Elvis Presley Rights Holders Sue Anthropic Over 274 Songs: What the $41M Claude Lawsuit Means for Brands

Elvis Presley's rights holders sued Anthropic on October 23, 2026, alleging Claude was trained on lyrics from 274 Presley-associated songs without a license and still produces near-copies of them. The plaintiffs, led by an Authentic Brands Group subsidiary, seek about $41.1 million in statutory damages. Anthropic says it disagrees and will defend the case, and it has argued elsewhere that AI training is fair use.

What exactly does the Elvis lawsuit claim?

The complaint says Anthropic copied Presley's song lyrics to train Claude and keeps copying them whenever users ask for lyrics or new versions. According to The Next Web, the suit was filed in the US District Court for the Northern District of California. It also alleges that Anthropic got around technical protections and stripped or changed copyright management information.

The songs named include:

• "Can't Help Falling in Love"
• "All Shook Up"
• "Don't Be Cruel"
• "A Little Less Conversation"
• "Return to Sender"

Rolling Stone reports the caption as ABE EPE LLC v. Anthropic and describes it as one of the first major cases in which an artist's estate has sued an AI company.

Who is suing, and is the $41 million figure accurate?

Most outlets report the same number, and it is the statutory copyright maximum. Plaintiffs are asking for up to $150,000 per song across 274 works, which comes to roughly $41.1 million. NME and TNW both cite $41.1 million, and Rolling Stone rounds it to about $41 million. Fox News adds that the plaintiffs also want further damages on other claims, so $41.1 million should be treated as the floor of the request rather than the total.

The plaintiff group is broader than "the Elvis estate":

1. A subsidiary of Authentic Brands Group, which controls much of Presley's IP, including his name, image and likeness.
2. Raleigh Music Publishing, plus several trusts and at least one individual, according to TNW.

Lead counsel is Richard Busch, who represented Marvin Gaye's estate in the "Blurred Lines" case. He has framed the suit as a defense of iconic songs.

Why do the "guardrails" matter so much?

The guardrails are the center of the complaint, because they turn a dispute about training into a dispute about outputs. After earlier publisher suits, Anthropic reportedly added filters to stop Claude from reproducing lyrics. The complaint says those filters fail. When asked to recreate songs "for a modern audience," Claude allegedly produced versions that kept the originals' structure and hooks. Rolling Stone reports that the complaint quotes Claude itself saying the renders kept "landing very close to the originals."

Training claims can be met with fair use; a claim that the product keeps producing derivative works on request is harder to dismiss that way.

What is Anthropic's side of the story?

Anthropic denies the claims. In a statement to Fox News Digital it said: "We disagree with the claims and intend to defend this case fully." Its position in related cases is known:

• Training as fair use. Anthropic has argued that training models on copyrighted lyrics and text is protected under US fair use law.
• A favorable precedent. In Bartz v. Anthropic, Judge William Alsup ruled in June 2025 that training Claude on books was "exceedingly transformative" and therefore fair use, Publishers Weekly reported.
• The limit of that precedent. In the same ruling, Alsup found that building a permanent library from pirated copies was not fair use. Anthropic later agreed to a $1.5 billion settlement with authors.

That precedent cuts both ways for the music cases. The training itself may be defensible, while how the data was acquired and what the model outputs may not be.

How does this fit with the other music cases against Anthropic?

This is the latest in a series of suits, and the plaintiffs' lawyers are coordinating across them:

1. October 2023: UMPG, Concord and ABKCO sued over Claude reproducing lyrics.
2. January 2026: Publishers led by Concord and UMG filed a separate suit over more than 20,000 works, saying damages could exceed $3 billion and alleging torrenting, TechCrunch reported.
3. March 2026: BMG filed its own case.
4. August 2026: Sony Music Publishing and Warner Chappell sued, and so did Round Hill Music.
5. October 2026: Reservoir Media sued, followed days later by the Elvis plaintiffs. Busch represents Round Hill, Reservoir and the Elvis group.

The timing also stands out. TNW notes the Elvis suit landed three days before Anthropic made its IPO filing public, which raises the profile of every pending claim. Background on the filing: Anthropic's S-1, revenue and losses explained.

What should brands and IP owners do when their work shows up in AI answers?

Start by measuring. Most brands have no idea how often AI assistants reproduce, paraphrase or riff on their protected material, and they can't make a legal or commercial decision without that evidence. The Elvis complaint rests on prompt-and-response evidence, a method open to any rights holder.

A practical playbook:

1. Run structured prompt audits. Ask major assistants to quote, summarize, "modernize" or imitate your signature assets: lyrics, slogans, characters, product copy. Log the outputs with dates and model versions.
2. Separate mention from misuse. A model that recommends your brand helps you. A model that reproduces your IP or invents content under your name hurts you. Track the two separately. MarkGrid's Model Share measures how often and how accurately AI models surface a brand compared with competitors.
3. Watch impersonation and fake attributions. AI tools can put real names on things they never made, as we covered with ChatGPT's fake New Yorker cartoons that carried real artist signatures. For creator-led brands, Creator Audit helps check whether creator associations are authentic.
4. Keep visibility and protection in balance. Blocking all AI access can erase you from the answers customers now rely on. Our GEO guide covers how to stay discoverable while limiting what can be copied.
5. Bring legal and marketing together early. Evidence from brand monitoring is often what counsel needs to judge whether licensing, takedowns or litigation make sense.

The aim isn't to sue every chatbot. It is to have a record before you need one.

FAQ

Who filed the Elvis Presley lawsuit against Anthropic?

The plaintiffs include a subsidiary of Authentic Brands Group, which controls Presley's name and likeness, along with Raleigh Music Publishing, several trusts and at least one individual. The case is reported as ABE EPE LLC v. Anthropic.

How much money are the Elvis rights holders seeking?

They want up to $150,000 per song in statutory damages across 274 songs, about $41.1 million in total, plus additional damages on other claims and an injunction.

What has Anthropic said about the lawsuit?

Anthropic told Fox News Digital it disagrees with the claims and intends to defend the case fully. In other cases it has argued that training on copyrighted material is fair use.

Is training AI on song lyrics legal?

No court has decided this for lyrics. In a 2025 books case, a federal judge found training on lawfully acquired books to be fair use but ruled that pirated copies were not. Music cases add claims about model outputs.

Why does the "for a modern audience" detail matter?

The complaint alleges that Claude produced near-copies of the songs even with guardrails in place. That moves the dispute from training data to infringing outputs, which is harder to defend as fair use.

How can a brand tell if AI models are reproducing its IP?

Run repeatable prompt audits across major assistants, log the outputs, and track mentions separately from reproductions or false attributions.

Last updated October 26, 2026.