Threat assessment: elevated. The owners of Elvis Presley’s music catalog have sued Anthropic. They say the company used 274 Presley songs to train its Claude models without permission. Hacker News reports that Authentic Brands Group, which owns Presley’s music and likeness, filed the suit on Friday. Billboard broke the story first.
They’re asking for $150,000 per song. Across 274 tracks, that comes to about $41 million.
📍 Situation Report
- Plaintiff: Authentic Brands Group, which licenses both the Presley catalog and his likeness.
- Defendant: Anthropic, the company behind Claude.
- The works: 274 songs, including “Can’t Help Falling in Love,” “All Shook Up” and “A Little Less Conversation.”
- The claim: Anthropic trained on these songs without a license, and Claude produced infringing output.
- The ask: $150,000 per track. That’s the top statutory amount for willful infringement under US copyright law.
- Lead counsel: Richard Busch.
⚔️ The Accusations
The complaint doesn’t hold back. Busch writes that Anthropic “has built its entire business model based on copyright infringement at a scale never seen before.”
The suit also 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.”
The sharpest argument is about guardrails. Anthropic added filters to limit copyrighted music output. The plaintiffs say those filters prove Anthropic knew there was a problem. In their words, the guardrails came “after being caught,” and Anthropic knew they were “ineffective,” which shows it “knowingly engaged and continues to willfully engage in copyright infringement.”
That argument matters. Willfulness is what unlocks the $150,000 maximum. The plaintiffs are treating Anthropic’s own safety measures as evidence against it.
🧭 Context: How We Got Here
The music industry has been fighting AI for a while now, but mostly through big institutions:
- The RIAA and major labels (Sony, Universal, Warner) have filed copyright claims against AI firms.
- Music publishers have already sued Anthropic over song lyrics in a separate case.
- Book authors went after Anthropic over training data, which ended in a huge settlement in 2025.
The new part is who’s suing. Hacker News describes this as one of the first major legal challenges brought directly by an artist’s estate against an AI developer. Earlier fights were label against platform. This one puts a single iconic artist’s catalog up against a frontier AI lab.
One note: Claude doesn’t generate audio. It’s a text model. So the infringement claims will probably focus on lyrics and compositions, not recordings. That difference could shape how the court looks at the case.
🎯 Why This Matters
What stands out here is the template. If an estate can bring a claim like this and win, every major catalog holder has a playbook. Think about the estates and rights owners behind other legacy artists. Each one could file its own suit, catalog by catalog, track by track.
This also lands while the industry is already pushing back hard on AI music:
- Musicians keep raising alarms about their work and voices being used without consent.
- AI bots have started impersonating real, living artists on streaming platforms.
- Spotify said last week it pulled about 50 million AI “slop” songs from its recommendations over the past year.
The mood has shifted. Rights holders aren’t waiting for regulators anymore. They’re going to court.
🛡️ Tactical Implications
- For AI labs: Training data provenance is now a balance-sheet risk. Expect more licensing deals and fewer “fair use” bets.
- For builders on Claude and similar models: Expect tighter output filters around lyrics and copyrighted text. Prompts that touch song content may start getting refused more often.
- For rights holders: The guardrails-as-evidence argument is new. If it works, it’ll show up in a lot of future complaints.
- For the industry: A ruling here could set precedent for how AI companies can use existing music catalogs.
🔭 Outlook
Anthropic hasn’t publicly answered the specific claims yet. Watch for its first filing, especially whether it argues fair use or challenges the output evidence. Either way, the King’s catalog has just become one of the more important test cases in AI copyright law. Full details are in the original report on Hacker News.