Blab Music
Sony Music Publishing and Warner Chappell File Lawsuit Against Anthropic Over Alleged Copyright Infringement
On August 28 2026, Sony Music Publishing and Warner Chappell Music filed a federal suit in the Northern District of California, alleging that AI startup Anthropic used pirated song lyrics and musical compositions to train its Claude large‑language‑model (LLM) system. According to the complaint, Anthropic illegally downloaded hundreds of copyrighted works—including tracks by The Beatles, Taylor Swift and Michael Jackson—from torrent sites and fed them to Claude, enabling the model to reproduce lyrics word for word and to generate new lyrics that rival the originals.
Sony and Warner seek statutory damages of up to $150,000 per infringed work and ask the court to prohibit Anthropic from using the plaintiffs’ copyrighted material. The complaint claims that Anthropic’s use of the works allowed Claude to produce verbatim reproductions of copyrighted lyrics when prompted and to churn out vast quantities of new lyrics that could act as harmful substitutes for the publishers’ legitimate catalogues.
Founded in 2021 and based in San Francisco, Anthropic has built its flagship Claude series on a training corpus that plaintiffs say contains unlicensed copyrighted text. The company previously settled a 2025 class‑action lawsuit from authors and publishers for $1.5 billion, resolving allegations that it had used copyrighted books without permission. This latest suit joins a growing wave of copyright‑infringement cases aimed at AI developers.
Universal Music Group (UMG) filed suit against Anthropic in 2023, claiming the company used UMG’s song lyrics for training. That case is still pending. Other publishers—EMI, Concord, ABKCO, BMG and Round Hill—have also sued Anthropic, alleging similar unauthorized use of their catalogs.
The complaint points out that Claude models can generate new lyrical content that is vastly similar to the copyrighted works used during training. Plaintiffs contend that this practice erodes the market for legitimate music publishing and infringes the exclusive rights of copyright holders.
Industry observers view the lawsuit as a litmus test for how copyright law applies to AI training data. Although the complaint does not assert that Anthropic’s use of the works qualifies as fair use, it cites the company’s reliance on pirated material as clear evidence of infringement.
Neither Anthropic nor the plaintiffs’ representatives have issued statements. The case remains pending in federal court, with no hearing dates yet set.
The lawsuit underscores the escalating friction between AI developers and copyright owners over the use of protected text in training large language models. As AI systems expand into commercial arenas, the legal framework surrounding training data is still evolving, and the case’s outcome could shape how AI firms source and license copyrighted material going forward.
Sony and Warner’s filing is the newest chapter in a series of legal actions that could redefine the relationship between the music publishing industry and AI technology providers. Stakeholders will monitor court rulings that clarify how far copyrighted works can be used in AI training without explicit permission.
Sony and Warner seek statutory damages of up to $150,000 per infringed work and ask the court to prohibit Anthropic from using the plaintiffs’ copyrighted material. The complaint claims that Anthropic’s use of the works allowed Claude to produce verbatim reproductions of copyrighted lyrics when prompted and to churn out vast quantities of new lyrics that could act as harmful substitutes for the publishers’ legitimate catalogues.
Founded in 2021 and based in San Francisco, Anthropic has built its flagship Claude series on a training corpus that plaintiffs say contains unlicensed copyrighted text. The company previously settled a 2025 class‑action lawsuit from authors and publishers for $1.5 billion, resolving allegations that it had used copyrighted books without permission. This latest suit joins a growing wave of copyright‑infringement cases aimed at AI developers.
Universal Music Group (UMG) filed suit against Anthropic in 2023, claiming the company used UMG’s song lyrics for training. That case is still pending. Other publishers—EMI, Concord, ABKCO, BMG and Round Hill—have also sued Anthropic, alleging similar unauthorized use of their catalogs.
The complaint points out that Claude models can generate new lyrical content that is vastly similar to the copyrighted works used during training. Plaintiffs contend that this practice erodes the market for legitimate music publishing and infringes the exclusive rights of copyright holders.
Industry observers view the lawsuit as a litmus test for how copyright law applies to AI training data. Although the complaint does not assert that Anthropic’s use of the works qualifies as fair use, it cites the company’s reliance on pirated material as clear evidence of infringement.
Neither Anthropic nor the plaintiffs’ representatives have issued statements. The case remains pending in federal court, with no hearing dates yet set.
The lawsuit underscores the escalating friction between AI developers and copyright owners over the use of protected text in training large language models. As AI systems expand into commercial arenas, the legal framework surrounding training data is still evolving, and the case’s outcome could shape how AI firms source and license copyrighted material going forward.
Sony and Warner’s filing is the newest chapter in a series of legal actions that could redefine the relationship between the music publishing industry and AI technology providers. Stakeholders will monitor court rulings that clarify how far copyrighted works can be used in AI training without explicit permission.
