Sony and Warner sue Anthropic over alleged illegal song scraping
Sony Music Publishing and Warner Chappell Music have filed a federal lawsuit against Anthropic, alleging the company used copyrighted music to train its Claude chatbot without authorization. The publishers are seeking statutory damages and an injunction, challenging Anthropic's fair use defense regarding generative AI training data.
Key Takeaways
- Publishers allege Anthropic used a 'brazen campaign' of torrenting and scraping to acquire training data for Claude.
- The lawsuit cites specific protected works including Marvin Gaye's 'Ain't No Mountain High Enough' and Mariah Carey's 'All I Want for Christmas is You.'
- Anthropic claims its training methods constitute 'transformative fair use' and cites the Bartz v. Anthropic settlement as a legal precedent.
- Legal counsel for the music publishers previously represented authors in a similar $1.5 billion copyright dispute against the AI firm.
Why It Matters
This litigation signals a hardening stance by major rights holders against generative AI firms that bypass traditional licensing frameworks. By targeting individual executives and seeking maximum statutory damages, Sony and Warner are attempting to establish that AI training on copyrighted catalogs requires explicit authorization rather than falling under fair use protections. This conflict mirrors ongoing disputes involving AI music generators Suno and Udio, suggesting a broader industry effort to force tech firms into revenue-sharing agreements. Watch for the court's ruling on whether 'transformative use' applies to large-scale scraping, as this will dictate the future cost structure for AI companies utilizing premium media content.
Additional Context
The Sony and Warner lawsuit against Anthropic is part of a broader wave of copyright enforcement actions targeting AI companies that train on creative works. In June 2024, the Recording Industry Association of America filed lawsuits against AI music generators Suno and Udio on behalf of major labels including Sony, Universal, and Warner, alleging the startups copied millions of recordings to build their generative models without permission. Those cases remain active in federal courts in Massachusetts and New York, and the outcome is expected to influence how judges treat the fair use defense in the Anthropic matter. The publishers' decision to name Dario Amodei and Benjamin Mann individually mirrors a strategy seen in the New York Times lawsuit against OpenAI and Microsoft filed in December 2023, which similarly sought to pierce the corporate veil and hold leadership personally accountable for training-data decisions.
On the licensing and business side, the music industry has begun negotiating structured deals with AI firms even as litigation proceeds. In early 2025, Universal Music Group announced a licensing agreement with AI music platform Udio, signaling that rights holders are willing to monetize AI training when terms are favorable. Warner Music Group has pursued a parallel path, with CEO Robert Kyncl stating in a February 2025 earnings call that the company is in active discussions with multiple AI companies about licensing its catalog. These negotiations suggest that the Sony and Warner lawsuit against Anthropic may be less about blocking AI training entirely and more about establishing leverage for eventual revenue-sharing terms. The statutory damages demand of $150,000 per work creates a financial ceiling that makes settlement economically rational for Anthropic if the court finds infringement.
From a technical and legal precedent standpoint, the Anthropic case will likely hinge on whether courts apply the reasoning from the Authors Guild v. Google decision, which held that mass digitization for search purposes constituted transformative fair use. However, legal scholars have noted that generative AI training differs materially from index-based search because models can produce outputs that substitute for the original works. A Stanford Law School analysis published in March 2025 argued that the transformative use doctrine faces its most significant stress test in AI training cases, with outcomes likely to diverge based on whether the AI system's outputs compete directly with the training data. For Anthropic, the risk is compounded by evidence that Claude can reproduce song lyrics and musical structures, which the publishers will argue demonstrates market substitution rather than transformation.
Read full article at variety.com
Enjoy our coverage?
Add StreamingMeme as a preferred source on Google to see more of our streaming news at the top of your Search results.
Add as preferred source