuit: Sony Music Publishing, Warner Chappell and other publishers accuse the AI lab of mass torrenting and scraping to train Claude.
Anthropic copyright lawsuit filed late Friday in federal court accuses the AI lab and its co‑founders of a “brazen campaign” of illegally downloading and scraping copyrighted works to train its Claude models. The complaint, lodged in the U.S. District Court for the Northern District of California, names Sony Music Publishing, Warner Chappell and numerous other music publishers as plaintiffs seeking multibillion‑dollar damages. Plaintiffs contend the conduct amounts to systematic theft of works including lyrics, sheet music and books used in model training.
Publishers file multibillion‑dollar complaint
The complaint alleges that Anthropic, along with co‑founders Dario Amodei and Benjamin Mann, engaged in large‑scale acquisition of protected content without authorization. Plaintiffs say the company used torrenting and automated scraping to amass millions of copies of copyrighted works, then incorporated those works into training datasets for Claude.
The filing frames the conduct as coordinated and repeated, and asks the court to hold Anthropic accountable through damages and injunctive relief. The publishers describe the case as one of the most significant alleged thefts of intellectual property in the AI era.
Allegations of illegal torrenting and mass scraping
At the core of the complaint are accusations that Anthropic obtained materials through piracy methods, including torrent networks and bulk downloading tools. Plaintiffs assert that these techniques allowed the company to gather millions of works quickly and covertly, including books that contain musical lyrics and notated sheet music.
The suit emphasizes that the alleged acquisition method, not merely the subsequent use in training, is unlawful. According to the complaint, even if training on copyrighted content can be lawful in some contexts, obtaining that content by illicit means remains actionable.
Connection to prior litigation and counsel overlap
The current action builds on previous litigation against Anthropic and involves some of the same legal teams that pursued earlier claims. Those counsel previously represented groups including Concord Music Group and Universal Music Group in a separate case filed earlier in the year.
The complaint also references the Bartz litigation, in which a judge concluded that while training on copyrighted works might be permissible, acquiring those works via piracy was not. That case resulted in a substantial monetary judgment against Anthropic, a fact plaintiffs highlight to underscore the alleged pattern of misconduct.
Scope of works targeted and damages sought
Publishers say the scope of the alleged downloads extends beyond isolated files and includes vast libraries of books, songs, lyrics and sheet music. The complaint characterizes the harm as not only economic but cultural, asserting that the unauthorized use of musical works undermines licensing markets and creators’ control over their material.
While the filing seeks multibillion‑dollar damages, it does not specify a single aggregate figure in the public summary. Plaintiffs are asking the court to assess losses, statutory damages, and other relief tied to the scale of the alleged infringement.
Potential precedents for AI training practices
Legal experts and industry participants will watch the case for guidance on how courts treat the intersection of data acquisition methods and AI training. The complaint draws a line between permissible model training and impermissible acquisition tactics, a distinction that could influence licensing negotiations and compliance practices across the sector.
Companies that develop large language and generative models may face heightened scrutiny over their data sourcing, and this litigation could sharpen obligations for audits, provenance tracking, and licensing agreements. The outcome may also affect how publishers and rights holders structure collective responses to unauthorized use.
Anthropic’s silence and litigation next steps
Anthropic was not available for comment at the time the complaint was filed, and the company had not responded publicly to the specific allegations contained in the suit. The filing initiates a sequence of pretrial procedures, including responses, discovery requests, and potential motions that could significantly expand the factual record.
If discovery proceeds, plaintiffs are likely to seek detailed logs, acquisition records and internal communications about dataset construction and ingestion practices. Those materials could be pivotal in resolving disputed facts about whether the works were obtained lawfully.
The case adds another high‑profile dispute to a growing docket of copyright lawsuits involving AI developers, and it reinforces the legal and commercial tensions that accompany rapid advances in model capabilities. As this litigation unfolds, stakeholders across music, publishing and technology will monitor outcomes that could shape licensing norms and enforcement strategies going forward.
