Anthropic copyright lawsuit roundup: Canadian news publishers sue in Ontario, and a judge weighs fair use for song lyrics

The newest Anthropic copyright lawsuit comes from Canada. Six news publishers, including The Globe and Mail, Postmedia, the Toronto Star and The Canadian Press, filed a statement of claim against Anthropic in the Ontario Superior Court on October 6, 2026. They allege the company scraped their articles to build Claude without a license. The suit was first reported by Cartt and was covered widely by October 19. Two days later, on October 21, a federal judge in San Jose heard arguments on whether training Claude on song lyrics is fair use, in a case music publishers filed in 2023. She didn't rule. The authors' case that came before both, Bartz v. Anthropic, ended in a $1.5 billion settlement, as our Anthropic explainer describes.

This story is based on the Concord court docket, reporting by The Canadian Press, Broadcast Dialogue, Reuters, Law360, Courthouse News and Music Business Worldwide, court-tracking pages and the Authors Guild, checked October 23, 2026. The Canadian Press is itself a plaintiff in the Ontario case. Nothing here is legal advice.

The short version

• Canada: six publishers sued Anthropic and its Canadian subsidiary on October 6, 2026, over about 13.5 million articles published since 2021, according to Broadcast Dialogue's reading of the claim.
• Claims: copyright infringement, breach of the websites' terms, circumvention of technical protections such as paywalls, and unjust enrichment. The allegations haven't been proven.
• Music: Judge Eumi Lee heard summary judgment motions in Concord Music Group v. Anthropic on October 21. Law360 reported she suggested AI "mash-up" lyrics might be fair use; Courthouse News reported she didn't indicate her thoughts. No decision had been issued as of October 23.
• Context: at least five other music suits are pending against Anthropic in California, and the authors' Bartz case has settled.

What are the Canadian publishers suing Anthropic for?

The plaintiffs are the Toronto Star, The Globe and Mail, Metroland Media, Postmedia Network, PNI Maritimes and The Canadian Press, according to Broadcast Dialogue, which credits its sister publication Cartt with the first report. The Canadian Press, in a story by Anja Karadeglija carried by CP24, lists the core claims:

• copyright infringement, using the publishers' content to develop Claude "without obtaining a valid licence";
• breach of the news sites' terms of service;
• circumvention of technological protection measures meant to stop copying;
• scraping both for model development and on an ongoing basis in response to user prompts;
• unjust enrichment. The claim says Anthropic "has been substantially and unjustly enriched to the detriment of the News Media Companies."

Broadcast Dialogue reports that the claim names Anthropic's ClaudeBot crawler and alleges it got around paywalls, login authentication and network-level protections. Anthropic says ClaudeBot respects robots.txt, as our guide to blocking AI crawlers documents. The claim's allegations about paywalls go beyond that question, and a court will have to weigh them. Many publishers have moved to block AI crawlers outright, a trend covered in our analysis of websites closing their doors to AI agents.

The publishers seek general damages, an accounting and disgorgement of profits or, as an alternative, statutory damages of up to $20,000 per infringed work, plus punitive damages and permanent injunctions, Broadcast Dialogue reports. The claim also cites the Bartz settlement.

Anthropic hadn't responded to The Canadian Press's request for comment. No statement of defence had been reported as of October 14, and Broadcast Dialogue notes that a defendant served outside Ontario has up to 60 days to respond. The same publishers, with CBC/Radio-Canada, sued OpenAI in 2024. That case hasn't been settled, according to The Canadian Press.

What happened at the Concord fair use hearing?

On October 21, 2026, Judge Eumi Lee of the US District Court for the Northern District of California heard competing motions in Concord Music Group v. Anthropic (case 5:24-cv-03811). The publishers, Universal Music, Concord and ABKCO, asked her to rule that Anthropic infringed. Anthropic asked her to rule that its training is fair use. She also heard Daubert motions over expert evidence; the court's minute entry says the hearing will be continued for the remaining Daubert motions.

What's been reported:

• Law360, in a story headlined "Anthropic's Lyric 'Mash Up' May Be Fair Use, Judge Says," reported that Lee appeared receptive to Anthropic's fair use argument during a daylong hearing, suggested AI-generated "mash-up" lyrics might be fair use, and questioned how to "slice and dice" the case. The rest of the story is paywalled.
• Reuters, as relayed by MediaNama, reported that Lee pressed both sides and questioned the publishers' claim that Claude's ability to write lyrics threatened their market.
• The arguments: Anthropic's lawyer, Sonal Mehta, said training produces "transformative new content." The publishers' lawyer, Matt Oppenheim, said Claude received nearly 120,000 requests for synthetic lyrics in six months, according to MediaNama and Südtirol News.
• Courthouse News read the hearing differently: Lee asked the publishers' lawyer whether his argument would give publishers a monopoly on lyrics, and took the issue under submission "without indicating her thoughts on summary judgment."
• No ruling: the court's docket showed no decision as of October 23, 2026; the latest entries were transcript orders filed October 22.

A judge's questions at a hearing aren't a decision. In March 2025, Lee denied the publishers' request for a preliminary injunction and said fair use remained an unsettled question, Reuters reported at the time. The case is the first district court summary judgment hearing on fair use for AI training since the June 2025 rulings in Bartz and in Kadrey v. Meta, according to ChatGPT Is Eating the World, a site that tracks the dockets. Its author disclosed joining an amicus brief supporting Anthropic.

What other copyright cases does Anthropic face?

In Bartz, Judge William Alsup ruled in June 2025 that training on books was fair use but that building a central library from pirated copies was not. The case settled for $1.5 billion, about $3,000 per work, and Judge Araceli Martínez-Olguín granted final approval on July 20, 2026, according to the Authors Guild.

The later music suits lean on that piracy finding. Concord II, filed in January 2026, alleges "illegal torrenting of copyrighted works" and names CEO Dario Amodei and co-founder Benjamin Mann, TechCrunch reported. Sony Music Publishing and Warner Chappell's August 28 suit also names both men, Music Business Worldwide reported. Round Hill and Sony seek up to $150,000 per work willfully infringed. Judge Lee is overseeing the related music cases, according to the docket tracker.

Why it matters

Lee's ruling, whenever it comes, will be the next US district court decision on whether training an AI model on copyrighted text is fair use. The Canadian case tests similar conduct under Canadian copyright law rather than US fair use. Both arrive as Anthropic prepares for a reported November IPO; what its draft prospectus reportedly shows is in our prospectus analysis.

What's confirmed, reported and uncertain

• Confirmed (court records and trackers): the Concord I case number, the October 21 hearing, no ruling on the docket as of October 23, the Bartz final approval on July 20, 2026.
• Reported: the Ontario plaintiffs, claims and remedies (The Canadian Press, Broadcast Dialogue); Lee's remarks (Law360, Reuters via MediaNama, Courthouse News); the other music suits (Reuters, TechCrunch, MBW).
• Uncertain: when Lee will rule, and which way she leans, since Law360 and Courthouse News read the hearing differently; how Anthropic will respond in Ontario. The statement of claim itself wasn't available to read, so its details come from press accounts.