On 13 May 2025, the Prime Minister of Canada, Mark Carney, announced the members of his new Cabinet, including the new Minister of Artificial Intelligence and Digital Innovation, Evan Solomon. An individual mandate letter, which would provide details on his role and responsibilities, has not been released. However, a mandate letter sent to all the ministers notes the “transformative nature of artificial intelligence (AI)” and states that the government must become much more productive, including by “deploying AI at scale.”
During October 2025, the government is holding a public consultation to “[h]elp define the next chapter of Canada’s AI leadership.” The government’s stated goal is to develop a new AI strategy by the end of 2025 in light of the priorities identified during the consultation. It is worth noting that the government’s report following the Consultation on Copyright in the Age of Generative Artificial Intelligence, held between October 2023 and January 2024, was published in February 2025.
Minister Solomon said he plans to address the issue of AI and copyright in his future AI strategy [in French], which is expected to focus on protecting cultural sovereignty and the role of creators. While he does not plan to introduce any bills in the House of Commons at this time, the minister noted that he is monitoring ongoing court cases and market developments.
One such case, Toronto Star Newspapers Limited v. OpenAI Inc., involves a coalition of media outlets suing OpenAI for using copyrighted content to train its ChatGPT chatbot. OpenAI has denied these allegations, saying that its models are “grounded in fair use and related international copyright principles that are fair for creators and support innovation.” The company is also challenging the jurisdiction of the Ontario Superior Court to hear the case.
In two recent decisions in the United States (U.S.), Bartz et al. v. Anthropic and Kadrey v. Meta Platforms, Inc., it was determined that the unauthorized use of copyrighted works to train AI models constituted fair use. However, it is important to note that the U.S. concept of “fair use” differs from the Canadian concept of “fair dealing.”
These cases show that AI-created works challenge certain fundamental assumptions about copyright. As AI increasingly contributes to visual, literary, musical and other works, there is a growing risk of confusion over what part of a work is created by a human being and what part is created by AI. This presents a challenge for the application of certain concepts, such as the originality of a work, authorship and the ownership of copyrights. Further, the training of AI models raises questions about whether text and data mining should be added to the exceptions provided for in the Copyright Act.
Originality of a Work, Authorship and Ownership of Copyrights
To be protected by the Copyright Act, a work must be original. In CCH Canadian Ltd. v. Law Society of Upper Canada (CCH), the Supreme Court of Canada clarified that, to be original, a work “must be more than a mere copy of another work” without needing to be “creative, in the sense of being novel or unique.” According to the Court, “[w]hat is required to attract copyright protection in the expression of an idea is an exercise of skill and judgment.” In addition, the exercise of this skill and judgment “must not be so trivial that it could be characterized as a purely mechanical exercise.” Since AI has the ability to demonstrate what appears to be skill and judgment, applying this criterion is becoming problematic.
Under Canadian copyright law, the author of a work must be a human being, and ownership of the copyright to this work stems from the authorship status granted to that person. While it is easy to determine the author of an algorithm, which is protected by the Copyright Act as a “computer program” equivalent to a literary work, the author of an AI-created work is less obvious. Is the AI programmer or provider the author, or is it the user? Are they considered co-authors of a work of joint authorship, which the Copyright Act defines as “a work produced by the collaboration of two or more authors in which the contribution of one author is not distinct from the contribution of the other author or authors”?
Copyright Exception for Text and Data Mining
The algorithms used to create AI-generated works learn automatically after being “trained” on large volumes of data. When this data includes protected works or other protected subject matter, the use of this data – as well as the distribution of the works created by AI as a result – may infringe copyright. Experts and lawmakers in other jurisdictions have proposed various solutions to protect rights holders while encouraging AI development, such as adding an exception for text and data mining in copyright legislation.
Article 4 of the European Union (EU) directive on copyright and related rights in the digital single market provides for an exception or limitation to copyright for text and data mining. This exception or limitation applies on condition that the use of works and other protected subject matter has not been expressly reserved by their rights holders, “in an appropriate manner, such as machine-readable means in the case of content made publicly available online.”
