tech

Canadian news publishers sue Anthropic over use of journalism to train AI

Postmedia, Torstar, The Globe and Mail and The Canadian Press allege their work was used without permission to build the Claude model. The allegations have not been proven in court.

The Canadian

Newsroom

October 23, 2026

3 min read

AI illustration of stacked newspapers beside a laptop showing abstract network lines, with a balance scale behind, no people present
AI-generated editorial illustration; not a photograph of the reported event.

Four of Canada’s largest news organizations have sued Anthropic in Ontario, alleging the artificial intelligence company used their journalism without permission to develop its Claude model. The Canadian Press reported on October 19 that the plaintiffs are Postmedia, Torstar, The Globe and Mail and The Canadian Press itself. The suit was filed on October 6, and the allegations have not been proven in court.

The case is part of a widening global legal fight over how AI companies obtain the text used to train large language models. News organizations argue that their reporting — expensive to produce and protected by copyright — has been absorbed into commercial products without consent or payment. AI companies have generally argued that training on publicly available text is lawful, and some have signed licensing deals with publishers instead.

The Canadian lawsuit is significant because it tests those arguments in an Ontario court, under Canadian copyright law, rather than in the United States where most similar cases have been filed. A ruling here could shape how Canadian courts treat the fair dealing defences that technology companies are expected to raise, and could influence whether Ottawa feels pressure to legislate on AI training data.

For the publishers, the stakes are existential. News outlets have watched advertising revenue migrate to digital platforms for two decades, and they now face a generation of AI products that can summarize and answer questions using journalism without sending readers to the original source. Licensing revenue from AI companies has become a potential lifeline — which is why the terms of use matter as much as the principle.

Anthropic, founded by former OpenAI researchers and backed by major technology investors, has positioned itself as a safety-conscious AI developer and has struck content agreements with some publishers. The Canadian plaintiffs will have to prove both that their works were used and that the use was infringing. Whatever the outcome, the case ensures that the question of who pays for the raw material of AI will be argued in a Canadian courtroom.

techReportingCanada

Up next

Keep readingOttawa tables MAID bill that permanently bars mental-illness-only cases and opens door to advance requests