Anthropic sued over alleged theft of 'tens of thousands' of songs
Sony Music Publishing and Warner Chappell filed a multibillion-dollar lawsuit against Anthropic, alleging the unauthorized use of tens of thousands of copyrighted songs to train Claude AI models The suit claims Anthropic scraped lyrics from pirate websites, legal platforms like Musixmatch and LyricFind, and archive datasets such as Common Crawl, while stripping attribution metadata from the works Damages could reach billions, calculated at up to $150,000 per infringed work and $25,000 per instan
Analysis
TL;DR
- Sony Music Publishing and Warner Chappell filed a multibillion-dollar lawsuit against Anthropic, alleging the unauthorized use of tens of thousands of copyrighted songs to train Claude AI models
- The suit claims Anthropic scraped lyrics from pirate websites, legal platforms like Musixmatch and LyricFind, and archive datasets such as Common Crawl, while stripping attribution metadata from the works
- Damages could reach billions, calculated at up to $150,000 per infringed work and $25,000 per instance of altered identifying information, with at least 20,000 compositions allegedly involved
- This follows Anthropic's recent $1.5 billion settlement with US authors over similar copyright claims, and coincides with separate lawsuits from Universal Music Group and Concord Music Group
- Anthropic has denied the allegations and stated it will defend itself robustly in court, even as it prepares for a potential $2 trillion stock market listing
Why It Matters
This lawsuit represents a major escalation in the ongoing legal battle between the AI industry and creative rights holders, specifically targeting the music publishing sector. It signals that copyright claims are expanding beyond text and authors into music lyrics and compositions, potentially setting precedents that could reshape how AI companies source and license training data across all creative domains.
Technical Details
- Anthropic is accused of "torrenting, scraping and downloading" copyrighted song lyrics and sheet music from multiple sources, including pirate websites (allegedly 7 million book copies), legal lyric platforms (Musixmatch, LyricFind), and public archive datasets (Common Crawl)
- The plaintiffs allege that Anthropic deliberately stripped identifying information and attribution metadata from the lyrics during data processing, denying copyright owners credit for their work
- The alleged training data includes high-profile compositions such as Mariah Carey's "All I Want for Christmas Is You," Survivor's "Eye of the Tiger," Bon Jovi's "Livin' on a Prayer," and Leonard Cohen's "Hallelujah"
- The lawsuit targets both the company and its leadership, naming CEO Dario Amodei and co-founder Benjamin Mann individually, and is filed in a California court
- The damages framework mirrors the recent authors' settlement: statutory damages of up to $150,000 per willful infringement and $25,000 per act of metadata removal, applied across an estimated 20,000+ works
Industry Insight
- AI companies will face increasing pressure to establish transparent, licensed data pipelines for creative content; the trend suggests that relying on scraped public data will become legally and financially unsustainable, especially as settlement benchmarks like the $1.5 billion authors' deal raise the cost of non-compliance
- The music industry is likely to accelerate coordinated legal action against AI firms, following the precedent set by the GEMA vs. Suno case and the growing coalition of publishers and rights holders, which could force industry-wide licensing standards
- Anthropic's $2 trillion valuation ambitions may be complicated by these mounting legal liabilities, and the outcome could influence how the broader AI sector approaches intellectual property—potentially driving investment toward licensed data markets or synthetic data alternatives
Disclaimer: The above content is generated by AI and is for reference only.