A coalition of Canada's largest news publishers is suing Anthropic. Postmedia, Torstar, The Globe and Mail and The Canadian Press filed a statement of claim in an Ontario court on October 6, 2026, saying Anthropic infringed their copyright by using their content to develop Claude. CBC News reported the filing on October 19, citing the news site Cartt.ca as the first to report it. Anthropic had not responded to a request for comment, and, as CBC notes, "the allegations have not been proven in court."
TL;DR
| Question | Answer |
|---|---|
| Who sued? | Postmedia, Torstar, The Globe and Mail, The Canadian Press |
| Against whom? | Anthropic |
| Where and when? | Ontario court, statement of claim filed October 6, 2026 |
| Core claim | Unlicensed scraping of news sites to develop Claude |
| Other claims | Breach of terms of service; circumventing technological protection measures; ongoing scraping in response to user prompts; unjust enrichment |
| Anthropic response | None reported yet |
| Status | Allegations only; nothing decided |
What does the claim say?
CBC quotes the statement of claim directly. On the commercial argument: "Anthropic has capitalized on the commercial success of its AI models, building an expansive suite of services, and raised significant capital — all without obtaining a valid licence from any of the News Media Companies." On enrichment: Anthropic "has been substantially and unjustly enriched to the detriment of the News Media Companies."
The publishers describe themselves as responsible for "a significant portion of Canada's journalistic content" with millions of readers collectively. They say Anthropic "deliberately 'scrapes' (i.e. accesses and copies) enormous amounts of text data from the internet and from the Plaintiffs' Websites," then uses it without authorization or consent to develop its models.
Three details stand out:
- Terms of service. The claim says Anthropic breached the websites' terms of service. That is a contract-style argument, separate from copyright.
- Technological protection measures. The claim says Anthropic circumvented technical measures the sites use to restrict copying. In Canada, circumventing such measures can be a separate violation under copyright law, which is why plaintiffs plead it.
- Ongoing scraping on prompts. The claim says Anthropic scrapes content on an ongoing basis in response to user prompts, not only in a one-time training run. That targets retrieval and browsing features, not just the training dataset.
Scales and a folder, representing the legal filing by Canadian news publishers against Anthropic
How does it compare with the OpenAI case?
The same publishers, together with CBC/Radio-Canada, launched a joint suit against OpenAI in November 2024; The Globe and Mail's release at the time described it as legal action over OpenAI using news content to train ChatGPT. BetaKit also covered it. CBC reports that case "has yet to be settled," and a related CBC link notes Ottawa is weighing plans on AI and copyright as OpenAI fights Ontario court jurisdiction.
Two differences are visible from the reporting. CBC/Radio-Canada, a public broadcaster, is named in the OpenAI case but not listed among the plaintiffs in the Anthropic report. And the Anthropic claim explicitly pleads the circumvention of protection measures and the prompt-time scraping, which the CBC summary of the claim spells out.
CBC also notes that in the US, the New York Times sued OpenAI, and the Justice Department sided with the tech company last month. We could not verify that detail beyond CBC's line, so read it as CBC's report.
Where this sits in the wider copyright fight
Anthropic is now a defendant in several media fights. Music publishers and labels are pursuing it, covered in Sony and Warner suing Anthropic over per-song damages. Newspapers are suing OpenAI, as in USA Today's $250 million claim. Our running counter of cases is in the AI lawsuit tracker post.
Anthropic's relationship with Canada is otherwise investment-oriented: the company committed funding to Canadian research, covered in Anthropic's CAD 10 million commitment. The suit shows that good relations with research institutions do not settle questions with publishers.
What would the case need to show?
We are not lawyers, and this is not legal advice. In general terms, a claim like this tends to turn on questions courts have been working through elsewhere:
- Copying. Did Anthropic make copies of the publishers' articles, and can the plaintiffs show which ones? Statements of claim assert it; proof comes with disclosure.
- Exceptions. Canadian law has fair-dealing exceptions, but they are narrower than the US concept of fair use, so the outcome may not track US decisions.
- Access controls. If the sites used paywalls or technical blocks, whether Anthropic's crawlers bypassed them is a factual question.
