Exploring Artificial Intelligence and Assessments

A guide for instructors considering artificial intelligence in their assessments.

Authors: Lorelei Anselmo, MEd; Ali Mikaeili, EdD (Student)
*Adapted from the original resource developed by: Lorelei Anselmo, MEd & Tyson Kendon, PhD

Last modified: July 29, 2026

Realizing the potential of generative AI in higher education means moving beyond automation and reimagining how teachers, students, and AI can learn and work together.

Qian, 2025

This short guide provides some context on Generative Artificial Intelligence for instructors, as well as some suggestions for ethically addressing artificial intelligence in teaching and learning. There are also some risks and limitations you may need to consider as well as some guiding questions and approaches which may help mitigate those risks and limitations.

A First Response to Assessment and ChatGPT in your Courses

Teaching and Learning with Artificial Intelligence Apps

The good, the bad and the artificial: Bringing artificial intelligence into the classroom

Artificial Intelligence

A catch-all term for techniques to help computers solve problems the way people would or to model computational problem solving on biological brains. (Russell & Norvig, 2020)

Machine Learning

A subfield of AI in which systems learn patterns from data and improve their performance on a task without being explicitly programmed with rules for every case. Machine learning underlies most modern generative AI tools.

Generative Artificial Intelligence

Techniques to allow AI systems to produce works based on the data they have been trained on. These can be text, images, translations, sound, video or more. Other types of AI systems might do things like classify inputs. (Chan & Hu, 2023)

Large Language Model

A type of AI system that generates text based on the probability of words in its training data and the text that it has already generated. LLMs can produce fluent and convincing text, but they do not understand truth, context, or course expectations the way a person does. (Kasneci et al., 2023)

Hallucination

False, fabricated, unsupported, or misleading AI output presented confidently. AI-generated text can sound authoritative even when it is inaccurate. (Ji et al., 2023)

Prompt engineering

The practice of designing prompts to shape the quality, format, tone, or relevance of AI output. Prompting can be treated as a learning skill when students are also asked to evaluate, revise, and verify the output. (White et al., 2023)

Multimodal AI

AI tools that can process or generate more than one type of content (text, images, audio, video, files). Matters for teaching because AI use now extends beyond written text.

Agentic AI / AI agents

AI systems that can plan and complete multi-step tasks with limited human prompting. Raises questions about what counts as student work and how assistance should be documented. Instructors are also using agentic tools themselves, for example to build course-tutor bots or automate routine tasks, so this is an opportunity to explore as well as a risk to manage.

Ambient AI

AI that operates continuously in the background of a device or environment, often without an explicit prompt, such as live captioning, writing suggestions, or meeting summarization tools built into everyday software. Because it runs by default, students and instructors may use it without deciding to or knowing.

Foundation and frontier models

Foundation models underlie many AI tools; frontier models refer to the most advanced models at a given time. Different products may share similar capabilities, limitations, and risks. (Bommasani et al., 2021) For example, tools such as ChatGPT and Gemini are both built on foundation models, which is part of why they can behave in similar ways even though they are made by different companies.

Informed Use:
Before using or requiring an AI tool, instructors and students should understand how it works, what it can and cannot do technically, and what risks or limitations it presents.

Transparent Use:
When AI is used in course work, students should identify the tool, how it was used, and what role it played in the submitted work. Instructors should be equally transparent about their own use of AI in preparing course materials, feedback, or grading.

Ethical Use:
Students should understand the difference between using AI to support learning and using AI to replace required student thinking, with what counts as acceptable assistance made clear at the course and assignment level. Instructors are equally responsible for clarifying these expectations for their own course or assignment.

Responsible Use:
Responsible use includes checking AI-generated content for accuracy and verifying claims against course materials, since detection tools alone are unreliable and academic judgment remains essential Both students and instructors remain responsible for any AI-generated content they use, regardless of whether detection tools flag it.

Adapted from Gutiérrez, J. (2023) Guidelines for the Use of Artificial Intelligence in University Courses. Version 4.3, Universidad del Rosario. https://forogpp.files.wordpress.com/2023/02/guidelines-for-the-use-of-artificial-intelligence-in-university-courses-v4.3.pdf

AI systems may not fit every teaching context or assessment purpose. When considering using generative AI in your teaching, here are some things you may need to consider. These include understanding context and nuance, accuracy and bias, privacy, feedback quality, interactivity, and over-reliance on technology.

