by Stuart Collins, Jennifer Harrison, Kevin Muir, and Dr. Kate Scarpena
As institutions shift responsibility for defining acceptable generative artificial intelligence (AI) use to individual instructors, faculty need practical, fair systems for managing it. This post presents the AI Use Log, a transparency-based disclosure tool in which students name the AI tool, record prompts, and explain how they verified or revised the output. It describes a dual-pathway model, first used in Purdue Global’s Digital Citizenship course (SS255), in which students choose an AI-Assisted Pathway, documented with a log and screenshots, or an equally legitimate Traditional Pathway. Practical design tips follow, along with common student misunderstandings that make early instruction on the log essential.
Generative AI has moved past the point where instructors can simply ban it and move on. Most institutions have shifted responsibility for defining acceptable use to individual instructors, which means the burden of building a workable, fair system now sits with the individuals teaching the course. One of the effective tools available for that job is deceptively simple: the AI Use Log.
Reframing the instructor-student relationship through a transparency-based disclosure model has been effective. In this model, students document how and why they used AI on a given assignment, naming the tool, sharing the prompt, and explaining how they changed or built on the output. This shifts the burden of disclosure onto the student and doubles as a lesson in ethical use.
A tool that supports the practice: A minimal, field-tested template
The simplest version of an AI Use Log asks students to record seven things: the tool used, the date, the assignment, the purpose of the AI use, the key prompts entered, what parts of the output were actually used, and what human checking or rewriting followed. The habit of keeping a short record while working, rather than reconstructing an explanation after the fact, is what makes disclosure credible.
The pattern is the same: Name the tool, capture the prompt, and record what the student did with the output afterward. That third element, the human verification step, is what separates a disclosure log from a permission slip.
The Dual-Pathway model
One structure worth sharing with colleagues is a dual-pathway framework, where students choose at the outset between an AI-Assisted Pathway and a Traditional Pathway for a given assignment or initial discussion post. Students who select the AI-Assisted Pathway commit to mandatory disclosure and documentation of their AI interactions; students who select the Traditional Pathway complete the work without AI assistance and document that choice instead. The value of this structure is that it does not treat AI use as a binary, either permitted or forbidden; rather, students choose their preferred approach (AI-assisted or traditional).
How it works
At the point an assignment or discussion initial post is given, each student chooses one of two pathways:
The AI-Assisted Pathway. The student may use AI tools to support the assignment, for brainstorming, outlining, feedback on a draft, or similar support, provided the use is disclosed. Disclosure has two required components: a completed AI Use Log entry and a screenshot of the AI interaction itself. The screenshot requirement matters more than it might at first appear; a written log alone is easy to reconstruct after the fact and summarize favorably, whereas a screenshot is a contemporaneous record of the actual prompt and output.
The Traditional Pathway. The student completes the assignment without AI assistance and documents that choice instead of an AI Use Log. This pathway is not a fallback or a default for students without access to AI tools. Instead, it is presented as an equally legitimate methodological choice, similar to how a research course might let a student choose a qualitative or quantitative approach.
Consistency across course materials
The dual-pathway language appears everywhere a student encounters the assignment, not only in a single policy document: the syllabus, the assignment instructions, the rubric, and the instructor feedback all reference the same two pathways by the same names. Students at Purdue Global may use Gemini Notebook or Gemini; the disclosure requirement is about the interaction, not the vendor. This consistency is what makes the framework legible rather than a one-time announcement students forget by Unit 4.
A Non-AI Option Is Not a Consolation Prize
When the Dual Pathway model was initially integrated into Purdue Global’s Digital Citizenship course (SS255), the faculty members teaching the course often had to catch themselves in terms of how to frame the Traditional Pathway; it is not the “old-fashioned” approach, or the choice for students who are not as tech-savvy. That framing does real damage.
Some of the most capable writers and careful thinkers choose the Traditional Pathway every single time, not because they cannot use AI, but because they do not want to for that particular assignment. Let that be a genuine choice. The moment a non-AI option starts to feel like a fallback, students stop choosing it for the right reasons (Garcia Ramos, 2025).
