Canada court lets publishers sue OpenAI over AI training

The Ontario Court of Appeal has dismissed OpenAI's appeal against a ruling that Canada's largest news publishers can sue the company in Ontario, clearing the way for the country's first copyright case over the use of journalism to train artificial intelligence systems. The appeal was dismissed on Friday, 13 November 2026, by a unanimous three-judge panel led by Justice B.W. Miller.
The publishers — Toronto Star Newspapers, Metroland Media, Postmedia Network, The Globe and Mail, The Canadian Press and CBC/Radio-Canada — filed their lawsuit in November 2024. They allege that OpenAI copied large amounts of their journalism without permission and used it to train the models behind ChatGPT, profiting from that use without paying for it. The allegations have not been proven in court.
OpenAI had argued that the case should be heard in the United States, saying the alleged breaches fell outside the reach of Canadian law and that the claim had been served on its Delaware-based companies outside Ontario. The Court of Appeal rejected that argument, finding that Ontario has a real and substantial connection to the alleged breaches and upholding an earlier ruling by Justice Jessica Kimmel of the Ontario Superior Court of Justice.
What the publishers allege
The publishers are suing for breaches of Canada's Copyright Act, breach of contract through their websites' terms of use, and unjust enrichment. They say OpenAI reproduced their journalistic works without permission, used them in datasets to train its GPT models, and exploited the resulting systems commercially. They are seeking damages, disgorgement of profits, punitive damages and a permanent injunction, with alternative statutory damages of C$20,000 per work.
The court also noted that OpenAI had filed no evidence showing that a judgment in Canada would conflict with any existing United States law in a way that would complicate enforcement. The judge found that the fair and efficient working of the Canadian legal system favoured allowing Canadian authors to bring claims against foreign companies for breaches of Canadian law.
What the ruling does not decide
Friday's ruling decides only where the case can be heard, not whether OpenAI infringed anyone's copyright. No court has yet ruled on that question, and the case will now return to the Ontario Superior Court for a hearing on the merits — the stage at which the court will hear whether OpenAI's use of the publishers' work actually breached Canadian copyright law. No date has been set for that hearing.
The appeal decision, cited as 2026 ONCA 814, upholds Justice Kimmel's ruling of 7 November 2025 that the Ontario court had jurisdiction. The Court of Appeal ordered OpenAI to pay $100,000 in legal costs to the publisher group.
Why the case matters
The lawsuit is the first Canadian case to address the practice of using copyrighted content to train generative AI systems, and its outcome is likely to shape how AI companies deal with Canadian publishers. Similar lawsuits are already under way in the United States, some dating back to 2023, including a case by publishers alleging that the Canadian AI company Cohere infringed their copyright by scraping their articles.
Earlier this year, Postmedia, Torstar, The Globe and Mail and The Canadian Press launched a separate lawsuit against the AI company Anthropic. News and cultural industry groups have called on the federal government to take a stand on how AI systems use copyrighted content, but Ottawa has not indicated any plans to update the Copyright Act.
For now, the publishers have won the right to make their case at home. Whether they can prove that training ChatGPT on their work broke the law is a question for the trial ahead. Sources: Broadcast Dialogue and The Canadian Press (via Winnipeg Free Press).