Assessing Student Learning

Principles of Good Assessment Design

• Prioritize principles of good assessment design, focusing on validity, transparency, and feasibility. 
  • Validity: Do the assessments measure the knowledge, skills, and competencies they are intended to measure, at the appropriate level, in alignment with the intended learning outcomes?  Will student learning be evident from how they complete the assessment? 
  • Transparency: Are the expectations, purpose(s), processes for completing the assessment, and marking criteria, clearly communicated? 
  • Feasibility: Are the assessments practical in terms of time, resources, and student numbers? Do students have the pre-requisite skills or do you need to take time to teach those? 
• To learn more: Read this Teaching Tip: Qualities of Effective Assessment and visit this website 2026 Assessment Redesign Framework (Hazel Farrell). 

Responding to the Impact of AI in Assessment (Re)design

• Reflect on your existing assessment structure and the extent to which AI may undermine the validity of your assessments or obscure the evidence of student learning. Focus on rethinking unsupervised assessments such as written essays, online discussion forums, and online auto-graded quizzes.
  • To learn more: Read this Teaching Tip: Redesigning Unsupervised Assessments in the GenAI Era.

• Identify assessments with compromised validity. Reflect on any patterns or concerns you have noticed in student grades and ask your TAs for their observations of student learning throughout the grading process.  

• Determine assessments to change. Assessments that fail to provide meaningful evidence of the intended learning outcomes are good candidates for redesign. Balance feedback from students and TAs with your capacity to rethink changes and anticipated workload (including additional grading considerations and TA training).

  • To learn more: Read Course and Curriculum Design

• Consider whether there are any AI-related competencies students should develop related to your course topic/discipline and intended learning outcomes. Explore opportunities for thoughtful integration.

  • To learn more: Read Integrating AI Use in Assessments below.

• Coordinate with colleagues within your program(s) to ensure that students experience a variety of assessment formats. (E.g., Are programs relying too heavily on oral assessments? Do students have opportunities to practice long-form writing?)

Important note: Remember that the assessment (re)design process is complex and continually evolving. Adopt a mindset that allows you to compromise, diverge, and iterate (Corbin et al., 2025). There is no single or uniform solution to an assessment (re)design in response to AI, and it is important to continue refining assessments based on your experiences as well as student and TA feedback.  

Mitigating AI Use in Assessments

• Revisit the assessment breakdown and ensure that most of the course grade comes from supervised assessments such as proctored tests and exams, live oral assessments (e.g., problem-solving discussions, case study analyses), poster presentations, and direct observations of process or procedures (e.g., clinical observations, fine arts studio sessions). 
  • To learn more: Read this Teaching Tip: Implementing Oral Assessments.
• Incorporate scaffolded assessments throughout the course to make learning visible and measure process alongside (or over) product. Focus on assessment design that shows evidence of students’ decision-making.
  • To learn more: Read this Teaching Tip: Redesigning Unsupervised Assessments in the GenAI Era.

• Incorporate reflection and authentic application of knowledge in discipline-specific contexts.
  • To learn more: Read these Teaching Tips: Using Authentic Assessments to Surface Learning,  Promoting and Assessing Critical Thinking, and Strategies for Addressing AI in Common Assessment Types. 
• Use student-instructor agreements that ask students to verify that they will not use (or have not used) AI in their assessed work.

• Recognize that AI detection tools are unreliable. They continue to produce false positives and negatives and cannot be relied upon to provide definitive evidence of academic misconduct.

Important note: The University of Waterloo has discontinued the use of AI detection in Turnitin: See the Discontinuing use of AI detection functionality in Turnitin announcement (AVPA, Sept 2025) 

• To learn more: Read these Teaching Tips: Encouraging Academic Integrity in Remote, Online, and In-person Learning Contexts and General Assessment Design Strategies.

Integrating AI Use in Assessments

• Provide clear parameters for where and when AI use is permitted in your assessments. In some courses, limited use may work best (e.g., have students use AI for some assessments or parts of assessments but not for others), while other courses may want to fully integrate AI into one or more assessments. Regardless of the degree of integration, be explicit about how students should use AI, paying specific attention to nuances and interpretation (e.g., define potentially ambiguous terms like “brainstorming” so that students understand what specific activities this refers to). 

Important note: Remember that it may not be necessary to fully incorporate AI into each assessment, as the degree of AI integration should always align with the course learning outcomes and activities.

• Identify the AI tool(s) that students are permitted to use in your assessments. AI tools have different capabilities and qualities of output, so be specific about the permitted tools and ensure you have factored time to learn the tool into your course planning.
  • To learn more: See Getting Started information regarding AI Tools for Learning

• Acknowledge different viewpoints on using AI, and unless AI is tied to course learning outcomes or is required in your course to address program outcomes, consider whether it is feasible to offer students with objections an option to complete assignments and activities without using AI (e.g., by generating an AI output yourself for students to analyze, or by transferring the weight to the final exam or other assessment).

• Demonstrate how students should document AI use in the context of the assessment. Provide examples of correct attribution or citation and convey the process that you and your TAs will take to review this component of the assessment.

• To learn more: Read The Artificial Intelligence Disclosure (AID) Framework and AI and the Writing Process – Documenting and Citing.
• Prioritize human judgment by modelling how students should interrogate AI outputs. Demonstrate how AI might be used in the context of your discipline or assessment and ask students to articulate what AI did well in addition to areas of concern.

• Highlight bias, harms, and inaccuracies inherent to AI by having students interrogate AI outputs. Asking students to think critically about the outputs helps strengthen critical AI literacy and ensures humans remain at the center of decision-making.

• Consider new assessments you could design that are built around applications of AI within your discipline.

• To learn more: Read this Teaching Tip: Integrating GenAI in Assessments and  Strategies for Addressing AI in Common Assessment Types.

Communicating Expectations

• Promote academic integrity by drafting a clear AI use policy for the course and for individual assessments. Ensure the type of use is clearly stated in your course outline and in your assessment descriptions.
  • To learn more: Read this Teaching Tip: Conversations with Students About GenAI Tools and refer to the UW Course Outline Suggestions for Generative Artificial Intelligence, which are boilerplate options in Waterloo’s Course Outline tool. Personalize policy statements to suit your course and aim to be as clear as possible since students are likely to encounter varying GenAI policies across their courses.

• Ensure that your course outline/course-level policy does not conflict with your assessment policies. For example, if your course outline says, “forbidden unless otherwise specified” and nothing is ever specified, your policy in effect prohibits AI use, but students may then search for the exceptions and be confused.

Communicating With Students

• Dedicate class time to explain how AI has shaped your assessment (re)designs and invite students to ask clarifying questions.
  • To learn more: Read this Teaching Tip: Conversations with Students about GenAI Tools and explore Talking to Students About Their Learning and GenAI: A Conversation Guide (MacPherson Institute)
• Explain the importance of following the AI use policy, implications for academic integrity, and negative consequences on student learning if students choose to disregard the policy.

Communicating With Teaching Assistants (TAs)

• Review the AI use policy for the course and individual assessments with your TAs. Describe your rationale for the policy.
• Explain how your TAs should approach the grading process for assessments that mitigate, permit, and/or integrate AI use.

• Be clear with your TAs about how they are allowed to use AI, or not, within their TA duties.
• Reserve time for your TAs to ask questions or express concerns. Suggest GenAI resources to help them learn and build confidence in their role.
• Share the Teaching Assistant Checklist: Questions to Ask resource with your TAs for suggested questions to ask about AI use during their onboarding meeting.