Recital 105 of the EU regulation on artificial intelligence, which came into force on 1 August 2024 (with most provisions becoming fully applicable two years after entry into force), specifies that “[w]here the rights to opt out [have] been expressly reserved in an appropriate manner, providers of general-purpose AI models need to obtain an authorisation from rightsholders if they want to carry out text and data mining over such works.”
In addition, article 53(1)(c) of the regulation specifies that general-purpose AI model providers must put in place a policy to identify and comply with a reservation of rights expressed by rights holders in accordance with article 4(3) of the directive.
To increase transparency regarding the data used in training general-purpose AI models, article 53(1)(d) of the regulation requires providers of these models to draw up and make publicly available a sufficiently detailed summary of the content used to train their models.
The EU approach runs up against certain obstacles. For example, unless they are aware that their work could be used, copyright owners may not know that they have the option to reserve their rights, which could prevent them from holding general-purpose AI model providers accountable and subsequently asserting their rights effectively.
Sections 29, 29.1 and 29.2 of the Copyright Act set out the fair dealing exceptions that apply to a work or other protected subject matter: private study, research, education, parody, satire, criticism, review and news reporting.
Adding a new fair dealing exception to the Copyright Act applicable to text and data mining would have to take into account the six factors set out in the CCH decision to determine whether a dealing is fair:
- the purpose of the dealing;
- the character of the dealing;
- the amount of the dealing;
- available alternatives to the dealing;
- the nature of the work; and
- the effect of the dealing on the work.
Next Review of the Copyright Act
Bill C-27, which proposed enacting a new Artificial Intelligence and Data Act (AIDA), among other matters, died on the Order Paper when Parliament was prorogued on 6 January 2025. The proposed AIDA did not deal with copyright, which was the subject of some criticism during parliamentary scrutiny of the bill. A government representative stated that AIDA was intended to be a law of general application, adding that the most effective way to address copyright concerns over AI is through the Copyright Act. To date, no legislation has been introduced incorporating AI into the Copyright Act.
Section 92 of the Copyright Act requires a parliamentary review of the Act five years after the day on which this provision came into force, which was in 2012, and every five years thereafter. The House of Commons Standing Committee on Industry and Technology carried out that review and presented a report in June 2019, which recommended providing clarity around the ownership of a computer-generated work. If this date is used as the starting point for the next five-year period, the next parliamentary review would have occurred in June 2024.
Rights of Indigenous Peoples
The United Nations Declaration on the Rights of Indigenous Peoples Act came into force in 2021. It requires that the Government of Canada, in consultation and cooperation with Indigenous peoples, take all measures necessary to ensure that the Act – like all federal statutes – is consistent with the United Nations Declaration on the Rights of Indigenous Peoples. The implementation of the right of Indigenous peoples to maintain, control, protect and develop their intellectual property is provided for in article 31 of the Declaration. A future review of the Act could also examine whether the Act is consistent with this provision.
The government report on copyright in the age of generative AI, published in February 2025, mentioned certain issues associated with respecting the rights of Indigenous peoples. In addition, the 2019 report on the parliamentary review of the Copyright Act contained recommendations aimed at protecting traditional Indigenous art forms and cultural expressions in Canadian legislation.
By comparison, it is worth noting that several aspects of New Zealand legislation discourage the registration of intellectual property that could be contrary to Māori values.
Further Reading
Azzaria, Georges. “Droit d’auteur et intelligence artificielle : un aperçu canadien.” In Céline Castets-Renard and Jessica Eynard, eds. In Un droit de l’intelligence artificielle – Entre règles sectorielles et régime général – Perspectives comparées, 2023.
Craig, Carys J. “AI and Copyright.” In Florian Martin-Bariteau and Teresa Scassa, eds. Artificial Intelligence and the Law in Canada, 2021.
Jonnaert, Caroline. Intelligence artificielle et droit d’auteur – Le dilemme canadien, 2024.
By Maxime-Olivier Thibodeau, Library of Parliament