A copyright lawsuit against AI start-up Anthropic puts a practical question at the center of the music industry’s debate about artificial intelligence: what happens when a chatbot responds to a prompt with recognizable song lyrics? Universal Music, Concord, and ABKCO allege that Anthropic used their songs without permission and that its Claude chatbot can produce copies of lyrics.
What the music companies allege
The companies filed their case in a US court in Nashville, Tennessee. They say Anthropic scraped songs without permission and used them to generate “identical or nearly identical copies of those lyrics.” They also claim the company had “never even attempted” to license their copyrighted work.
The complaint describes responses to prompts as evidence of the alleged problem. When asked for the lyrics to Gloria Gaynor’s “I Will Survive,” Claude allegedly returned text that was nearly word for word. The companies also say the chatbot can incorporate lyrics into answers to prompts that do not simply ask for a song’s words.
For example, the filing says a request for short fiction in the style of Louis Armstrong led Claude to use lyrics from “What a Wonderful World.” The allegation raises a distinction between a chatbot discussing a musician’s style and reproducing material associated with that musician. The lawsuit argues that the latter can happen even when a user’s prompt is framed as a different kind of creative task.
Training and chatbot answers are part of the dispute
The companies’ claims address both the material they say was used by Anthropic and what Claude produces in response to users. That puts the focus on two connected parts of an AI service: the content used in its development and the text that appears in a conversation.
The examples in the filing describe outputs that the companies say contain protected lyrics, whether a user asks for a song directly or requests writing in an artist’s style. Their complaint is that copyrighted work appearing online does not, by itself, make it free to use. Anthropic did not immediately respond to a request for comment.
The lawsuit is one part of a wider struggle between the music business and AI technology. The industry is also confronting AI-generated “deepfake” songs that imitate established musicians’ voices, lyrics, or sound. Earlier that year, a song mimicking Drake and The Weeknd spread online, drawing attention to how AI-generated music can evoke artists without being an ordinary release by those musicians.
Licensing offers another path
The legal claims sit alongside efforts to establish permission-based uses of music in AI. Universal announced a partnership with music platform BandLab to approach copyright use in AI “ethically,” with proper copyrights and permission. The announcement presents licensing as one way to address the use of music while engaging with AI tools.
Universal was also separately working with Google on licensing artists’ voices and melodies for AI-generated songs, according to the Financial Times, which reported on the effort in August. Those initiatives show that the dispute is not only about whether AI can be used in music. It is also about who can authorize the use of particular work and under what terms.
Universal had earlier asked Spotify and other streaming services to cut off access to its music catalogue for developers using it to train AI technology. That request and the Anthropic lawsuit both reflect concern about music being used in AI systems without the companies’ permission. The BandLab and Google efforts, meanwhile, point to licensing as a possible framework for some uses.
A familiar fight over new music technology
The conflict recalls the music industry’s copyright battles in the 2000s over technology services such as Napster, which enabled music piracy. The comparison in the source is about the industry’s recurring struggle to respond to new ways of distributing or using music. The technologies and allegations differ, but copyright remains central to the dispute.
Anthropic was founded in 2021 by researchers who left Microsoft-backed OpenAI over a disagreement about the company’s direction. Amazon and Google have invested in the start-up. The lawsuit puts the company’s chatbot into a broader debate over how AI businesses use creative work and how the music industry can protect or license it.
For the music companies, the alleged appearance of lyrics in Claude’s answers is a concrete example of why they are challenging AI use of their catalogs. For the industry more broadly, the case underscores the gap between AI’s capacity to generate music-related content and the permission arrangements rights holders say should govern that use.