Lack of understanding context and nuance

Generative AI might not be trained on the specific context related to your course. Outputs may be too general, oversimplified, or misaligned with what students are expected to demonstrate.

Inaccuracies and bias in information

Generative AI can produce hallucinations, fabricated citations, and confident-sounding errors, and may reproduce biases in its training data, leading to discriminatory output.

Lack of feedback and guidance

AI feedback may be generic, inaccurate, or misaligned with course expectations. It is most useful as a starting point for review and revision, not as a replacement for instructor, TA, or peer feedback.

Privacy and data security concerns

Information entered into public AI tools may be stored, retained, or used for model training depending on the tool's terms. Do not ask students to enter identifiable, confidential, or graded work into public AI tools unless the tool is institutionally approved. At the University of Calgary, privacy obligations fall under Alberta's Protection of Privacy Act (POPA; https://www.alberta.ca/protection-of-privacy-act) and Access to Information Act (ATIA; https://www.alberta.ca/access-to-information-act), which replaced FOIP in June 2025. This also applies to instructors and teaching assistants, who should not enter identifiable student information or their own confidential materials into these tools either.

Limited interactivity and discussion

Generative AI can simulate conversation and support brainstorming or surface-level analysis, but it cannot reliably replace adaptive academic discussion with an instructor, TA, or peer group. AI dialogue is most valuable when students critique, test, or revise it.

Overreliance on technology

An overreliance of Generative AI might reduce a learners’ ability to develop transferable, creativity and critical thinking skills. The risk is greatest when AI is used as an unguided shortcut This may also mean students are not achieving the intended learning outcomes of the course or program, which raises its own ethical concerns.

Use clear guidelines and expectations

Provide students with clear guidelines about the role of Generative AI in your course and assessments. Make it clear what Generative AI can and cannot do and specify how it could be used as a resource, emphasizing that Generative AI is a tool to aid understanding not a replacement for critical thinking.

Maintain human grading and review

While Generative AI can assist in certain aspects of student work, human graders are responsible for evaluating student work, providing context-specific feedback. Final grading decisions should remain the responsibility of the instructor.

Provide feedback on AI-generated responses

When Generative AI responses are used in assessments, offer feedback on the quality and accuracy of those responses. This helps students understand the limitations of AI-generated content and encourages them to critically evaluate the information provided.

Address bias and fairness

Include discussion on the ethical use of AI and encourage students to critically assess the information provided by AI to consider its accuracy, bias, and implications for different communities and contexts.

Discuss how AI-assisted assignments will be assessed

(for assignments where AI use is permitted or encouraged)
Demonstrate to students how AI-generated responses will be evaluated including how they should disclose the extent to which AI was used and how they cross-referenced AI-generated information with reliable sources. The Academic Integrity Reflection Tool offers a starting template for this kind of disclosure.

Mix AI-assisted and non-AI assisted questions

Design assessments that incorporate a mix of questions suited for AI-assistance and questions that require deep analytical skills. This allows learners to practice with GAI while still allowing for the development of critical thinking and creativity skills.

Provide balanced resources

Ensure that students have access to a variety of resources, including institutionally-supported AI tools, library guides, scholarly sources, student supports, and workshops and other materials so students learn to make informed decisions about when to use Generative AI and when to rely on other sources.

Generative AI can support teaching and learning in many ways when its role is clear and students are expected to evaluate, verify, or reflect on its output. The categories below are starting points — adapt them to your discipline, course, and learning outcomes. These are options for instructors to consider, not requirements. Where instructors do use AI in these ways, they should be transparent with students about how and why, for example by noting AI assistance in course materials, assignment instructions, or feedback.

For course design and preparation

Lesson and module planning: Draft outlines, learning activities, or pacing guides, then refine against your course outcomes.
Reading guides: Produce comprehension questions, glossaries, or summaries to accompany course readings.
Case scenarios: Generate discipline-specific cases, vignettes, or problem sets students can analyze.
Accessibility supports: Produce plain-language summaries, alternative explanations, or text-to-audio versions of course materials.

Question banks: Draft multiple-choice, short-answer, or scenario-based questions, then review for accuracy and alignment. For example, instructors might write the correct answers themselves and ask GenAI to generate plausible incorrect distractors.
Sample exemplars: Produce strong, weak, or partial sample responses for students to compare and critique.
Authentic tasks: Brainstorm context-rich, discipline-specific tasks that are harder to outsource to AI.
Staged assessments: Design proposal, draft, revision, and reflection components that make process visible.

