U.S. Judge Denies Music Publishers’ Bid to Halt AI Training by Anthropic

A California federal judge has denied a request from Universal Music Group (UMG), Concord, and ABKCO to stop AI startup Anthropic from using copyrighted song lyrics to train its chatbot, Claude. U.S. District Judge Eumi Lee ruled Tuesday that the music publishers failed to prove they would suffer irreparable harm and that their request was too broad in scope.

The lawsuit, originally filed in 2023, accuses Anthropic of using lyrics from over 500 songs — including those by Beyoncé, The Rolling Stones, and The Beach Boys — without permission. The publishers claim this infringes on their copyrights and damages their licensing market.

Judge Lee noted that the publishers’ motion prematurely asks the court to define the boundaries of a licensing market for AI training, even though the key legal issue of “fair use” has not been resolved. Her ruling did not address whether Anthropic’s use qualifies as fair use under U.S. copyright law.

Anthropic welcomed the court’s decision, calling the publishers’ request “disruptive and amorphous.” The company, like other tech giants including OpenAI, Microsoft, and Meta, argues that using copyrighted materials to train AI systems is transformative and protected by fair use.

The publishers remain confident in their broader case against Anthropic, despite the setback. Their lawsuit is part of a growing wave of legal challenges against AI developers accused of using protected content — from books to music — without consent or compensation.

This case highlights the ongoing legal uncertainty surrounding AI training and copyright, especially as courts begin to weigh the balance between innovation and intellectual property rights in the age of generative AI.