Anthropic seeks dismissal of personal claims against CEO and output-infringement claims in $3B music publisher suit

What happened

Anthropic and its CEO Dario Amodei petitioned a California federal court to dismiss certain aspects of a $3 billion copyright infringement lawsuit filed by major music publishers including Universal Music Group and Concord Music Group. The company challenges the claims against Amodei personally — arguing there is insufficient evidence to hold him personally liable for the alleged infringements — and disputes the allegation that the outputs generated by its AI models constitute direct copyright infringement.

The underlying suit (filed January 2026) accuses Anthropic of unlawfully acquiring and using more than 20,000 copyrighted musical works — songs, sheet music, and lyrics — to train its AI models, calling it "flagrant piracy." The publishers seek up to $150,000 per infringed work, which could total more than $3 billion. Anthropic says it will defend itself vigorously and maintains its training practices are lawful.

Why it matters to creators

This is the first procedural test of two questions that will matter across every AI training-data lawsuit: whether founders and executives can be held personally liable for training-data decisions, and whether model outputs alone (as opposed to training ingestion) can constitute direct infringement. If the personal-liability claims survive, it raises the stakes for individuals authorizing data acquisition inside AI companies; if outputs are excluded, it narrows what publishers can recover.

What is not yet known

The court has not yet ruled on the motion to dismiss; these are arguments, not findings. The separate August 2026 Sony/Warner Chappell music-publisher suit against Anthropic is unaffected and continues in parallel.