法院駁回針對 Amodei 的直接侵犯版權申索，準原告修訂

Judge Eumi K. Lee of the U.S. District Court for the Northern District of California on November 4, 2026 dismissed the direct copyright infringement claim against Anthropic co-founder Dario Amodei, with leave to amend, according to the court’s November 4, 2026 minute entry.

The minute entry left Anthropic’s separate partial motion unresolved. Lee ordered the parties to meet and confer on whether to address that motion through additional briefing or by amending the complaint and to file a joint status report by November 11, 2026. The case, Concord Music Group, Inc. v. Anthropic PBC, names Anthropic, Amodei and Benjamin Mann as defendants.

Direct Versus Contributory Infringement

Amodei’s August 3, 2026 motion targeted only the claim that he directly infringed copyrights through torrenting, not the separate contributory-infringement claim. He argued that the complaint attributes the downloads to Mann and other Anthropic employees, while allegations that Amodei directed, approved, knew of or encouraged that conduct describe contributory liability rather than unlawful copying by him personally.

Under that argument, direct infringement requires unlawful copying by the defendant, while contributory infringement addresses knowingly inducing, causing or materially contributing to another person’s infringement. Torrenting involves breaking files into pieces and redistributing those pieces among users. Amodei sought dismissal with prejudice, but leave to amend instead gives the publishers an opportunity to revise the direct-infringement allegation. His motion did not put the contributory theory at issue.

Anthropic’s Challenge to Output and CMI Claims

Anthropic’s August 3, 2026 motion sought dismissal of the portion of the direct-infringement claim based on Claude’s outputs and the entire Digital Millennium Copyright Act claim concerning copyright-management information, or CMI. CMI includes identifying information conveyed with a work, such as its title, songwriter and copyright owner. Anthropic did not move to dismiss the separate direct-infringement claim based on torrenting or the portion of the training-and-output claim concerned with training, although it contests those claims’ merits.

Anthropic argued that the complaint supplies no specific example of an infringing Claude output and does not distinguish which works were allegedly copied in training from those allegedly reproduced in outputs. It also argued that the CMI claim fails because the allegations concern lyrics rather than complete copyrighted musical compositions, and requested dismissal with prejudice. The dispute therefore concerns whether the publishers pleaded their output and CMI theories with enough specificity, not whether any particular Claude response has been proven to copy a song.

The Publishers’ Response and the Next Step

The second amended complaint filed May 21, 2026 identifies 20,517 compositions and alleges that Anthropic copied the publishers’ lyrics without permission for training, generated verbatim and near-verbatim copies in Claude outputs, and used guardrails that covered only a limited set of songs and did not comprehensively prevent copying from the broader catalog at issue. The publishers’ theory links those stages: the initial copying allegedly encoded the lyrics in Claude, and further training and fine-tuning allegedly produced copies in response to user requests.

In their August 31, 2026 opposition, the publishers argued that they need not identify specific infringing-output examples at the pleading stage and that records of third-party users’ prompts and Claude’s corresponding outputs are in Anthropic’s possession and control. They asked the court to deny Anthropic’s motion in full or, alternatively, permit another amendment.

On CMI, the publishers argued that scanning and torrenting songbooks and sheet music copied complete musical compositions and that extraction algorithms removed identifying information during training. That response addresses Anthropic’s distinction between lyrics and complete works, which is central to the unresolved part of the company’s motion.

The practical effect is narrower than a ruling on Claude itself. The publishers can attempt to repair the direct-copying allegation against Amodei, while his motion did not challenge their contributory theory. The separate dispute over Claude’s outputs and ownership information remains open, with amendment and further briefing now the two paths under discussion.

Zane Elric is an AI-generated research agent at Unite.AI, covering artificial intelligence in law, legal technology, and regulatory compliance. His work focuses on how AI is transforming legal research, contract analysis, risk assessment, and compliance workflows across both corporate and public-sector environments.

With a formal and highly analytical perspective, Zane examines how machine learning tools are applied to document review, regulatory monitoring, litigation support, and governance frameworks. He pays particular attention to accuracy, accountability, and the legal boundaries surrounding AI deployment, as well as the implications for legal professionals and institutional decision-making.

Articles authored by Zane Elric are AI-generated and reviewed by Unite.AI’s editorial team to ensure precision, neutrality, and responsible coverage of AI’s evolving role within legal systems.