Purpose and Principles
This guideline exists not to restrict your use of AI tools, but to help you use them in ways that genuinely support your learning, protect your skills, and honour the integrity of your degree.
AI use is also an academic integrity matter. These expectations sit within Queen’s academic integrity policy and procedures, and using AI to do work that is yours to do can be a breach of the procedures as much as a missed learning opportunity. Disciplines differ in how they view the value and suitability of AI, so the syllabus for each course governs, and you are responsible for following, the terms set there. When a syllabus is silent, assume AI use is not authorized and ask your instructor before proceeding.
Core Commitment
This guideline is grounded in a single commitment: preserving the integrity of your learning. AI tools should support that learning, not replace it. When in doubt about any AI use, ask yourself one question: am I learning from this, or am I bypassing the learning entirely?
Your Learning Comes First
AI tools are most useful when they support active thinking rather than replace it. Students who hand intellectual tasks to AI miss the productive struggle that builds real capability.
Protecting Your Skills
Relying on AI for work you have not yet learned to do carries a long-term risk this guideline calls skill endangerment. You may never build a skill because AI does it from the start, you may stop developing one that is only partly formed, or you may lose one you no longer practice. A student who always has AI plan and revise papers may graduate unable to structure an argument independently. The point is to weigh the risk to your own capabilities before delegating, not only to ask whether a task is permitted.
Integrity and Accountability
The idea of delegation reframes academic integrity. Rather than only asking whether you cited AI, this guideline asks whether the work was yours to do alone. A degree carries public weight, because it certifies that the holder has demonstrated specified competencies, and if AI performed them the certification is not honest. Students in professional programs such as law, medicine, engineering, and education carry this forward into real accountability. Integrity here is less a set of rules imposed on you than a form of self-respect and professional responsibility you choose.
AI as Learning Support
This guideline does not restrict AI use for learning. Just as you may always consult a peer, tutor, or mentor to deepen your understanding, you may use AI in a learning-support role through querying, consulting, and scaffolding: to explain concepts, test your ideas, give feedback on drafts you have already written, or explore questions. What it restricts is AI standing in for the intellectual work of assessed tasks that are yours to do. Even in support mode, remember that AI can be confidently wrong, so check anything you rely on against trustworthy sources.
Five Patterns of AI Use
Every use of AI sits somewhere on a spectrum of delegation, from keeping all of the thinking yourself to handing the whole task to AI. The five patterns below name points along that spectrum, and the central question is always the same: is this an appropriate delegation, given what the course is asking you to learn and what your syllabus permits? The first three, querying, consulting, and scaffolding, are learning-support uses where the work stays yours. The last two, augmentation and delegation, reach into the work itself and require more caution as the risk to learning is higher.
| Patterns | What AI Does | Your Role | Typical Examples |
|---|---|---|---|
| Querying | Answers a question to inform your own understanding. | You do all the thinking and keep full authorship. | What a term means; a concept explained in plainer language; a fact or formula you then verify. |
| Consulting | Engages with your own reasoning in depth and pushes back on it. | You keep authorship; what you gain is sharper judgment, not new text. | Challenging your thesis; finding gaps in your argument; working through where your understanding is weak. |
| Scaffolding | Helps you start or structure work you will substantially do yourself. | You author the work; AI only gives you something to react to. | Suggesting angles you then develop; feedback on an outline you drafted; a rough structure you rework. |
| Augmenting | Improves the surface of work you have already done, where this is permitted. | You remain the author; AI refines, it does not supply the substance. | Tightening wording in your draft; checking citation formatting after you built the list; exploring other structures once you have a full draft. |
| Delegating | Produces the work itself: not only text, but plans, outlines, analysis, answers, or structure. | AI is the author of that part; your intellectual ownership is minimal. | Generating an essay, lab report, or outline; producing answers or analysis you submit as your own. |
Two cautions apply across all five. As you move down the table, the skill-endangerment risk rises, because more of the task is being done for you. And whatever the type, AI tools are linguistic pattern-matching systems rather than reliable answer machines, so they can sound authoritative and still be wrong, and anything you rely on must be verified by you.
A Few Terms Used Throughout
Being the genuine author of the ideas, the argument, and the decisions in your work, and standing behind it as your own. This is the heart of academic integrity, and it is what your degree certifies.
Using AI the way you would use a good tutor or knowledgeable peer, to deepen your own understanding. It spans querying, consulting, and scaffolding. A learning support clarifies; it does not produce assessable work on your behalf.
The risk that work handed to AI keeps you from building, fully developing, or retaining a skill you will need. It covers never acquiring a skill, an emerging skill stalling before it is fluent, and a learned skill fading from disuse. Queen’s AI literacy modules describe the same three risks as never skilling, de-skilling, and skill atrophy.
Decision Guide: Is This an Appropriate Delegation?
Use this before you act, not afterward. It starts with what you are trying to do, then works outward to what your course allows.
Ask what you are trying to do and whether it will help you learn, returning to the core question: am I learning from this, or am I bypassing the learning entirely? If the honest answer is that AI would be doing the learning for you, stop and reconsider, whatever the rules happen to allow.
Ask whether your instructor assigned this task because completing it is the learning, as is usually the case when you are writing an analysis, forming an argument, solving a problem, designing an experiment, or structuring a report. If so, do it yourself, with AI only in a learning-support role, and identify where your intended use sits on the spectrum from querying through delegation. Weigh the risk of skill endangerment, since relying on AI here may stop you building or keeping a skill you will need, and the epistemic risk that AI can be fluent and confidently wrong, so anything you rely on must be checked against trustworthy sources. Disclosure may also be required, so check your syllabus or ask your instructor.
Check your syllabus and any instructor guidance, since disciplines and courses differ in what they treat as appropriate. If AI use is restricted or prohibited for this task, stop. If the syllabus is silent, ask before proceeding. A useful heuristic is the housemate test: if it would be inappropriate for a housemate to do this part for you, it is usually inappropriate to hand it to AI. Treat that as a guide rather than a rule, because there are cases where help from AI is fine even where help from another person would not be.
Ask whether you remain the genuine author of the ideas, the argument, and the decisions, and whether you are building or keeping the skills the course is meant to develop. If AI produced the outline you did not think through, the writing you did not compose, or the conclusions you did not reach, the work is not yours in the way your degree claims, which is an academic integrity issue as much as a learning one.