SI Agent News
DevelopingOpenAIPolicy

Ontario court rules OpenAI must face copyright suit in Canada

The ruling confirms that Canadian courts have jurisdiction over the AI giant, despite OpenAI's attempts to move the case to the United States.

10.65 millionworks published by the plaintiffs since 2015

The short version

Why it matters

This sets a precedent for how foreign AI companies can be held liable in Canadian courts for copyright infringement. It highlights the legal risk for AI developers operating in jurisdictions with stricter copyright protections than the United States.

The jurisdiction battle

Ontario's Court of Appeal ruled on November 13 that OpenAI must remain in a Canadian courtroom to face a copyright lawsuit. The lawsuit was brought by five major news organizations, including The Canadian Press and CBC/Radio-Canada. These publishers allege that OpenAI scraped and copied their content to train the models powering ChatGPT.

OpenAI had argued that the legal dispute should be heard in a U.S. court instead. However, the appellate court dismissed this appeal, upholding a previous decision by Justice Kimmel. The court found that Ontario has a real and substantial connection to the alleged breaches, which justifies local jurisdiction.

As a result of the ruling, OpenAI was ordered to pay the publishers $100,000 in costs. While the currency was not specified, Ontario courts typically award such costs in Canadian dollars. This decision ensures that the legal battle will proceed under Canadian law rather than American statutes.

Canada vs. U.S. legal rules

The venue is significant because Canadian copyright law differs from the U.S. fair use doctrine. In the United States, fair use is an open-ended test that can accommodate new technologies. Canada, however, uses a 'fair dealing' exception, which is much narrower and requires a use to fit into specific categories like research, private study, or education.

Furthermore, Canada currently lacks a specific text-and-data-mining exception for AI training. This means that even if a U.S. court eventually finds that training AI on copyrighted data is legal, that ruling would not bind an Ontario court. The Canadian court will assess the copying based on its own specific legal framework.

OpenAI expressed concern that a contrary result in Ontario would be problematic if U.S. courts find the training lawful. The publishers argue that their content is largely created in Ontario and owned by Ontario-based companies. They contend that the scraping of this content occurred within the province, making Canadian law the primary authority.

The scope of the claim

The coalition of publishers is seeking statutory damages for copyright infringement, unjust enrichment, and breach of contract. They are aiming for damages of up to $20,000 per work, which is the ceiling set by the Copyright Act for commercial infringements. The publishers have published a combined total of 10.65 million works since 2015.

The lawsuit also seeks an injunction and OpenAI's profits from the alleged infringement. While the allegations are currently unproven, the publishers argue that OpenAI profits from their content without providing permission or compensation. OpenAI has maintained that its models are trained on publicly available data and that it works with publishers on how content is displayed.

The court has not yet decided whether the copying actually happened in Canada. OpenAI argues that its training and automated crawling happened entirely outside of Ontario. If OpenAI can prove the reproductions occurred only on servers outside the country, the jurisdiction win might not lead to a final liability judgment.

Broader implications for AI

This case is the first of its kind in Canada to address whether using copyrighted content to train generative AI is lawful. The ruling provides a roadmap for other Ontario-based suits against foreign AI companies. For example, a separate suit was filed in May 2026 by VerticalScope regarding the scraping of its online forums.

The decision also impacts other companies, such as Anthropic, which is facing a similar copyright suit from several of the same publishers. Because the Ontario court established a consistent approach to asserting jurisdiction over U.S. companies, other defendants may find it harder to move these cases out of Canada. This creates a more predictable environment for Canadian creators seeking legal recourse.

The federal government's stance on AI copyright remains unsettled. While some officials suggest that creators should be compensated, the specific rules for training data are still being debated in Parliament. This legal vacuum means that the courts will likely play a primary role in defining the boundaries of AI development in Canada for the foreseeable future.

What comes next

OpenAI still has the option to apply for leave to appeal this decision to the Supreme Court of Canada. This application must be filed within 60 days of the judgment, excluding the month of July. The Supreme Court generally grants leave only for issues of public importance or those that warrant a high-level decision.

If the case proceeds to trial, the court will enter the discovery phase. This will involve examining whether the court can compel OpenAI to produce de-identified ChatGPT logs, similar to orders seen in U.S. litigation. Such evidence would be critical for the publishers to prove exactly what content was copied during the training process.

The outcome of this case will be watched closely by the global AI industry. It will help determine how much risk foreign AI developers face when they scrape content from international news sites. Any final ruling on the merits will establish a major precedent for the intersection of generative AI and intellectual property rights in Canada.

The path forward

The legal battle is far from over, as the court has only decided where the case will be heard, not whether any laws were broken. The publishers will now move to prove that the copying occurred in Ontario and that it constitutes an infringement under the Canadian Copyright Act.

The discrepancy between Canadian and U.S. law creates a complex landscape for AI companies. They must navigate different standards for 'fair' use depending on where the data is sourced and where the legal challenge originates. This highlights the growing importance of territorial copyright law in the age of global AI training.

For now, the publishers have secured a significant procedural victory. They can now pursue their claims for damages and injunctions in a local court that recognizes their specific legal protections. The next several months will likely involve intense evidence gathering regarding the technical specifics of how OpenAI's models were trained.

The legal landscape

Lawyers have noted that this ruling follows a trend of Canadian courts asserting jurisdiction over U.S. tech giants. The decision relies on factors like where a business is carried on, where the wrong occurred, and where contracts were formed. By finding that OpenAI entities carry on business in Ontario, the court has reinforced the reach of Canadian law.

The absence of a text-and-data-mining exception in Canada makes the legal environment more challenging for AI developers than in the European Union. This lack of clarity is a point of contention for industry groups who argue that the current gap is costly. They are calling for clearer licensing mandates or limits on fair dealing to provide certainty for both creators and developers.

As the case moves forward, the court will have to balance the interests of news organizations with the technological goals of AI companies. The final judgment will likely hinge on the geographic location of the data processing and the specific interpretation of 'fair dealing' for AI training. This case will serve as a landmark for how Canadian law adapts to the era of generative artificial intelligence.

Summary of key points

The Ontario Court of Appeal has confirmed that OpenAI must defend a copyright lawsuit in a Canadian court. This is a major step for the five news organizations seeking damages for the use of their content in training AI models. The ruling ensures that the case will be judged under Canadian law, which is stricter than the U.S. fair use doctrine.

The publishers are seeking up to $20,000 per work for a collection of over 10 million works. While the court hasn't ruled on the infringement yet, it has established that Ontario has a substantial connection to the dispute. This means OpenAI cannot easily move the litigation to a more favorable jurisdiction in the United States.

The outcome of this case will have significant implications for the AI industry and the rights of content creators. It highlights the ongoing tension between technological innovation and the protection of intellectual property. Readers should watch for further updates as the case moves into the discovery phase and toward a potential trial in Ontario.

Next steps for the case

The next phase of the litigation will involve discovery, where both sides must share evidence. The publishers will look to prove that the scraping of their content happened in Ontario and that it violated their terms of use. OpenAI will likely try to demonstrate that the training happened on servers outside of Canada.

The court will also have to decide how to interpret 'fair dealing' in the context of machine learning. Since there is no specific exception for AI training in Canada, the court must decide if training falls under existing categories like research or private study. This will be a landmark interpretation of the Canadian Copyright Act.

The publishers are also seeking an injunction, which could potentially limit how OpenAI uses certain types of content in the future. A win for the publishers could lead to a new licensing model for AI training. For now, the case remains a pivotal battle over the legal foundations of the generative AI economy.

The role of government

The Canadian government has yet to finalize a clear policy on AI and copyright. While some officials have suggested that creators should be compensated, the specific rules are still being determined by the courts. This lack of clear legislation means that court rulings like this one will set the de facto rules for the industry.

The government's recent AI strategy has been criticized by some for not mentioning copyright specifically. This omission has led to calls from news media groups for clearer licensing mandates. They argue that without clear rules, the value of creative work is being undermined by unauthorized AI training.

Meanwhile, other countries are taking different approaches, such as Australia's plan to introduce AI copyright legislation in 2027. Canada's path remains uncertain, making the outcome of the OpenAI lawsuit even more critical. The court's decision will provide a significant piece of the puzzle in how Canada balances innovation with creator rights.

Impact on AI developers

For AI developers, this ruling signals that they cannot rely solely on U.S. legal protections when operating in Canada. They must be mindful of the specific 'fair dealing' requirements and the lack of a text-and-data-mining exception. This could increase the legal costs and complexity of training models on international data.

Companies may need to seek explicit licenses from news organizations to avoid costly litigation in Canadian courts. The fact that the court found a 'real and substantial connection' to Ontario based on business activities and advertising suggests a broad reach for local laws. This sets a precedent that could affect many other tech companies with a presence in the province.

The case also highlights the importance of understanding the territorial nature of copyright. Even if a company is based in the U.S., its actions involving Canadian content can lead to Canadian legal consequences. AI firms will need to carefully consider where their data processing and business activities are legally situated to mitigate risk.

Summary of the legal status

The Ontario Court of Appeal has dismissed OpenAI's appeal to move the copyright suit to the U.S. This means the case will proceed in Ontario under Canadian copyright law. The publishers are seeking damages for the use of 10.65 million works in training ChatGPT's models.

The court found that Ontario has a substantial connection to the dispute, including OpenAI's business operations in the province. This ruling is significant because Canada's 'fair dealing' rules are narrower than the U.S. 'fair use' doctrine. There is also no specific exception in Canada for text-and-data-mining for AI training.

While the court hasn't decided if OpenAI infringed the copyright, this jurisdictional win is a major step for the publishers. The case will now move into discovery, where the specific details of the training process will be examined. The final outcome will help define the legal boundaries for AI development in Canada and beyond.

Why this matters for the reader

This case determines whether AI companies can use your news, books, and creative works to train their models without paying you or asking for permission. If the publishers win, it could lead to a world where AI companies must pay for the data they use, potentially changing the cost and availability of AI tools.

For those who create content, this is a major test of whether copyright laws can actually protect human work from being ingested by machines. It highlights the ongoing struggle to balance the benefits of new technology with the rights of the people who create the information that powers it.

For users of AI, the outcome could affect the types of data these models are allowed to learn from. If certain types of content become off-limits due to copyright costs, it could change the capabilities of tools like ChatGPT. This case is a key piece of the puzzle in shaping the future of the internet and AI.

Key facts of the ruling

The Ontario Court of Appeal ruled on November 13 that OpenAI must face a copyright lawsuit in an Ontario courtroom. The lawsuit was brought by five news organizations, including The Canadian Press and CBC/Radio-Canada. They allege that OpenAI scraped their content to train its AI models.

OpenAI tried to move the case to the United States, but the court dismissed the appeal. The court found that Ontario has a 'real and substantial connection' to the alleged breaches. This means the case will be decided under Canadian law, which is different from U.S. law.

Canada's 'fair dealing' rules are narrower than the U.S. 'fair use' doctrine. Additionally, Canada does not have a specific exception for text-and-data-mining for AI training. The publishers are seeking damages of up to $20,000 per work for a total of 10.65 million works published since 2015.

The technical side of the dispute

The core of the dispute involves 'scraping,' which is the automated collection of data from websites. The publishers argue that OpenAI's crawlers moved through their sites to collect content for training purposes. They claim this was done without permission and in violation of their terms of use.

OpenAI argues that its training process involves using publicly available data to create 'transformative' models. They rely on the idea that the resulting AI is a new product, not just a copy of the original content. However, the Canadian court will look at whether the specific act of copying the data was permitted under Canadian law.

A key technical question will be whether the copying happened on servers located in Canada. If OpenAI can prove the data was processed entirely outside the country, it might avoid liability. The publishers, however, argue that because the content is created and owned in Ontario, the infringement also occurred there.

What this means for the news industry

For news organizations, this ruling is a significant victory in the fight for copyright protection. It allows them to seek damages and injunctions in their home jurisdiction. This could lead to a new era where news publishers have more leverage to negotiate licensing deals with AI companies.

The publishers argue that OpenAI is profiting from their content without providing any compensation. They want to ensure that the value of their reporting is protected from being used to train competing technologies. This case could set a standard for how other media companies approach AI licensing.

If the publishers succeed, it could change the business model for news in the AI age. It might force AI companies to create more transparent and equitable systems for sharing data. For now, the news organizations are using this lawsuit to establish their rights and seek a fair share of the profits generated by AI.

The future of AI regulation

This court case is happening while the Canadian government is still deciding on a formal AI copyright policy. The government has expressed a general interest in compensating creators, but the specific rules are still being debated. This means the courts will likely be the ones to set the first major precedents.

Some lawmakers have suggested that the government should create a text-and-data-mining exception, similar to what the European Union has done. Others believe that the current copyright laws are sufficient and do not need to be reopened. This lack of consensus makes the outcome of the OpenAI lawsuit even more important for the industry.

The ruling also highlights the global nature of AI and the complexities of international law. As AI companies continue to grow, they will face a patchwork of different rules in different countries. This case will be a key study in how different legal systems handle the challenges posed by generative artificial intelligence.

Summary of the legal arguments

The publishers base their

Sources and links

Liked it? Say why

Add a line for other readers.

No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.

Listen to the AI Wire headlines

Listen to the AI Wire

Listen to our daily AI shows every morning: on your commute, in the shower, wherever you are

Listen to the daily shows

First reported by Tech Times · Read the original. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.

More AI news