BMG Sues Anthropic Over Alleged Use of Copyrighted Lyrics in AI Training
Music company BMG has filed a lawsuit against Anthropic, alleging that the artificial intelligence firm used copyrighted song lyrics without authorization to train its chatbot, Claude.
The complaint, filed in federal court in California, claims Anthropic incorporated lyrics from artists including Justin Bieber, Bruno Mars, Ariana Grande, and The Rolling Stones as part of its training data. According to BMG, the material was collected through automated scraping of online sources, including websites and unauthorized repositories.
In the filing, BMG alleges both direct copyright infringement and secondary liability. The company argues that Anthropic not only used protected material without permission but also enabled users of its Claude models to reproduce those works.
The complaint cites examples in which the chatbot allegedly generated full or substantial portions of well-known songs when prompted. Among the tracks referenced are “Uptown Funk,” “You Can’t Always Get What You Want,” and “What a Wonderful World.”
BMG states that it never granted Anthropic permission to use its catalog and that the company failed to respond to a cease-and-desist letter sent in December 2025. According to the lawsuit, no licensing discussions took place despite the request.
A central issue in the case is how Anthropic sourced and used training data. BMG is seeking a court order requiring the company to disclose details about its datasets, including whether and how copyrighted material was incorporated.
The lawsuit also requests information about the chatbot’s design and capabilities, particularly in relation to how it generates text that may resemble or reproduce existing works. BMG is seeking statutory damages of up to $150,000 per infringed work, along with injunctive relief aimed at preventing further use of its material.
The case is one of several recent legal actions targeting artificial intelligence firms over the use of copyrighted content. As generative AI systems have expanded, questions about how training data is collected and whether it complies with existing copyright law have become a central issue across multiple industries.
Anthropic, founded in 2021 by former members of OpenAI, has positioned itself as an AI company focused on safety and ethics. BMG’s complaint challenges that positioning, arguing that the company’s practices contradict its stated principles.
Rights holders are increasingly seeking clarity on how their content is used in training models, while technology companies argue that large-scale data use is essential to building advanced systems.
Anthropic has not publicly responded to the lawsuit.
