How is AI being used?
The use of AI content is increasing in academia. The availability and ease of use of generative AI tools have led to questions about copyright. This page will provide a brief synopsis of AI, its risks and recommended strategies related to copyright.
Generative AI can help in the creation or adaptation of instructional content, but there is legal uncertainty when it comes to using these tools. As this topic evolves, information on this page will change over time.
AI and copyright involve:
Input: AI content is generated through large amounts of data. These systems are "trained" to copy content without permission from the rights holders. This has resulted in several lawsuits due to copyright infringement. For education, it remains unclear how fair dealing can be applied since this topic is evolving.
Output: The author of a work is not well-defined under the Canadian Copyright Act. Generally, an "author" refers to a person who created the work or put it in a fixed form. The ownership of content created through AI is unclear as there are varying degrees of human input, making it difficult to determine the author and owner of the work.
What to Consider if Bringing AI into the Classroom
Exercise caution when using AI tools in the creation of course content. This topic is still developing in the Canadian courts and through policy development. Concrete guidelines are emerging.
While Generative AI tools are helpful, there are ethical risks and concerns to consider when using these resources in an academic environment:
Lack of transparency: AI tools are trained on data, but the programming logic is not always available to the public. Further, companies that develop these tools may not explain how they work or how their tools arrived at the output they did.
Content source: With a lack of transparency around programming logic, it is unclear whether generated content is infringing on copyrighted material. Currently, there are lawsuits claiming that generative AI tools are creating unauthorized derivatives.
Developing case law: Canadian legal rulings around AI content is developing. Shared international copyright and trade agreements continue to inform how generative AI impacts intellectual property.
Although this aspect of copyright is emerging, here are some best practices:
- Use caution: AI generated content is an area that is in flux and guidelines and recommendations are emerging.
- Be transparent: As with other areas of attribution, include a statement letting others know that you have used generative AI. This could include: what was created, what tools were used (including a link), how the tool was used, date of generation and steps used to review the content for accuracy or suitability.
- Be aware of changes: Legal consensus is developing and fair dealing may not be appropriate if creating an unauthorized derivative through generative AI tools.
- Be mindful of inputs: Do not use content that may include corporate logos or patents. If using AI to create a summary of another work, be sure that you're doing so fairly. Never input significant portions of intellectual property. You cannot use full-text library-licensed resources. Mohawk Library signs license agreements with vendors and publishers. These terms and conditions define user responsibilities and permitted terms of use, including generative AI. To avoid these risks, consider using public domain and openly licensed content.
- Review Terms of Use: These terms will dictate ownership and may be subject to change. While these tools aid in content creation, you may not necessarily hold or own the copyright on generated works.