📰 Key Takeaways

Sony Music Publishing, Warner Chappell, and several other music publishers have sued Anthropic and co-founders Dario Amodei and Benjamin Mann, accusing the AI lab of a “brazen campaign of illegal plunder, downloading and scraping massive amounts of copyrighted works via BitTorrent.” The lawsuit was filed late Friday in the U.S. District Court for the Northern District of California, first reported by Music Business Worldwide. The publishers allege Anthropic used thousands of copyrighted works to train its AI model Claude, constituting “brazen theft.” An Anthropic spokesperson responded that the company disagrees with the publishers’ claims and will vigorously defend itself in court.

This isn’t the first time Anthropic has faced an IP lawsuit. Some of the lawyers on this case also represent Concord Music Group and Universal Music Group (UMG), which sued Anthropic back in January. The same legal team previously led Bartz v. Anthropic, in which a group of authors accused Anthropic of using copyrighted works to train Claude and other products. In that landmark case, the judge ruled that training on copyrighted works is itself legal, but obtaining that content through piracy is not — Anthropic ended up paying $1.5 billion in damages.

This new lawsuit goes further, building on the existing case by accusing Anthropic of “rampant piracy” via illegal BitTorrent downloads to obtain millions of book copies, including lyrics and sheet music.

💬 JudyAI Lab Take

Sony Music Publishing, Warner Chappell, and other music publishers have officially sued Anthropic, accusing it of illegally obtaining massive amounts of copyrighted work via BitTorrent and other means to train Claude — heating up again right after the $1.5 billion Bartz judgment.

This case extends the core principle established in Bartz v. Anthropic: training an AI model on copyrighted content isn’t illegal by itself, but obtaining that content through piracy is infringement. This lawsuit pushes the scope further into lyrics and sheet music, showing that the music publishing industry is using the prior ruling as leverage, going hard after the “legitimacy of the source” for training data. For AI builders, this points to an increasingly clear industry line: the legal battlefield for model training has shifted from “can you use copyrighted content” to “how did you obtain it” — traceability of data sources is becoming just as much a competitive factor as model capability.

When you see cases like this, it’s worth checking whether your own project has clear, accountable records of where your training or fine-tuning data came from.

📅 Original Source

🔗 Further Reading