Canadian news publishers sue Anthropic in Ontario over AI scraping

The lawsuit

Postmedia, the Toronto Star, The Globe and Mail, Metroland Media, PNI Maritimes and The Canadian Press filed the claim in Ontario Superior Court on 6 October against Anthropic and its Canadian subsidiary. CARTT.ca first reported the filing. The publishers allege Anthropic scraped millions of their works to train and operate Claude without permission, licensing or payment, and that the scraping continues.

The claim pleads breach of contract, copyright infringement and unjust enrichment. It accuses Anthropic of "systematic harvesting and misappropriation" of the publishers' work, and says the company used automated crawlers, including one named ClaudeBot, to collect content for commercial purposes. According to the claim, Anthropic bypassed paywalls, personal-use restrictions, network-level protections and login authentications meant to block copying, and breached the websites' terms of service and circumvented their technological protection measures. It says Anthropic built an expansive suite of services and raised significant capital without a valid license from any of the publishers, and was "substantially and unjustly enriched."

The publishers say the scraping did not stop when training ended. The claim says Anthropic collects content on an ongoing basis in response to user prompts, and that despite other lawsuits and public criticism the company has done nothing to stop the conduct. It describes the outlets as generating "a significant portion of Canada's journalistic content," from daily print broadsheets to round-the-clock digital coverage, and says they collectively have millions of readers.

Scale of the claim

The Canadian Press accounts for the largest share of works allegedly scraped, at 4.2 million, Broadcast Dialogue reported. Postmedia and PNI Maritimes follow at 1.75 million each.

Damages and remedies

The publishers seek general damages, an accounting and disgorgement of profits, or statutory damages of up to $20,000 per infringed work, plus punitive and exemplary damages for what the claim calls willful conduct. Broadcast Dialogue put the total sought at potentially billions. The $20,000 figure is a statutory ceiling per work, so exposure scales with the number of articles the court counts.

They also seek permanent injunctions barring Anthropic from accessing, scraping or using their current and future works without written consent. That remedy would reach how Claude answers requests from now on, not only how it was trained.

The claim's harm theory is commercial. It says the copying diverts readers from the publishers' own sites, reduces paying subscribers and discourages advertisers from buying space in their papers. It ties that loss to Anthropic's balance sheet, stating that the company had raised approximately $125 billion USD in capital as of May and was valued above $965 billion USD, with more than a million users signing up for Claude each day. It also cites a US District Court for the Northern District of California finding that Anthropic downloaded more than seven million pirated books, followed by a $1.5 billion USD settlement.

On licensing, the claim says Anthropic, in contrast to OpenAI and nearly every other major AI company, has signed no public content deal with major publishers, and has treated litigation as a precondition for paying creators.

Context

The same companies, with CBC/Radio-Canada, sued OpenAI in 2024; that case is unsettled. The Canadian Press reported that the New York Times has sued OpenAI over its use of copyrighted material and that the US Justice Department sided with the tech company in September. The publishers have also pressed Ottawa for rules on how AI companies use copyrighted content, after the national AI strategy did not address the issue.

What happens next

The case is pending and none of the allegations has been proven in court. Anthropic did not respond to a request for comment.