The CIEL blog

Generative AI: citation and disclosure

In a previous post, we discussed making course-level AI statements that align with your learning outcomes. Once you have established when students can use generative AI (AI) tools, a next step is to help students acknowledge how they used them. This post considers several approaches, including citation, attribution, and disclosure. In developing guidance for this topic, we connected with colleagues at the Vancouver Island University (VIU) Library to draw on their expertise in citation practices and information literacy.

Depending on the assignment and learning goals, students may need to acknowledge AI through formal citations or disclosures that make their use of AI transparent. Practices are evolving, and there is no single approach that works in every context.

Citing generative AI

In some cases, students may need to formally cite generative AI tools. Because AI-generated outputs are unique to each session and non-retrievable by other readers, they are typically treated as non-recoverable sources and formatted according to specific style frameworks like APA, MLA, or Chicago. Students and instructors can refer to the following resources for current guidance on citing AI-generated content:

Formal citations are useful in some situations. However, it may not fully explain how AI supported a student’s work. Many instructors want to understand the learning process, not just the final product. They are interested in the student’s thinking, choices, and judgment when using AI. A disclosure statement can make this process more visible. It shows which tools were used, how they were used, and what steps were taken to check that the information was accurate and reliable.

Disclosing AI use

For assignments that involve collaboration between a student and an AI tool, disclosure practices can provide evidence of how students used AI to support their learning. These approaches help students be transparent about their use of AI while giving instructors insight into the student’s learning process. Consider including one or more of the following options in your assignment instruction.

AI tool use statements

Ask students to include a brief statement with their assignment explaining how they used AI. The statement should identify the tool used, describe how it supported the work, and explain the steps taken to verify information and correct any errors. Instructors may also choose to use an attribution template, framework, badge, or icon system to help students communicate AI use in a clear and consistent way.

Example

I used Microsoft Copilot to brainstorm potential topics for my assignment and organize my ideas into an outline. I also used it to identify areas where I could improve the clarity and flow of my writing. All content was reviewed and revised by me, and I verified information using course materials and library resources. Any inaccurate or incomplete suggestions were corrected before submission.

Disclosure frameworks and resources

  • The AI Attribution Toolkit is an open-source attribution framework that helps users document and communicate how AI was used in their work. Inspired by Creative Commons licensing, it provides a structured way to disclose AI contributions and clarify the relationship between AI-generated content and human authorship. 
  • AI Flags provides simple, voluntary AI content labels that help users communicate how AI contributed to their work. The framework offers a clear and accessible way to disclose AI involvement and support transparency. 
  • The Credtent Badge System provides a set of visual badges that can be used to communicate the extent of AI involvement in content creation. The badges support transparency by helping readers understand the relationship between AI-generated content and human contribution. 
  • Disclosure Statement Generators are tools that ask for information about how you used generative AI and then generate a disclosure statement that you can copy and paste into your work.

Prompt logs and transcripts

For assignments that involve significant AI use, ask students to include their prompts and the AI responses, or export their chat history. This helps make both the student’s process and the AI’s role more transparent.

Example

  • Tool and model used. ChatGPT (GPT-4o, OpenAI). Available at: https://chatgpt.com/
  • Purpose and scope. Generative AI was used during the preparatory phase to brainstorm counterarguments, organize draft outlines, and clarify core concepts.
  • Prompt submitted. “Act as a critical peer reviewer. What are three potential counterarguments to the thesis statement that digital taxation reduces foreign direct investment?”
  • Transcript and chat record. The full, unedited chat transcript is attached in Appendix A: AI Chat Transcript. Alternatively, the record can be viewed through the shared chat log.
  • Verification. All factual claims and historical dates generated by the AI were independently cross-referenced against peer-reviewed course readings before inclusion.

Process notes and reflections

Ask students to submit drafts, outlines, or reflections that show how their work developed over time.

Example

Example reflection activity

  • In approximately two hundred and fifty to five hundred words, reflect on your process for completing this assignment. Describe any tools, resources, or strategies that supported your work, including AI if applicable. Explain key decisions you made throughout the process, how you evaluated and verified information, and how your own analysis, judgment, and academic voice contributed to the final submission.

Assessment and feedback

  • Emphasize metacognition, factchecking, source verification practices, and demonstrate ownership of the final work.

Example reflection

  • Throughout the drafting process for my essay on digital service taxes, I used ChatGPT (GPT-4o) to help me brainstorm ideas and organize my arguments. One of my first prompts asked the tool to suggest counterarguments to my thesis. Two of the suggestions, about compliance burdens and international trade challenges, helped me strengthen my outline. However, another suggestion claimed that digital taxes always reduce foreign direct investment in developing countries. After comparing this claim with our course readings, I found that the issue was much more complex and depended on the country and policy context. As a result, I chose not to use the AI-generated statement and revised the section using evidence from class materials.
  • During the literature review stage, the AI also provided a citation that appeared relevant to my topic. When I searched for the source through the university library, I discovered that the article did not exist. This experience reminded me that AI-generated references must be verified before they are included in academic work. I removed the citation and replaced it with peer-reviewed sources that I located through the library database.

