GenAI, Copyright and Intellectual Property

This page provides guidance for Toronto Metropolitan University (TMU) faculty, staff, and students on the appropriate and responsible use of Generative Artificial Intelligence (GenAI) tools in relation to copyright, intellectual property, and licensed electronic resources..

Please note that the terms of use for the TMU Gemini Responsible use guidelines explicitly discourages the uploading of any copyrighted content to the TMU Gemini instance. SEE: It isn’t appropriate to share copyrighted information nor the intellectual property of others with Gemini – including your students’ work and resources provided through the TMU Libraries.
Under current Canadian law, copyright protection can only be granted to a natural person and requires human authorship. As a result, content generated entirely by GenAI tools is generally not eligible for copyright protection in Canada.
Where human involvement is limited to entering prompts, the resulting output is unlikely to have any copyright protection.
GenAI outputs may not be unique and may be generated for multiple users.
Because GenAI-generated content is not protected by copyright, users should not claim copyright ownership over such outputs. For example, copyright notices or symbols should not be applied to images or text created solely using GenAI tools.

GenAI Tools and Terms of Use, Opt outs, and Privacy

Before using any GenAI tool, users must review and comply with the tool’s terms and conditions, including:

What types of content may be uploaded;
How uploaded content is stored, reused, or disclosed; and
Whether uploaded content may be used to train GenAI models.

TMU strongly recommends configuring GenAI tools to opt out of training or reuse of uploaded content wherever this option is available. For example, in ChatGPT accounts (including free accounts), users can disable model training under Settings/Data Controls/Improve the model for everyone. When this setting is turned off, content uploaded while logged in is not used to train the model. TMU’s enterprise Gemini environment already includes these protections.

Users should not upload confidential, personal, proprietary, unpublished, or sensitive information including student data, research data, or community-governed knowledge into GenAI tools.

GenAI and Fair Dealing in Canada

Fair Dealing is an exception in the Canadian Copyright Act which allows you to use other people’s material for the purpose of research, private study, criticism, review or news reporting, education, and satire or parody without having to seek permissions. Fair dealing is a long-standing and important user right under the Canadian Copyright Act. It allows a short excerpt of copyrighted protected works to be used without permission. Fair dealing is assessed case by case, taking into account factors such as the purpose of the use, the amount taken, and the nature of the work. The TMU Fair Dealing Guideline governs fair dealing at our institution for the use of material for teaching and research.

Legal requirements for attribution

Under the Canadian Copyright Act, when a fair dealing amount of a work is used for the purposes of criticism, review, or news reporting, the source and the author (where possible) must be clearly acknowledged. In addition, authors retain moral rights, including the right to be properly attributed and the right to the integrity of their work.

GenAI Model Training and Fair Dealing

Most large-scale GenAI models have been developed by copying and analyzing very large collections of text and data. Because Canada does not currently have a specific text and data mining exception in its copyright law, developers have generally relied on copyright exceptions such as fair dealing, or on licensed and publicly available content, to justify this activity.

There is ongoing legal and policy debate, both in Canada and internationally, about whether existing copyright exceptions, including fair dealing, are sufficient to support the training of GenAI models, particularly where models are developed or deployed for commercial purposes.

While fair dealing may support limited uses of copyrighted material by individuals for purposes such as research, private study, and education, it does not automatically extend to large-scale copying or model training, nor does it override licence terms or technological protection measures. If you plan to use copyrighted content to train a GenAI model, you may need to:

License the content or datasets; or
Conduct a thorough fair dealing analysis of the proposed use.

For guidance, contact the TMU Copyright Librarian Ann Ludbrook at aludbrook@torontomu.ca, who can assist or refer you to the appropriate university contact.

Technological Protection Measures (Digital Locks) and Fair Dealing

The Copyright Act contains provisions that restrict the circumvention of technological protection measures (TPMs), also known as digital locks. These restrictions apply even when the intended use might otherwise qualify as fair dealing.

When a work is protected by a TPM, users may be legally barred from accessing, copying, or reusing the material regardless of purpose. In addition, digital works are governed by licence agreements that may further restrict permitted uses. Together, TPMs and licence terms can significantly limit how fair dealing operates in digital environments.

Because of this, TMU staff and students should be cautious and avoid uploading copyrighted materials into GenAI tools unless they are clearly allowed to do so. Because GenAI tools store and process content in ways that are not always transparent, students and staff should use fair dealing cautiously and avoid uploading copyrighted content unless they are confident the use is permitted.

Fair Dealing and Print Materials

Print materials from sources such as physical books or journals are generally not subject to licence restrictions or digital locks, making fair dealing more straightforward to apply. Fair dealing will often apply to the use of short excerpts from a print works, such as:

A chapter from a book
A journal article
Selected pages or passages from a printed source not exceeding 10% of the work
An image, illustration, or map or graph, etc.