For student learning activities

Brainstorming: Students generate initial ideas and explain which they accepted, rejected, or developed.
Draft and revise: Students produce an AI draft, then document what they verified, changed, or corrected.
AI-output critique: Students evaluate AI output against rubrics and identify errors, bias, or fabricated citations.
Concept explanation: Students ask AI to explain a concept multiple ways, then compare with course readings.

For feedback (with human review)

Drafting feedback comments: Generate first-pass feedback language for instructor revision before sharing with students. Before doing so, ensure no identifiable student information is entered into the tool. Some students may have concerns about their work being used this way even when anonymized (Luo, 2025).
Expanding short notes: Turn brief instructor annotations into clearer student-facing comments.
Feedback consistency: Draft rubric-aligned comment banks for teaching teams to adapt.
Translation and accessibility: Reformat feedback for plain language, translation, or audio versions.

AI should not make grading decisions; human judgment remains central to all grading decisions. Regardless of what tools or methods are used to calculate or draft a grade, the instructor of record remains solely responsible for the final grade assigned. If an AI tool produces an incorrect, incomplete, or biased grade or piece of feedback, that responsibility falls to the instructor to catch and correct, not to the tool. Using AI for grading or feedback is further complicated by privacy and intellectual property: uploading student work to a third-party AI tool means sharing that work, and the student’s intellectual property, without their consent, and the tool may retain that content with no student oversight. Prefer AI tools that have been institutionally reviewed and approved for this purpose (for example, Gradescope is an approved campus grading platform) over uploading student work to general-purpose public tools such as ChatGPT.

Set written expectations

State permitted use, disclosure, and verification expectations in the syllabus and on each assignment.

Model AI use in class

Demonstrate prompting, compare strong and weak outputs, and show verification steps.

Discuss privacy

Cover what should not be entered into public AI tools and where institutional guidance applies.

Provide alternatives

Offer pathways for students who cannot or do not wish to use a specific tool. This should be considered from the start of assessment design, not as an afterthought, with clear alternative pathways for both individual and group work.

These questions can be used individually or as a team discussion guide before finalizing an assessment.

What learning outcomes should this assessment measure, and would AI use support or interfere with them?
Is AI use prohibited, permitted with disclosure, or encouraged for a specific learning purpose?
Have I provided examples of acceptable and unacceptable AI use for this assignment?
How should students cite, acknowledge, or document their AI use?
Does the task require students to show process, reasoning, judgment, or application?
How will students verify AI-generated content for accuracy, bias, and relevance?
Have I considered privacy, data governance, and University of Calgary obligations under POPA and ATIA?
Have I considered equity of access, including paid tools, and provided alternatives?
How will I keep human judgment central in any AI-assisted feedback or grading?
How will I revise this assessment after seeing how students actually use AI?

Bommasani, R., Hudson, D. A., Adeli, E., Altman, R., Arora, S., von Arx, S., … Liang, P. (2021). On the opportunities and risks of foundation models. arXiv. https://doi.org/10.48550/arXiv.2108.07258

Chan, C. K. Y., & Hu, W. (2023). Students’ voices on generative AI: Perceptions, benefits, and challenges in higher education. International Journal of Educational Technology in Higher Education, 20, Article 43. https://doi.org/10.1186/s41239-023-00411-8

Gutiérrez, J. (2023). Guidelines for the use of artificial intelligence in university courses (Version 4.3). Universidad del Rosario. https://forogpp.files.wordpress.com/2023/02/guidelines-for-the-use-of-artificial-intelligence-in-university-courses-v4.3.pdf

Ji, Z., Lee, N., Frieske, R., Yu, T., Su, D., Xu, Y., Ishii, E., Bang, Y. J., Madotto, A., & Fung, P. (2023). Survey of hallucination in natural language generation. ACM Computing Surveys, 55(12), Article 248. https://doi.org/10.1145/3571730

Kasneci, E., Seßler, K., Küchemann, S., Bannert, M., Dementieva, D., Fischer, F., Gasser, U., Groh, G., Günnemann, S., Hüllermeier, E., Krusche, S., Kutyniok, G., Michaeli, T., Nerdel, C., Pfeffer, J., Poquet, O., Sailer, M., Schmidt, A., Seidel, T., … Kasneci, G. (2023). ChatGPT for good? On opportunities and challenges of large language models for education. Learning and Individual Differences, 103, Article 102274. https://doi.org/10.1016/j.lindif.2023.102274

