Canada’s largest news publishers have sued Anthropic in Ontario, accusing the AI company of scraping millions of their articles to build and run Claude without a licence. The statement of claim was filed in the Ontario Superior Court of Justice on October 6, and The Canadian Press — itself one of the plaintiffs — confirmed the case on Monday.
The plaintiffs are Postmedia Network, Torstar, The Globe and Mail and The Canadian Press, joined by Metroland Media and PNI Maritimes LP. The claim names Anthropic and its Canadian subsidiary, and Broadcast Dialogue reports the publishers are seeking potentially billions in damages. Anthropic had not responded to requests for comment, and none of the allegations has been proven in court.
Claude’s Training Data: What The Publishers Say Was Scraped
The claim says Anthropic “deliberately scrapes” huge amounts of text from the open web and from the publishers’ own websites, and uses that proprietary content to develop its models “without authorization or consent”. It says the company obtained no licence “from any of the News Media Companies”.
The scraping is alleged to be ongoing rather than a one-off training run. Because Claude also draws on material when answering user prompts, the publishers say their content continues to be copied in the day-to-day operation of the product, not only when a new model is trained.
The publishers further argue that Claude repeats their work — verbatim, paraphrased or summarised — inside its answers. That, they say, draws readers away from their own sites, puts off advertisers and reduces the need for subscriptions.
The Four Claims: Copyright, Contract, Circumvention And Unjust Enrichment
The first claim is copyright infringement: copying the journalism to train and run Claude is said to breach Canada’s Copyright Act.
The second is breach of contract. Automated crawlers such as ClaudeBot are accused of ignoring the publishers’ website terms of use, which limit their content to personal use.
The third is circumvention: the claim says Anthropic got round the paywalls, logins and other technological protection measures — including network-level protections and login authentication — that were put in place to stop copying.
The fourth is unjust enrichment. “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,” the statement of claim says. It adds that Anthropic “has been substantially and unjustly enriched to the detriment of the News Media Companies”, and that the company “has done nothing to curtail its unlawful behaviour”.
Why It Matters
Anthropic is the central entity here, and the case tests three of its attributes at once: how Claude’s training data is gathered, how ClaudeBot behaves against publisher terms and paywalls, and whether the licensing market now forming around news content will be built by deals or by courts. The publishers have been pressing Ottawa on exactly this point after the federal government’s national AI strategy failed to address how AI companies use copyrighted content.
Canada is also a distinct legal test. Where US cases turn on fair use, this claim is built on the Copyright Act and on Canada’s rules against circumventing technological protection measures — ground we mapped in our guide to the copyright fights reshaping AI. A ruling against Anthropic would land on a company that this month committed 5% of its shares to a public benefit fund and gave METR researchers employee-level access to assess Claude’s risks — a safety-first public posture that sits awkwardly beside claims of systematic unlicensed scraping.
Background
This is the second time the same group has gone to court over AI training data: together with CBC/Radio-Canada, the publishers sued OpenAI in Ontario in November 2024, and that case has yet to be settled. The Anthropic claim extends the fight from one AI lab to the next, on the theory that the whole industry built its models the same way.
Source: MobileSyrup, reporting The Canadian Press; Broadcast Dialogue, reporting CARTT.ca
0 Comments