In these circumstances, fair dealing can support using limited portions of copyrighted print material with GenAI tools, for example, to summarize, explain, or assist with studying the material, provided the use is limited, non-commercial, and for personal academic or research purposes. The TMU Fair Dealing Guideline governs fair dealing at our institution for the use of material for teaching and research.

Fair Dealing and Works on the Internet that are publicly accessible

As long as there is not a clearly visible notice that prohibits reuse you can use a fair dealing amount of content from a public website under some conditions: .]

Please make sure the works you access are legally posted.
Please make sure that you do not use images that have watermarks or are from stock photo agencies.

Openly licensed content

Material that is found on a public webpage, including open access books, open access journals, images, charts, maps and illustrations that have Creative Commons licences applied to them can be used in their entirety.

Openly licensed content and Attribution

When using content licensed under Creative Commons or other open licences, users must follow the specific attribution requirements of the licence, including crediting the creator, providing the licence name, and linking to the licence if that is in the terms.

Using Electronic Library Resources with GenAI

In most cases, TMU Library electronic resources may not be used with GenAI tools. TMU Libraries license electronic content from vendors and publishers under contractual agreements that specify permitted uses. TMU users are required to comply with these licence terms.

Many licences explicitly prohibit:

Uploading licensed content into GenAI tools;
Using licensed content to train GenAI models; and
Automated, large-scale, or non-human uses of database content (such as the use of webcrawlers)

Some licences may permit the use by authorized users of fair dealing amounts (i.e., short excerpts) for allowable purposes such as education or research. These acceptable uses under fair use (the US term) or fair dealing are typically narrow, non-commercial, and require appropriate attribution. Users must review the terms of use for each database to understand the specific limitations that apply. See the TMU Libraries Electronic Resources Terms of Use. for further details.

Attribution and GenAI outputs

GenAI tools may paraphrase, summarize, or remix content without providing reliable citations. Users remain responsible for ensuring that appropriate attribution is provided where required, regardless of whether GenAI tools were used to assist with the work.

When using copyrighted material with GenAI tools, users should:

Keep track of the sources they input into the tool.
Clearly acknowledge original authors and sources where required by law.
Avoid presenting GenAI assisted paraphrases as original work when they are taken from identifiable sources.

What material can I use with GenAI Tools

We recommend using content that is openly licensed or in the public domain whenever possible, and also works that may qualify under fair dealing under certain circumstances (see above). According to the Creative Commons FAQs on AI you can use Creative Commons licensed works as long as the license conditions are followed. This means you will need to attribute each Creative Common licensed work according to the licence terms.

Before using content with GenAI, ask:

Is it openly licensed or public domain?
Is it subject to a licence or digital lock?
Does the GenAI tool allow me to opt out of LLM training?
Am I using only a short excerpt under fair dealing?
Can I provide appropriate attribution?

GenAI Upload Permissions for Faculty

Material Type	Can I upload this to a GenAI tool?
Material on public Webpages	Yes, in accordance with the TMU Fair Dealing Guideline
Print Journal Article, Print Book Chapter	Yes, in accordance with the TMU Fair Dealing Guideline
Print Image, Illustration, Map, or Graph; images on public Webpages	Yes, in accordance with the TMU Fair Dealing Guideline. Webpage images are subject to certain conditions.
Open Access material on public web	Can use whole work, please track and retain attribution information.
Library Electronic Book (Ebook) or Chapter	If protected by digital locks, in most cases, no, but there are exceptions. Please contact aludbrook@torontomu.ca for further details.
Library Electronic Journal Article	In most cases, no, but there are exceptions. Proquest One and some other journal databases have their own AI summary tools which are available to use with articles contained in the database. More information. Please contact aludbrook@torontomu.ca for further details.
Library licenced Digital Images, Illustrations, or Graphs	In most cases, no, but there are exceptions. Please contact aludbrook@torontomu.ca for further details.
Any Material Protected by Digital Locks (TPMs)	No
Licensed Library Databases/Datasets	In most cases, no, but there are exceptions. Please contact aludbrook@torontomu.ca for further details.

Important Requirements

Even when an upload is permitted (such as from a print source), the following Canadian legal requirements still apply:

Attribution: The source and author must be clearly acknowledged.
Moral Rights: The author’s right to the integrity of their work must be respected.
Purpose: Use must remain limited, non-commercial, and for personal academic or research purposes.

Note on Electronic Materials: Digital works are governed by specific license agreements and Technological Protection Measures (Digital Locks) that often override standard Fair Dealing exceptions. Always verify before uploading digital content.