Anthropic Defends AI Training as Fair Use in Music Publishers Copyright Lawsuit Over Claude Chatbot

Anthropic Faces Music Publishers in Pivotal AI Copyright Case

Anthropicdefended its AI training practices before US District Judge Eumi Lee in San Jose, California, arguing that using millions of song lyrics to train its chatbot Claude qualifies as fair use under U.S. copyright law1

1

2

2

. Music publishers including Universal Music Group, Concord, and ABKCO brought the copyright lawsuit in 2023, alleging that Anthropic infringes copyright on hundreds of songs by artists including Beyoncé, the Rolling Stones, and the Beach Boys1

1

. Judge Eumi Lee pressed both sides during Wednesday's hearing but did not issue a ruling, questioning whether Claude's lyric-generation capabilities genuinely threaten the market for original lyrics2

2

.

The Fair Use Defense and Its Implications for AI

Anthropicargued through lawyer Sonal Mehta of WilmerHale that its AI training creates transformative content, positioning it squarely within fair use doctrine protections2

2

. The company maintains that using copyrighted material helps Claude understand human language and enables progress in science, business, and education1

1

2

2

. Anthropic emphasized that allowing copyright holders to veto transformative technology would contradict the Supreme Court's mandate that copyright's primary objective serves the public rather than rewarding authors1

1

. Mehta contended that AI has not reduced demand for original lyrics and that music publishers cannot identify specific instances of market harm from Claude or similar large language models2

2

.

Music Publishers Challenge AI-Generated Content Impact

The music publishers counter that Anthropic misuses their lyrics to generate competing content that harms human songwriters2

2

. Attorney Matt Oppenheim of Oppenheim + Zebrak revealed that Claude received nearly 120,000 requests to generate synthetic lyrics over a six-month period2

2

. The publishers argue that Anthropic's use of song lyrics is not fair because it enables Claude to generate AI-generated lyrics that dilute the market for their works1

1

. Oppenheim warned that large language models could overwhelm the market with AI-generated content very easily, pointing to AI songs created with tools like Claude saturating streaming platforms and climbing Billboard charts1

1

2

2

.

Precedent-Setting Implications for AI Industry

How Judge Eumi Lee rules on the fair use defense will likely determine the fate of dozens of high-stakes lawsuits brought by authors, news outlets, visual artists, and others against tech giants including OpenAI and Meta Platforms over their AI training practices1

1

2

2

. Two Northern California judges have already split on whether generative AI training generally makes fair use of copyrighted material, creating uncertainty for tech companies using millions of copyrighted works to train their systems without content owners' permission1

1

2

2

. Now-retired US District Judge William Alsup sided with Anthropic last year, ruling that its AI training made fair use of books, while US District Judge Vince Chhabria ruled for Meta in a similar lawsuit but cautioned that AI training in many circumstances would not qualify as fair use1

1

2

2

. Legal observers will watch closely as this legal battle could establish whether AI companies can continue training models on copyrighted material or face substantial liability for using transformative content without authorization.