Provisional Recommendations for the Use of Generative AI

Teaching in the Context of AI

To:  Educators at the University of Guelph 

From:  Brian Husband, Acting Associate Vice-President (Academic) 

Sean Lyons, Assistant Vice-President, Graduate and Postdoctoral Studies 

Date:  September 3, 2026 

Subject: Teaching and learning with generative AI this fall 

Dear Colleagues, 

As we head into the fall semester, the use of generative AI in courses continues to prompt many questions from instructors. In support of both instructors and students, we are sharing some general guidelines, aimed at aligning AI use in course design and delivery with current best practices and associated university policies. These guidelines update the recommendations currently posted on the OTL website (Provisional Recommendations for the Use of Generative AI, 2023) and will continue to evolve with changes to institutional policies, AI models, and best practices.  

Top Four Guidelines: 

  1. Talk with your students are AI in the first week and for each assessment and learning activity. Explain how it will be used (or not) and why you are taking this approach in your course. 
  2. Substantive academic evaluations should never be made independently by AI: a grade is a human decision at this University and Senate-approved grading procedures require the exercise of discretion by instructors. Clear communication can reduce the likelihood of grade appeals related to assessment (Methods and Criteria or Misapplication of an Academic Regulation or Procedure). 
  3. Read your outline once more before the term starts and make sure you have explained your AI expectations for each assessment
  4. Ask for help. The Office of Teaching and Learning (otl@uoguelph.ca) takes any teaching question about AI and has programming this fall to support you; the School of Continuing Studies (scs@uoguelph.ca) takes questions related to DE courses and the McLaughlin Library offers resources on AI literacy (How the Library Can Help; library@uoguelph.ca). Information Governance and Compliance can support questions about privacy risks, and CCS for security risk assessment. 

Note that several colleges and departments have developed their own guidelines. Please inquire with your Associate Dean, Academic, or Chair/Director. 

Detailed Guidelines 

  1. Authority for AI use: As an instructor, you determine whether and how AI is incorporated into your courses. You may prohibit it, permit it in specified ways or for some tasks, or require it. However, be aware of how usage differs across courses within a program as this may influence students’ understanding of what is expected or appropriate. In graduate courses and other graduate learning contexts, expectations should also be considered alongside relevant research and supervisory expectations. 

  1. How AI is used: Ultimately, how you incorporate student use of AI should be guided by course learning outcomes. For example, are you aiming to improve AI awareness, promote critical and responsible use of AI, or prepare students to use AI in an authentic workplace setting? Depending on the learning outcome you may or may not choose to incorporate AI. Review Teaching and AI resources or Tools for Allowable Uses of AI or contact otl@uoguelph.ca for support.  

  1. Talk with your students: Two findings recur in most surveys of student AI use: use is now nearly universal, and students themselves are uneasy about what the tools are doing to their learning. Please remember that most students want to talk about AI and their learning. Take the time to explain your approach to AI and how those decisions support their learning. Note the library guide, Introduction to GenAI for Students, introduces generative AI, and helps students make responsible decisions for AI use. 

  1. Setting Expectations: Differences among courses in AI usage can be confusing for students. Please use your course outline, CourseLink site, as well as in-class conversations to review and explain the expectations for using AI in your course (Course Outline Statements on AI). Articulate your expectations for each assessment as to how students should demonstrate their learning and acknowledge AI use. You may be unsure or uncomfortable about talking with your students about AI. If this is the case, please reach out to OTL (otl@uoguelph.ca) or the Library (How the Library Can Help, library@uoguelph.ca).  

  1. Academic Integrity: Ensuring students are learning with integrity can be challenging in the age of AI. The calendar's existing academic misconduct provisions (Unauthorized Aids and Assistance) cover unauthorized use of AI. Being explicit about your expectations regarding AI use in individual assessments and having students declare sources and prompts for AI use can be helpful promoting a culture of integrity. Use of tools for detecting AI use are not approved at Guelph and cannot be used to support an allegation of a breach of academic integrity. Continue to talk with your students and consider ways of re-designing assessments that maximize learning while minimizing opportunities for integrity breaches. This question guide may be helpful for engaging in those conversations regarding suspected misuse of Generative AI. 

