A group of Canada's biggest news publishers has taken Anthropic to court. On October 19, 2026, The Canadian Press and other outlets reported that Postmedia, Torstar, The Globe and Mail and The Canadian Press had sued the maker of Claude in Ontario, alleging copyright infringement. The claim was filed on October 6 and first reported by the trade outlet Cartt before the wider coverage landed. (Source: The Globe and Mail; BlogTO)
Copyright suits against AI companies are no longer rare. What makes this one worth reading closely is how much of the claim is not about training at all. It targets scraping that allegedly continues every time a user asks Claude a question, and it leans on website terms and technical protections, not just copyright law.
Quick summary: Postmedia, Torstar, The Globe and Mail and The Canadian Press sued Anthropic in an Ontario court. The statement of claim was filed October 6, 2026, and reported widely on October 19. The publishers allege copyright infringement, unjust enrichment, breach of website terms of service, circumvention of technological protection measures, and ongoing scraping in response to user prompts. The reports I reviewed do not state a damages amount, and Anthropic had not commented. None of the allegations has been proven in court.
Who is suing, and what they say
The plaintiffs are Postmedia, Torstar, The Globe and Mail and The Canadian Press, a coalition that says it is responsible for a significant portion of Canada's journalistic content. They allege that Anthropic infringed their copyright by using their content to develop Claude. (Source: The Globe and Mail)
The claim, as quoted by the Canadian Press, says Anthropic "scrapes" text from the internet and from the plaintiffs' websites, and did so without a "valid licence" from any of the publishers. It also says Anthropic was "unjustly enriched" to the publishers' detriment. (Source: CP24; The Globe and Mail)
The five legal theories in one place
- Copyright infringement: using the publishers' content to develop Claude without permission.
- Unjust enrichment: the claim says Anthropic was enriched at the publishers' expense.
- Breach of terms of service: the publishers allege Anthropic broke the terms of their news websites.
- Circumvention of technological protection measures: they allege Anthropic got around tools meant to stop copying.
- Ongoing scraping: they allege content is still being scraped in response to ongoing user prompts, not only during past training.
The last item is the one to watch. Most earlier AI copyright fights centered on whether training on text is allowed. A claim that scraping happens live, at the moment a user asks a question, reaches a different part of how modern chatbots work, the part where a model retrieves current web content to answer. The reports I read do not explain the technical basis for that allegation, so treat it as an allegation only. (Source: BlogTO; The Globe and Mail)
What the reports did not say
Several details are missing from the coverage I opened. None of the reports I reviewed stated how much money the publishers want, and Anthropic had not responded to requests for comment. The OECD's AI Incidents Monitor lists the event as an "AI Incident," while noting its labels are AI-generated and not official OECD views. (Source: CP24; OECD AI Incidents Monitor)
Some secondary outlets circulate a per-work damages figure and a count of millions of articles. I could not confirm either against the claim or a primary report, so I am leaving them out. If you see those numbers elsewhere, ask for the document they come from.
How this fits with earlier cases
The same Canadian publishers, together with CBC, sued OpenAI in 2024, and that case has not settled. (Source: The Globe and Mail) In the United States, Anthropic agreed in 2025 to pay $1.5 billion, about $3,000 per book for an estimated 500,000 books, to settle a class action by authors who said it used pirated copies to train Claude. The judge in that case, William Alsup, found that Anthropic acquired more than 7 million digitized books it "knew had been pirated," but also ruled that training AI on copyrighted books was not illegal. (Source: Michigan Public / AP)
| Item | Bartz v. Anthropic (U.S.) | Canadian publishers v. Anthropic |
|---|---|---|
| Who sued | Book authors (class action) | Postmedia, Torstar, The Globe and Mail, The Canadian Press |
| Where | U.S. federal court, San Francisco | Ontario court |
| Content at issue | Books, including pirated copies | News articles on the publishers' websites |
| Core theory | Piracy of training material | Infringement, scraping, terms-of-service breach, circumvention of protections, unjust enrichment |
| Outcome so far | $1.5 billion settlement, about $3,000 per book | Allegations only; no ruling, no reported damages figure |
The comparison shows why this case is different. The U.S. case turned on how the books were obtained. The Canadian claim adds website terms and technical protections, and it reaches current-use scraping. Whether Canadian courts will accept those arguments is untested: the reports I reviewed do not say how a Canadian court will treat them. (Source: The Globe and Mail)
What to watch next
- Anthropic's formal response, which will show whether it contests jurisdiction or the merits.
- Whether the plaintiffs specify damages or seek an order limiting how Claude accesses their sites.
- Progress in the unresolved OpenAI case, which involves the same publishers plus CBC.
- Whether the federal government moves on AI and copyright, which could change the ground under both suits.
If you publish content online and care how AI systems use it, check your site's terms of service and your robots.txt rules, and keep a dated copy of both. Claims like the ones in this suit lean heavily on what a site's terms and technical protections said at the time.
For readers who simply use Claude, nothing changes today. A filed claim is not a ruling, and Claude remains available. But the case is a reminder that the rules for how AI tools gather and use journalism are still being written, country by country.
01Who filed the lawsuit against Anthropic?
A coalition of Canadian publishers: Postmedia, Torstar, The Globe and Mail and The Canadian Press. It was filed in Ontario on October 6, 2026 and reported widely on October 19.
02What do the publishers allege?
Copyright infringement, unjust enrichment, breach of website terms of service, circumvention of technological protection measures, and ongoing scraping in response to user prompts.
03How much money are they seeking?
The reports I reviewed do not state a damages amount. Treat any specific figure you see elsewhere as unconfirmed until it is tied to the claim itself.
04Has Anthropic responded?
The coverage says Anthropic had not responded to a request for comment. None of the allegations has been proven in court.
05How is this different from the $1.5 billion authors settlement?
The U.S. case centered on pirated books used for training. The Canadian claim covers news articles and adds website terms, technical protections and alleged ongoing scraping.
Legal fights like this one will shape which AI tools are available and how they behave. LumiChats gives you 40+ premium AI models side by side in one subscription, so you are never dependent on a single provider when the rules, or the headlines, change.
