More than 100 authors have hauled Anthropic into federal court in San Francisco, accusing the AI company and two of its co-founders of using pirated books to train the Claude family of models. The suit, filed June 17 in the U.S. District Court for the Northern District of California, demands statutory damages, injunctive relief and a jury trial. According to plaintiffs’ filings, the complaint comes with an exhibit that lists roughly 500 registered works the authors say were taken without permission. The timing is notable, arriving as Anthropic reportedly gears up for an IPO and adding yet another high-stakes chapter to the ongoing copyright battles over AI training data.
The case is styled Shakespeare et al. v. Anthropic PBC, No. 4:2026cv05931, and names Anthropic, CEO Dario Amodei and co-founder Benjamin Mann as defendants. Justia Dockets & Filings shows the complaint was filed on June 17 and confirms that the plaintiffs attached an exhibit identifying the copyrighted works at issue.
What the complaint alleges
In the complaint, the authors claim Anthropic built what they describe as a "vast central library" by downloading files from shadow libraries and torrent sites, then feeding that collection into training datasets for Claude. A press release and the complaint itself allege that material came from datasets and sites including Books3, Library Genesis and the Pirate Library Mirror, and that some company officers personally took part in torrenting the files. EIN Presswire published the plaintiffs’ summary of those allegations when the suit was filed.
Who signed on
The filing is a crowded one, pulling together writers across genres and formats, including fiction, nonfiction, memoir and children’s literature, with both marquee names and smaller presses in the mix. Legal trackers and case lists show dozens of plaintiffs, and note that many of the authors opted out of an earlier class settlement so they could press their own claims instead. Mishcon de Reya and other docket aggregators list the roster and outline the strategy of pursuing separate statutory-damages claims.
What they’re asking for
The authors are seeking statutory damages, injunctive relief and attorneys’ fees. Plaintiffs’ filings say they want statutory damages of up to $150,000 per infringed work and include registration information for the titles they say were copied. Under U.S. law, courts can award up to $150,000 per work for willful infringement, a framework explained by Cornell’s LII. The plaintiffs’ press materials note that the complaint lists roughly 500 registered works, a number that, at least in theory, could put maximum statutory exposure in the tens of millions. EIN Presswire published the plaintiffs’ summary of the damages claim.
How this fits in the wider fight
This new case does not arrive in a vacuum. It follows the separate Bartz v. Anthropic litigation, which produced a proposed $1.5 billion settlement and a 2025 ruling that split the legal issues: a judge held that training on lawfully acquired books can qualify as fair use, while also finding that keeping pirated copies was not protected. That mixed ruling, along with the settlement, has shaped later suits and objections from both authors and publishers. As reported by AP News and in coverage of Anthropic’s internal data practices, those earlier decisions and disclosures form the backdrop for this latest complaint.
What happens next
The docket currently shows that summonses were issued on June 17 and that an initial case-management order is on file. The court has assigned the case to U.S. District Judge Jon S. Tigar and set scheduling deadlines tied to early case-management steps. Both the complaint and its registration exhibit have been entered on the public docket. Justia Dockets & Filings lists the filings and procedural entries.
Why it matters
Attorneys and industry watchers say the suit highlights how the rules for training-data provenance are still being hammered out in court, and that AI developers will keep facing pressure to document and license the texts that power their models. A growing line of complaints from authors, publishers and the music industry against Anthropic and other labs is putting fresh focus on licensing, audit trails and corporate disclosure around training pipelines. Legal Clarity and other legal trackers are watching how opt-outs and individual suits might reshape settlement talks and expectations for future licensing.