  1. Privacy and Data security: Student work that can identify, or reasonably be linked to, an individual student is personal information under Ontario's Freedom of Information and Protection of Privacy Act (FIPPA) and must be handled accordingly. Information Security's Artificial Intelligence Acceptable Use Guideline (May 2026) notes that identifiable student work is confidential information and cannot be shared in public AI tools (e.g. ChatGPT, Claude, and Gemini), whether the account is free or a paid personal one. Prompts in those tools have left the University's custody and may train someone's next model. If a teaching/assessment task needs an external tool, you should strip the prompt of any identifying information, ask Information Security to review the tool (if not previously reviewed; infosec@uoguelph.ca), or build the task so students enter their own work themselves (while offering them an opt-out).  

Microsoft Copilot, signed in with your University single sign-on, can be used for student information. Authenticated Copilot runs inside the University's Microsoft environment with enterprise data protection (Copilot: Get Started). Faculty and staff use various tools for other purposes, such as drafting, brainstorming, summarizing public material, and that is fine; but do not share student information with third-party tools. For more information, please review the Information Security guidelines for differentiation among data types.  

  1. Using AI for assessment: Substantive academic evaluations should never be made independently by AI: a grade is a human decision at this University and Senate-approved grading procedures require the exercise of discretion by instructors. Where an AI tool assesses student work, it is important to be transparent with students about how the tool is being used and the role it plays in the assessment process. Clear communication can help ensure that students understand how their work is being assessed and reduce the likelihood of grade appeals relating to assessment (Methods and Criteria or Misapplication of an Academic Regulation or Procedure). If you move toward secured formats for assessments (e.g. invigilated exams), be sure to consider students registered with Student Accessibility Services, and that for some students AI is assistive technology. 

  1. Use of 3rd party platforms. You may wish to have students use a 3rd party learning resource that uses AI. For platforms that are new to the university, confirm with the Office of Information Governance & Compliance, University Secretariat, that there are no privacy risks associated with the platform and CCS to ensure there are no security risks. Also, work with your College Dean or designate to ensure you are complying with the Ancillary Fees policy, which requires their use to be clearly communicated in the course outline, assessments conducted to constitute no more than 30% of the final grade, cost to be reasonable, and where feasible offer an alternative learning resource or grading scheme for students experiencing financial hardship.  

  1. Using AI for course preparation: Instructors may use AI tools for their own teaching preparation. Please double check AI products for accuracy and bias as you are accountable for your teaching materials and are modeling academic integrity for students. To consider copyright obligations in using AI tools, talk to colleagues in the Library (copyright@uoguelph.ca). If you are unfamiliar with the current capabilities and limitations of AI tools consider attending an OTL workshop or reach out to otl@uoguelph.ca to schedule an individual consultation.  

  1. Teaching assistants: Decide what constitutes acceptable use of AI by TAs in support of the course (e.g., developing class materials, grading) and discuss these expectations with your TAs. Ensure TAs understand how students will use AI so communication is consistent across the teaching team. 

  1. Distance education: Assessment formats that provide the most reliable measures of learning in the age of AI, such as supervised exams and oral exchanges, can present challenges for DE courses. The School of Continuing Studies provides a variety of supports to DE course instructors and academic departments in the design and delivery of assessments including: a) workshops and webinars (e.g., Best Practices for Online Assessments in the AI Era); b) individual support for DE course instructors; c) full administrative support of Off-Campus Final Exams for students studying away from the Guelph campus (per the Academic Calendar). Direct any questions or requests for support to Megan Pickard, Manager, Digital Learning Design and Innovation (meganp@uoguelph.ca). Questions specific to off-campus final exams can be directed to the SCS Exam Coordinator (scs-exams@uoguelph.ca). 

We appreciate that guidelines for AI use in the classroom can’t come soon enough. We hope this memo provides some clarity for this fall and trust you will find the support you need among your colleagues, the OTL, SCS, the Library and our offices. We will continue to work with campus partners and academic units to refine the guidance as the AI landscape changes. Please reach out with feedback and questions. Have a good start to the semester. 

Sincerely, 

Brian Husband 

Acting Associate Vice-President, Academic 

Sean Lyons 

Assistant Vice-President, Graduate and Postdoctoral Studies 

This document was drafted with much guidance from the Office of Teaching and Learning (OTL) and Office of Quality Assurance (OQA) and input from several support offices on campus, referred to herewithin. 

 

 

Teaching & Assessment Resources