Universal Music Group May Have Admitted “Dilution” Theory Is Base


In a surprising development on March 6, Universal Music Group’s (UMG) Executive Vice President and Chief Digital Officer Michael Nash may have admitted that foundational legal claims in UMG’s multi-plaintiff copyright lawsuit against generative AI developer Suno have no factual basis.

UMG and plaintiffs have advanced a novel legal theory in several cases known as “dilution.” Dilution theory holds that generative AI users can produce so many new works that the value of existing works will be diluted, and that such general harm should be considered under copyright law. In its June 2024 complaint against Suno, UMG stated that its catalogue would be “overrun” by new works created by Suno users, which “compete for plays against real, copyrighted sound recordings” and “substantially dilute the royalty pools that are paid out to artists.”

But speaking to investors on UMG’s 2025 Q4 earnings call, Nash contradicted UMG’s legal filing, saying “we are seeing no indication that AI royalty dilution is a material issue for UMG from a revenue perspective.” Nash assured investors that dilution would not financially harm UMG’s revenues, now or in the future, saying:

“First, misunderstandings have resulted from anecdotal press reports that AI-generated content has somehow overtaken the charts. Nothing could be further from the truth … In the aggregate, the most prominent AI content barely registers even in the leading market for this English language repertoire.” 

“Some commentators say, ‘That’s right now. What about the future?’ We don’t have to theorize about the future of AI saturation as it’s become a marketplace reality with 60,000 AI tracks being uploaded a day at present … So despite the huge volume of AI uploads, the aggregate organic consumption of AI content by actual consumers is less than 0.5% based on the best available data.”

 Nash later sharpened the point on March 10 at a conference in London, noting: “we feel like the case for AI dilution risk has been massively over-extrapolated from a few anecdotes.

Dilution Theory

UMG v. Suno pits one of the world’s largest music companies against a generative AI startup. UMG has argued that dilution’s harm to whole categories of works should be considered under copyright law. Suno contends that using large numbers of copyrighted works to train and produce their AI models without permission does not harm any specific work. Its training is thus legal fair use under §107 of the Copyright Act.

Dilution theory is both unprecedented and far reaching, as Professor Edward Lee of Santa Clara Law explains:

Copyright dilution is sweeping. It flouts copyright law’s fundamental limitation on the scope of copyright infringement to substantially similar copies that copy protected elements from the plaintiff’s work. It proposes to treat every output of AI generators as dilution of every copyrighted work used to train the AI.”

 

In Bartz v. Anthropic, that sweeping impact was why Judge William Alsup rejected dilution as proof of legitimate harm. He wrote in his order granting Anthropic’s summary judgment motion that even if generative AI models produced an “explosion of competing works,” that is “not the kind of competitive or creative displacement that concerns the Copyright Act. The Act seeks to advance original works of authorship, not to protect authors against competition.”

In Kadrey v. Meta, Judge Vincent Chhabria was taken with dilution theory, but he points to no other case in which dilution played a role. Unlike Judge Alsup’s ruling on dilution citing previous cases, Chhabria’s writing is what lawyers call “dicta” – it carries no precedential authority. Dilution theory is also briefly discussed by the U.S. Copyright Office in the unofficial printing of Part 3 of its “Copyright & AI” report on generative AI training, which found no support for dilution in copyright statutes or case law, noting that dilution theory takes U.S. law into “uncharted territory.”

UMG v. Suno

UMG’s case is complicated by Alsup and Chhabria’s rulings that generative AI models are “quintessentially transformative” under the first heavily weighted factor of the statutory fair use test. In his ruling, Alsup drew on Sega v Accolade, in which the 9th Circuit held “an attempt to monopolize the market by making it impossible for others to compete runs counter to the statutory purpose of promoting creative expression and cannot constitute a strong equitable basis for resisting the invocation of the fair use doctrine.”

The ruling in Sega, and many other cases regarding new technologies, derives directly from the copyright clause of the Constitution, Art. I, §1, Clause 8, which is the only clause in the document that expressly states its purpose: “to promote the Progress of Science and useful Arts.”

As Professor Lee again explains, by thwarting that goal in multiple ways, dilution theory may be unconstitutional:

“[Dilution theory] stretches copyright beyond the permissible bounds in the Copyright Clause, which limits the grant of copyright, or exclusive rights, to authors in only ‘their respective writings.’ Authors cannot own entire genres, and the non-infringing works of others are not theirs. Moreover, because copyright dilution alters the traditional contours of copyright, both fair use and the idea-expression dichotomy, it must be subject to strict scrutiny under the First Amendment. Copyright dilution fails. As the Supreme Court recognized, ‘Mere speculation of harm does not constitute a compelling state interest.”

 

With a theory as unprecedented and far reaching as dilution, Alsup and Chhabria’s recent rulings will almost certainly present a challenge to UMG.

Returning to Nash’s comments on UMG’s earnings call and in London, the remarks have already appeared to make an impact in the case. At a March 12 hearing, Suno counsel read Nash’s comments into the record. UMG lawyers argued the remarks were irrelevant and asked Magistrate Judge Paul Levenson to deny discovery by Suno into Nash’s comments and the data behind them.

Judge Levenson denied the request.



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *