By Julius Melnitzer | May 11, 2026
“In Canada, fair dealing is a creature of statute with a closed list of eligible purposes. Courts have described it as a ‘users’ right’ that provides a ‘fsafety valve’ on the monopoly of copyright.” — Tamara Céline Winegust
Canada, Australia, New Zealand, Hong Kong, China and most of Southeast Asia have their own variations of the copyright law doctrine of “fair use” that prevails in the United States.
This is the second of a series of articles that elucidates the jurisdictional approaches in Australia and Canada.
As it turns out, nothing less than a philosophical divide is at the core of the US doctrine of “fair use” and the Canadian doctrine of “fair dealing”.
“In Canada, fair dealing is a creature of statute with a closed list of eligible purposes. Courts in Canada have repeatedly framed it as a ‘users’ right’ that provides a safety valve on the monopoly of copyright,” says Tamara Céline Winegust, Principal in the Toronto office of Smart & Biggar, a member of the IPH network. “By contrast, ‘fair use’ is a common law doctrine codified into the US Copyright Act. U.S. Courts tend to frame ‘fair use’ as a defense to infringement and an ‘exception’ to owners’ exclusive copyrights.”
Under Canada’s Copyright Act, purposes eligible for fair dealing are research, private study, education, parody, satire, criticism, review, and news reporting.
“The Supreme Court of Canada has made it clear that this is a closed universe of purposes, and only dealings that fall within that list qualify for consideration as to whether they are ‘fair’”, Winegust says.
By way of example, Winegust cites a 1990s decision that denied a fair dealing claim based on parody because “parody” wasn’t, at that time, one of the enumerated exceptions.
“But since 2012 amendments to the Copyright Act added ‘parody’ to the list, it now qualifies,” Winegust says.
If a dealing falls into one of the enumerated purposes, courts move on to the second step of the test and assess fairness by way of an overall approach using the following criteria as a framework: the dealing’s purpose, its character, the amount, the alternatives, the nature of the work, and the effect of the dealing on the work.
The brunt of judicial analysis falls on the second stage.
“The Supreme Court of Canada has mandated an approach that takes a ‘large and liberal interpretation’ in determining whether a dealing falls under the enumerated categories, with the conceptual heavy hitting taking place when determining fairness.
In the US, fair use is a defence to infringement claims as opposed to a standalone users’ right.
“There’s no closed universe in the US, as any kind of use could theoretically qualify,” Winegust says. “It’s a straight ‘fairness’ test”.
Another difference is that American courts have at times placed a heavy emphasis on the concept of “transformative use” (i.e., whether a use adds something new—like a new purpose, message, or character—to an original work, rather than substituting for it) as part of the consideration of the purpose and character of the use; however, Canadian Courts have not historically recognized the “transformativeness” of a dealing as an explicit or necessary factor in the analysis.
Still, the distinctions between the American approach and the Canadian approach go beyond the merely theoretical.
“Despite some similarities, the differences between fair use and fair dealing have the potential to produce divergent outcomes,” Winegust says.
It is an open question whether this could become the case in relation to the controversy surrounding artificial intelligence and copyright.
In June 2025, the United States District Court for the Northern District of California found, on a summary judgment motion, that the AI developer Anthropic’s practice of purchasing legal copies of books, then digitizing and using them to train its large language model, CLAUDE, amounted to fair use and did not infringe copyright (the ruling did not apply to Anthropic’s use of illicit digital copies, which the court found to be infringing and not excused).
It’s not clear, however, where Canadian law will land with respect to AI in the copyright context.
“In the intellectual property space, the Supreme Court of Canada has made decisions with outcomes both consistent with and contrary to outcomes in the U.S. ” Winegust says.
The most famous example of divergent outcomes in IP cases is the Harvard Mouse case, which involved an “oncomouse” that was genetically engineered. (Harvard College v. Canada (Commissioner of Patents), 2002 SCC 76)
There, the U.S. Patent and Trademark Office granted a patent on the mouse itself—not just the genetic engineering method and the modified gene. The Supreme Court of Canada, however, rejected the patentability of higher life forms.
In the AI and copyright space, it remains to be seen whether Canada will apply reasoning that produces a similar outcome to that in Anthropic.
“There could be a split similar to the one in Harvard Mouse when it comes to dealing with AI,” Winegust says. “The outcome will be driven largely by the particular facts and equities of the case. Current jurisprudence suggests it could go either way.”
Julius Melnitzer is a Toronto-based writer who focuses on law, legal affairs, and the business of law. Follow him on LegalWriter.net or email him at julius@legalwriter.net.