Luo (Jess), J. (2025). How does GenAI affect trust in teacher–student relationships? Insights from students’ assessment experiences. Teaching in Higher Education, 30(4), 991–1006. https://doi.org/10.1080/13562517.2024.2341005

Qian, Y. (2025). Pedagogical applications of generative AI in higher education: A systematic review of the field. TechTrends, 69, 1105–1120. https://doi.org/10.1007/s11528-025-01100-1

Russell, S. J., & Norvig, P. (2020). Artificial intelligence: A modern approach (4th ed.). Pearson.

White, J., Fu, Q., Hays, S., Sandborn, M., Olea, C., Gilbert, H., Elnashar, A., Spencer-Smith, J., & Schmidt, D. C. (2023). A prompt pattern catalog to enhance prompt engineering with ChatGPT. arXiv. https://doi.org/10.48550/arXiv.2302.11382

This guide was developed with reference to current scholarship on generative AI in higher education. The following sources informed the framing, examples, and recommendations throughout, even where they are not cited directly.

Dabis, A., & Csáki, C. (2024). AI and ethics: Investigating the first policy responses of higher education institutions to the challenge of generative AI. Humanities and Social Sciences Communications, 11, Article 1006. https://doi.org/10.1057/s41599-024-03526-z

Eaton, S. E. (2022, December 9). Sarah’s thoughts: Artificial intelligence and academic integrity. Learning, Teaching and Leadership. https://drsaraheaton.wordpress.com/2022/12/09/sarahs-thoughts-artificial-intelligence-and-academic-integrity/

Eaton, S. E., & Anselmo, L. (2023, January 12). Teaching and learning with artificial intelligence apps. Taylor Institute for Teaching and Learning, University of Calgary. https://taylorinstitute.ucalgary.ca/teaching-with-AI-apps

Farazouli, A., Cerratto-Pargman, T., Bolander Laksov, K., & McGrath, C. (2024). Hello GPT! Goodbye home examination? An exploratory study of AI chatbots’ impact on university teachers’ assessment practices. Assessment & Evaluation in Higher Education, 49(3), 363–375. https://doi.org/10.1080/02602938.2023.2241676

Farrelly, T., & Baker, N. (2023). Generative artificial intelligence: Implications and considerations for higher education practice. Education Sciences, 13(11), Article 1109. https://doi.org/10.3390/educsci13111109

Moorhouse, B. L., Yeo, M. A., & Wan, Y. (2023). Generative AI tools and assessment: Guidelines of the world’s top-ranking universities. Computers and Education Open, 5, Article 100151. https://doi.org/10.1016/j.caeo.2023.100151

Office of the Information and Privacy Commissioner of Alberta. (n.d.). ATIA/POPA resources. https://oipc.ab.ca/atia-popa-resources/

Perkins, M., Roe, J., Postma, D., McGaughran, J., & Hickerson, D. (2023). Detection of GPT-4 generated text in higher education: Combining academic judgement and software to identify generative AI tool misuse. Journal of Academic Ethics, 22, 89–113. https://doi.org/10.1007/s10805-023-09492-6

Sardi, J., Darmansyah, Candra, O., Yuliana, D. F., Habibullah, Yanto, D. T. P., & Eliza, F. (2025). How generative AI influences students’ self-regulated learning and critical thinking skills: A systematic review. International Journal of Engineering Pedagogy, 15(1), 94–108. https://doi.org/10.3991/ijep.v15i1.53379

Tong, S. T., DeTone, A., Frederick, A., & Odebiyi, S. (2025). What are we telling our students about AI? An exploratory analysis of university instructors’ generative AI syllabi policies. Communication Education, 74(3), 261–282. https://doi.org/10.1080/03634523.2025.2477479

University of Calgary. (n.d.). Artificial intelligence. https://ai.ucalgary.ca/

Xia, Q., Weng, X., Fan, O., Lin, T.-J., & Chiu, T. K. F. (2024). A scoping review on how generative artificial intelligence transforms assessment in higher education. International Journal of Educational Technology in Higher Education, 21, Article 40. https://doi.org/10.1186/s41239-024-00468-z