- Jurisdiction. CBC points to OpenAI fighting Ontario's jurisdiction. Anthropic may do the same.
- Remedies. The CBC summary does not state a damages figure, so none should be assumed.
What it means for builders and publishers
If you run a site, the practical lessons are the same ones we raised when covering crawler traffic: state your terms, keep access controls, and keep logs. Our posts on Meta's commercial terms for crawling publisher sites and on the flood of AI bot traffic hitting small sites cover the mechanics of robots.txt and rate limits.
If you build with Claude, nothing changes today. But if your product summarizes or reproduces news, expect licensing to become a line item, and ask your vendor what it says about indemnity.
Claimed versus confirmed
Because this story rests on a single wire report of a court filing, it helps to separate what is on the record from what is not.
| Item | Status |
|---|---|
| Filing date October 6, 2026 | Reported by CBC/Canadian Press, first by Cartt.ca |
| Plaintiffs: Postmedia, Torstar, The Globe and Mail, The Canadian Press | Reported; we have not seen the filing |
| Quoted language about licences and enrichment | Quoted by CBC from the statement of claim |
| That Anthropic scraped the sites, bypassed protections or breached terms | Allegation only |
| Damages sought | Not stated in the CBC report |
| Anthropic's position | Not yet given |
| Court outcome | None; the case is at the pleading stage |
How to check for yourself: the Ontario court record is the primary source, and Cartt.ca carried the first report. Any outlet that quotes figures not in the CBC piece should cite where they come from.
Why the "on prompt" scraping allegation matters
Most copyright cases against AI labs focus on training data collected in the past. The Canadian claim also says Anthropic scrapes content on an ongoing basis in response to user prompts. That targets a different layer of the product: features where Claude fetches a live page to answer a question.
For a publisher this matters because live retrieval can substitute for a visit without leaving a record in any training set. For a lab it matters because browsing and search tools are now core to assistants, and any ruling on whether fetching a page on demand is copying would reach well beyond one dataset. We do not know yet how the court would treat that, and the claim as reported does not say which Claude features it means.
A brief history of publishers versus AI labs
- November 2024. Postmedia, Torstar, The Globe and Mail, The Canadian Press and CBC/Radio-Canada sue OpenAI in Ontario, per the Globe release. They allege OpenAI scraped their content to build ChatGPT.
- Since then. CBC says that case is unsettled, and OpenAI is contesting Ontario's jurisdiction.
- US. The New York Times sued OpenAI over use of its copyrighted material; CBC says the Justice Department sided with the tech company last month, which we have not independently confirmed.
- September 2026. Sony and Warner went after Anthropic over songs, covered above.
- October 6, 2026. The Canadian publishers file against Anthropic.
The direction is steady: the same plaintiffs repeat the template against each new defendant, and each case adds a jurisdiction. For labs, that means the legal map is widening to national courts with different exceptions, not just US fair use.
Questions readers ask
Will Claude stop answering questions about Canadian news? Nothing in the reporting suggests any change to the product.
Could this end in a licensing deal? It could; deals are a common endgame in publisher disputes, though nothing in the CBC piece points to talks.
Does the claim target users? No. The defendant is Anthropic, and the claim concerns how it gathered content.
Why now? The reporting does not say. The suit was filed October 6 but surfaced publicly only on October 19.
What to watch next
- Anthropic's statement of defence, or a motion on jurisdiction.
- Whether other Canadian publishers, including CBC/Radio-Canada, join.
- Ottawa's plans on AI and copyright, flagged in CBC's coverage.
- Any settlement or licensing deals, which have been the usual endgame in other publisher cases.
- The OpenAI Ontario case, which will likely set the first procedural precedents.
Related reading
- Sony and Warner sue Anthropic over per-song damages
- USA Today sues OpenAI for $250 million
- AI lawsuit tracker: 248 cases, 76 companies
- Anthropic commits CAD 10M to Canadian AI research
- Meta commercial terms for crawling publisher sites
- AI bot traffic flood and robots.txt limits
Details are accurate as of October 20, 2026 and rely on CBC and Canadian Press reporting of the statement of claim. We have not seen the claim itself.