Version history and draft records

For writing or coding projects, ask students to submit a version history that shows how their work evolved. Techniques such as colour-coding, tagging or tracking can highlight where AI suggestions were used, changed, or rejected. This can help demonstrate their contributions throughout the project and show that they remained in control of the final product.

Example

Options for tracking in Microsoft Word

  • Turn on tracking before editing AI drafts so every insertion, deletion, and rewording is automatically recorded

  • Use comments (Ctrl + Alt + M) to add marginal notes to log exact prompts, record source verification steps, and explain why false/hallucinated AI claims were deleted.

  • Compare the raw AI file (Draft1_AI.docx) against the final paper (Draft2_Final.docx).

  • Use document tables to track workflow. Record prompts, key outputs, fact-checking activities, and reflections to document how generative AI contributed to the development of the work. See examples from the University of Waterloo.

Takeaways

There is no single approach that will work for every course or assignment. Choose a citation or disclosure approach that aligns with your learning outcomes and is manageable for both students and instructors. In some cases, a formal citation may be sufficient. In other contexts, a disclosure statement may provide more meaningful insight into how AI supported the work. Some instructors may choose to use both. The most important step is to clearly communicate your expectations and encourage students to explain how AI supported their learning. Multiple attribution frameworks and templates are available, giving instructors a range of options depending on the level of transparency and detail they want students to provide.

Attribution

Algonquin College Learning and Teaching Services. (n.d.). How to attribute use of generative AI. https://www.algonquincollege.com/lts/ai/how-to-attribute-use-of-generative-ai/

Anselmo, L., & Mikaeili, A. (2026, July 29). Exploring artificial intelligence and assessments. Taylor Institute for Teaching and Learning, University of Calgary. https://taylorinstitute.ucalgary.ca/resources/exploring-artificial-intelligence-and-assessment

Bradley University Library. (n.d.). Disclosing AI usage. Generative AI usage. https://bradley.libguides.com/genAI/disclosure

Centre for Teaching, Learning and Technology. (2025). Assessment design using generative AI. University of British Columbia. https://ai.ctlt.ubc.ca/assessment-design-using-generative-ai/

Firr, A., & Fenton, A. (2026, August 12). Here’s how to show students what responsible AI use looks like. Times Higher Education. https://www.timeshighereducation.com/campus/heres-how-show-students-what-responsible-ai-use-looks

He, J., Houde, S., & Weisz, J. D. (2025). Which contributions deserves credit? Perceptions of attribution in human-AI co-creation. In Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems (Article 540, pp. 1–18). Association for Computing Machinery. https://doi.org/10.1145/3706598.3713522

IBM Research. (2025). AI Attribution Toolkit. https://aiattribution.github.io

Monash University. (n.d.). Acknowledging the use of AI. Student Academic Success. https://www.monash.edu/student-academic-success/ai-hub/acknowledging-the-use-of-a

Nguyen, G. (2024). GenAI in teaching and learning toolkit: The least you need to know as educators. BCcampus. https://opentextbc.ca/teachingandlearningwithai

OpenED Manitoba. (n.d.). Supporting students to use AI effectively. In Generative artificial intelligence: Practical uses in education. https://pressbooks.openedmb.ca/aiineducation/chapter/supporting-students-to-use-ai-effectively/

University of British Columbia Library. (n.d.). Generative AI. APA citation style guide (7th ed.). https://guides.library.ubc.ca/apa-7/generative-ai

Writing and Communication Centre. (n.d.). AI and the writing process: Documenting and citing. University of Waterloo. https://uwaterloo.ca/writing-and-communication-centre/Resources-AI-Citing

Use of generative AI

Artificial intelligence tools: Gemini Notebook (Google) and Microsoft Copilot (GPT-5, Microsoft), used September 16 to October 09.

Conceptualization: Gemini Notebook (Google) was used to generate all examples of AI use disclosures. The generated examples were reviewed, adapted, and edited by the author to reflect the context and intent of this blog post.

Writing review and editing: Microsoft Copilot (GPT-5, Microsoft) was used to edit this blog post. I prompted Copilot to act as an expert blog editor and review the post through the lens of accessibility to help make the writing clear, concise, and understandable to a broad audience. I also asked Copilot to preserve my voice and intended message without introducing new content. AI-generated edits were reviewed, verified ,and further revised to remove unnecessary wording and ensure accuracy.

Disclosure preparation: The Artificial Intelligence Disclosure (AID) Builder was used to develop this disclosure statement.

Posted

in

,

by

Jessica Gemella