The dual-pathway model does not eliminate the need for an instructor to define what counts as acceptable AI use within the AI-Assisted Pathway; it establishes the disclosure mechanism, not the content boundary. Instructors adopting it still need to decide and clearly state what AI is allowed to do (brainstorm, outline, give feedback) versus what it is not allowed to do (generate final submitted prose) within that pathway.
Tips for building your own AI Use Log
For faculty ready to design a version for their own course(s), a few practical pointers:
- Keep the log short. A few lines per use is enough. A log that feels bureaucratic will not get filled out honestly.
- Require the prompt, not just the tool name. “Used Gemini” tells an instructor almost nothing. The actual prompt, even abbreviated, shows the nature of the assistance.
- Make the verification step explicit. Ask students what they checked, corrected, or rejected from the AI output. This is where genuine learning shows up, and it is the strongest evidence that a student engaged critically with the tool rather than simply accepting its output.
- Consider requiring a screenshot alongside the written log. A screenshot is harder to reconstruct after the fact than a written summary, and it gives the instructor a direct view of the actual prompt and output rather than the student’s characterization of them.
- Attach the log to the assignment, not the course. A single end-of-term disclosure is too vague to be useful. Logs tied to specific assignments produce a record an instructor can actually act on.
- Offer a non-AI pathway, and make it equally legitimate. Present the choice as methodological rather than remedial, so students who decline AI assistance are not treated as though they are missing out or falling behind.
- Use tool-agnostic language throughout. Ask students to name whichever AI platform they used rather than assuming a specific one. This keeps the template usable across institutions with different tool access and across international settings with different data-privacy regimes.
An AI Use Log will not solve every question generative AI raises for higher education. But it does something detection tools cannot: it turns AI use into a documented, reflective part of the learning process, and it gives instructors a much better window into how their students are conceptualizing, rather than only whether they cheated (Jin et al., 2025).
To be frank, the AI Use Log is a tough sell for some students. Some misunderstand what counts as “using AI,” while others log a basic spell-check pass as though it were a collaborator. Others use AI to outline an entire argument and do not think that counts as disclosure-worthy, because to them, the actual sentences were still their own (Du & He, 2026). For courses with the AI use log, therefore, it’s vital that faculty show students how to use the log early in the course.
References
Du, Y., & He, H. (2026). Enabling and inhibitory pathways of university students’ willingness to disclose AI use: A cognition-affect-conation perspective (Version 1) [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2604.21733
Garcia Ramos, J. (2025). Development and introduction of a document disclosing AI use: Exploring self-reported student rationales for artificial intelligence use in coursework: A brief research report. Frontiers in Education, 10, Article 1654805. https://doi.org/10.3389/feduc.2025.1654805
Google. (2026). Gemini (Gemini 2.5) [Large language model]. https://gemini.google.com/
Jin, Y., Yan, L., Echeverria, V., Gašević, D., & Martinez-Maldonado, R. (2025). Generative AI in higher education: A global perspective of institutional adoption policies and guidelines. Computers and Education: Artificial Intelligence, 8, Article 100348. https://doi.org/10.1016/j.caeai.2024.100348
About the Authors
Stuart Collins is a full-time faculty member in Humanities and Social Sciences, teaching American Government, Digital Citizenship, and Sustainable and Resilient Communities. He serves on Purdue Global’s AI Taskforce and the university’s Strategic Foresight Committee. He lives in Evergreen, Colorado, and is president of the Evergreen Mountain Area Historical Society.
Jennifer Harrison is a full-time faculty member in the Humanities & Social Sciences department and course lead for American Women and Digital Citizenship. She specializes in American women’s history and civics, with research interests involving the use of AI tools and the presentation of women in literature, film, and television.
Kevin Muir is a full-time Purdue Global faculty member with over 17 years at the institution and college teaching experience dating back to 1999. As course lead for Citizen-Driven Public Policy and Colonial and Revolutionary America, he manages curricular design, focusing on student engagement and AI Use in higher education.
Kate Scarpena serves as Department of Humanities and Social Sciences, fostering lifelong learning through interdisciplinary study of human societies and culture. A former K–12 social studies teacher and history professor, her expertise spans higher education administration, curriculum design, program development, and faculty development.
Leave a Reply