Elvis Catalog Owner Sues Anthropic Over 274 Songs, Seeks $41M

Authentic Brands Group is demanding $150,000 per infringed track — roughly $41 million, or about £31 million — from AI developer Anthropic for allegedly training its Claude models on 274 Elvis Presley songs without permission. The lawsuit, first reported by Billboard, was filed on Friday in what marks one of the first major legal challenges brought directly by an artist's estate against an AI developer.

Authentic Brands Group holds the rights to both Presley's music and his likeness. The suit accuses Anthropic of infringing on 274 Presley songs, including Can't Help Falling in Love, All Shook Up and A Little Less Conversation. Multiply the $150,000 statutory-claim figure by 274 tracks and the demand lands at approximately $41 million — a sum large enough to signal intent, though a fraction of what full catalog litigation could reach.

What does the lawsuit actually allege?

Lead attorney Richard Busch did not hedge. "The Elvis plaintiffs bring this action to enforce these exclusive rights against the defendant that has built its entire business model based on copyright infringement at a scale never seen before," Busch wrote in the filing.

The complaint claims that "from its inception, Anthropic's AI model, 'Claude,' has created infringing output of copyrighted material of virtually every song from every catalog available on the internet." That is a sweeping assertion. Notably, the suit's own language refers to outputs rather than proving the training data itself — a distinction courts will have to parse.

The filing also targets Anthropic's internal safeguards. "Even Anthropic's institution of guardrails (albeit ineffectively) to purportedly limit outputs which infringe upon music copyrights, after being caught, along with its knowledge that its guardrails were ineffective to stop creating outputs which infringed music copyrights, demonstrates that Anthropic knowingly engaged and continues to willfully engage in copyright infringement of others' musical works," the lawsuit states. The word "knowingly" matters here: willfulness can raise statutory damages under US copyright law. The suit does not detail when Anthropic installed those guardrails or how their failure was measured.

Why is this case different from prior AI music suits?

AI firms already face copyright infringement claims from the Recording Industry Association of America (RIAA) and major labels including Sony, Universal and Warner Records. Those cases pit corporations against corporations. The Presley suit reframes the fight as an artist's estate confronting an AI developer directly, and its outcome could set a broader precedent for how AI companies use existing music catalogues.

The mechanics under dispute are well established. AI music platforms analyze large amounts of recorded music to learn the sounds, structures and expressions of songs, then let users generate music from text or speech prompts. Whether that analytical ingestion counts as infringement, fair use or something courts have not yet named remains unresolved.

Is there wider pressure on AI-generated music?

Musicians have repeatedly said their music and voices are being used without permission by AI models, and have voiced fears that AI-generated "slop" music could replace human-made songs. Those concerns are not hypothetical. Some AI bots have begun impersonating real, living artists on streaming platforms.

The industry is responding at scale. Last week, Spotify announced it had removed about 50 million songs it considered AI slop from its user recommendations over the past year — 50 million tracks in twelve months, a volume that illustrates how much machine-generated content the platforms now filter.

What comes next?

Anthropic has not publicly responded to the specific allegations, and no trial date or settlement discussion appears in the filing. The $41 million demand is the opening number, not the likely final one. If a court accepts the estate's framing — that training on copyrighted songs and then producing infringing output constitutes willful infringement — every AI developer with music in its training data inherits a new legal template.