Judge to hear key fair-use arguments in music publishers' AI training lawsuit against Anthropic

What happened

On October 21, 2026, U.S. District Judge Eumi Lee in San Jose, California, was scheduled to hear arguments on a key defense by Anthropic in a copyright lawsuit brought by music publishers, including Universal Music Group, Concord, and ABKCO, over the use of their song lyrics to train Anthropic's chatbot Claude.

The case

The publishers sued Anthropic in 2023, alleging that its use of their song lyrics to train Claude infringed their copyrights in hundreds of songs by artists including Beyoncé, the Rolling Stones, and the Beach Boys. Anthropic has argued that its AI training makes transformative use of the lyrics, saying it helps Claude "understand human language and enable progress and productivity in science, business and education."

Context

The dispute turns on whether generative-AI training makes "fair use" of copyrighted material. Two other Northern California judges have already split on the question in closely watched AI-training cases: now-retired Judge William Alsup sided with Anthropic last year on its argument that training on books was fair use, while Judge Vince Chhabria ruled for Meta in a similar lawsuit but warned that AI training "in many circumstances" would not be fair use. How judges like Lee rule on the fair-use defense could shape dozens of similar lawsuits brought by authors, news outlets, and visual artists against AI companies.

Note: this item records the hearing as reported; no ruling on the arguments had been reported at the time of writing. All statements about the parties' positions are allegations and arguments, not findings.