Elvis Presley rights holders sue Anthropic over 274 songs used by Claude

Companies and trusts that own the rights to Elvis Presley's songs have sued the artificial intelligence (AI) company Anthropic, alleging that its chatbot Claude was trained on 274 Presley compositions without permission. The lawsuit was filed on Friday, October 23, 2026, in the federal court for the Northern District of California. The plaintiffs include Authentic Brands Group, Raleigh Music Publishing and Abilene Music, along with several trusts and individuals who say they own or control the songs. Anthropic said it disagreed with the claims and intended to defend the case fully.

Large language models, the technology behind chatbots like Claude, ChatGPT and Gemini, learn by processing huge amounts of text taken from books, websites and other sources. Much of that material is protected by copyright. AI companies argue that training on it is "fair use", a doctrine in US law that allows limited use of copyrighted work without permission for purposes such as research or commentary, as long as it does not harm the market for the original. Rights holders argue that copying their work to build a commercial product, without paying, is plain infringement.

Song lyrics are a sharp test of that argument. Lyrics are short, so a chatbot that reproduces them can hand a user the whole work in one answer. In the publishers' earlier case against Anthropic, the complaint gave the example of Claude producing a near-identical copy of Katy Perry's "Roar" when asked. In January 2025, the court approved an agreement requiring Anthropic to keep "guardrails", built-in limits meant to stop Claude from reproducing copyrighted lyrics. The Presley case adds a famous catalogue and a new set of owners to that fight.

The sums involved are not the largest Anthropic has faced. Anthropic agreed last year to a $1.5 billion settlement with authors who said the company had used pirated copies of books to train its models. A series of such lawsuits, against Anthropic and its rivals, is slowly forcing AI companies to decide whether to defend every case, settle, or sign licensing deals with publishers and record labels.

For India, the outcome matters in two ways. India has a large film and independent music industry whose songs, in many languages, are the kind of material AI systems can learn from. And Indian AI start-ups building chatbots in Hindi, Tamil and other languages face the same question of whose content they may legally train on. How US courts treat fair use in AI training is being watched by rights holders and developers well beyond America.

Short-term: Anthropic must respond to the complaint in court, and the case joins several other copyright suits it is already fighting over music and books.

Long-term: Rulings in these cases will help decide whether AI companies can train on copyrighted songs and texts without permission, or whether they must license them, which could change how AI models are built and what they cost.

Who is affected: AI developers and their users, songwriters and music publishers, estates of famous artists, and creators and start-ups in India and elsewhere who depend on copyright rules for their work or their products.

Questions and Answers