case-law

Concord, UMG, and BMG v. Anthropic: The Music Piracy Suits

By Nora Whitfield · August 8, 2026

Anthropic already paid $1.5 billion to settle a book-piracy case. Now the music industry wants its turn, using a lot of the same evidence, against the same company, and asking for even more money.

The first suit: Concord, UMG, and ABKCO

Universal Music Publishing Group, Concord Music Group, and ABKCO first sued Anthropic in October 2023, alleging the company copied and used their copyrighted musical compositions, primarily lyrics, to train Claude without permission or a license. The original complaint covered 499 copyrighted works. This case has been running in parallel with the book-publishing litigation against Anthropic for years, and by March 2026, the music publishers moved for partial summary judgment, asking the court to rule outright that Anthropic infringed their copyrights and to reject Anthropic's fair-use defense entirely. Their filing described the evidence as "overwhelming," an aggressive posture that suggests they believe the same fact pattern that sank Anthropic in the book-piracy case applies just as clearly here.

Amicus support has been substantial. The RIAA, the National Music Publishers Association, and other industry groups filed briefs backing the music publishers, characterizing Anthropic's unlicensed copying as "inexcusable," language clearly meant to echo the tone that helped drive the $1.5 billion book settlement toward resolution rather than trial.

The second suit: a much bigger number

In January 2026, the same core group of plaintiffs, UMG, Concord, and ABKCO, filed a second, far larger lawsuit against Anthropic, this one covering more than 20,000 songs and seeking over $3 billion in statutory damages. This isn't a modest expansion of the first case. It's an order of magnitude larger in scope, reportedly naming artists and catalogs spanning performers like Katy Perry, Beyoncé, and The Rolling Stones, among many others, and it uses the maximum statutory damages framework rather than the smaller, actual-damages approach.

The strategic logic is straightforward once you see the pattern from the book-publishing side: the original Concord suit tested the legal theory on a manageable set of works. Once that theory looked strong, plaintiffs' counsel scaled it up dramatically, betting that the same underlying facts, unlicensed copying from music catalogs, sourced in a manner similar to the piracy findings that sank Anthropic's book defense, would produce another enormous settlement or judgment.

BMG joins in March 2026

A third, related lawsuit followed in March 2026, filed by music publisher BMG, covering 493 compositions and alleging the same core conduct: Anthropic training its models on copyrighted song lyrics and sheet music without a license. BMG's suit is smaller in scope than the January UMG/Concord filing but follows the identical legal theory, and its timing, right on the heels of the larger suit, suggests coordination across the music publishing industry in how it's pursuing Anthropic specifically.

Why this echoes Bartz v. Anthropic so directly

The connection to the book-piracy litigation isn't incidental. The music publishers' filings draw explicitly on evidence and legal reasoning developed in the Bartz case, the same distinction Judge Alsup drew there between legitimately acquired training material and material sourced through piracy or unauthorized bulk copying. If Anthropic's practices around acquiring song lyrics and sheet music resemble its practices around acquiring books, meaning bulk collection from sources without proper licensing, the same legal exposure that produced a $1.5 billion settlement in the book context could apply here, at a larger dollar figure given the higher per-work statutory damages available and the sheer number of works claimed.

This is worth understanding as a broader pattern, not just an Anthropic-specific story. Once one industry establishes a working legal theory and gets a nine-figure or ten-figure result from it, other industries with similar fact patterns move quickly to apply the same theory to their own claims. Publishing led with Bartz. Music is now following with Concord, UMG, and BMG. It's a reasonable bet that other content industries with plausible piracy-sourced training claims against major AI companies are watching this sequence closely.

What this means for AI authors

If you're a musician, songwriter, or composer, this litigation is the clearest sign yet that the music industry is treating AI training on unlicensed lyrics and compositions as an actively litigated, well-funded legal fight, not a theoretical concern. If your work is part of a catalog represented by a major publisher, there's a real chance it's already implicated in litigation like this without any action required on your part, which is both reassuring (organizations with resources are pursuing this on your behalf) and a reason to keep your own records of what you've published and when, in case you ever need to participate directly in a claim.

For anyone using AI tools in a music-related creative process, whether generating melodies, getting lyrical suggestions, or using AI-assisted production tools, this case is a reminder that the underlying training data question is genuinely unresolved and actively expensive when it goes wrong. It doesn't mean avoid these tools. It means understand that the legal ground under them is shifting, and that your own authorship record, showing which words, melodies, and arrangements are genuinely yours regardless of what any given AI tool's training data situation turns out to be, remains valuable no matter how the industry-level litigation eventually resolves.

Copyrightable is not a law firm and doesn't provide legal advice. All three of these actions remain active and unresolved. Confirm current status at official court dockets before relying on this summary.