Image Credit: Completemusicupdate.com
The ongoing legal battle between major music publishers and the AI firm Anthropic has reached a critical juncture, as the startup moves to have a high-profile copyright lawsuit dismissed. In its latest filing, Anthropic is currently using Universal Music Group’s (UMG) own words against it, arguing that the publisher’s recent statements to investors undermine its claims of financial harm. The dispute centers on whether training the AI model Claude on copyrighted song lyrics constitutes “fair use” or an illegal dilution of the music market.
At the heart of the publishers’ lawsuit is the claim that AI-generated content is diluting the revenue pools for human-made songs. However, Anthropic pointed out in its Monday filing that UMG executives told investors just last month that they are seeing “no indication that AI royalty dilution is a material issue” for their revenue. By quoting this earnings call, Anthropic argues that the publishers have failed to show any undisputed market harm, which is a key requirement for defeating a fair use defense in court.
Anthropic’s defense goes beyond financial technicalities, leaning into a more philosophical interpretation of the US Copyright Act. The company compares the training of Claude to how a human learns by reading and re-reading books and lyrics to internalize themes and styles. Anthropic maintains that its AI does not simply replicate original works but creates something entirely new—a “transformative” use that allows the model to reason, code, and explain. The company insists that the primary objective of copyright is to serve the public by encouraging such innovation, rather than solely rewarding the author.
The publishers—including UMG, Concord, and ABKCO—contend that Anthropic committed copyright infringement on a massive scale by copying more than 20,000 compositions without permission. While they argue that Claude competes directly with licensed lyric sites, Anthropic counters that the “fair use inquiry” is anchored on the problem of substitution, not general competition. No single output from Claude is a direct substitute for a specific song, and the startup claims that its internal “guardrails” are designed to prevent the chatbot from regurgitating exact lyrics on demand.
This case is currently defining the economic reality for the $2 billion-plus music rights industry. If the judge agrees with Anthropic that AI training is “highly transformative,” it could set a massive precedent for how tech companies use creative content. However, the publishers still have another legal card to play; a separate lawsuit recently targeted Anthropic for its alleged use of pirated datasets. As the court weighs these arguments, the outcome will likely dictate how effectively major labels can protect their catalogs in an era of rapid AI expansion.
About the Author
