Unauthorized aids, generative AI, and collaboration

The instructor draws the line, in writing

Two of Policy 71's labels carry a qualifier: unauthorized co-operation or collaboration and unauthorized aids or assistance are each as defined by the course instructor. That qualifier is the whole lesson. There is no university-wide list of permitted help; the assignment's instructions — and the course outline behind them — say what is allowed, and anything outside them is unauthorized. The same act (discussing an approach with a classmate) is fine on one assignment and an offence on the next. So the first thing to read on any assignment is not the question; it is the line that says what help is permitted.

For generative AI the Office of Academic Integrity's guidance (read September 2026) says exactly this: instructors are asked to state explicitly whether and how far AI tools may be used, and a student who does not follow the instructor's rules is subject to Policy 71 and an investigation into academic misconduct. The policy's own home for it is unauthorized aids or assistance — and a submitted AI draft is also not your work, which reaches plagiarism.

Worked example — one instruction line, four uses

An AFM 112 assignment carries this line: "Individual assignment. You may discuss concepts with classmates but not answers. Generative-AI tools are not permitted on this assignment." Rule on four things students actually did.

1. A study group that talks through what a pivot table is for, then each member builds their own. Permitted — this is discussing concepts. Evidence of the line: each spreadsheet is built independently and the numbers, labels and layout differ.
2. A tutor who checks your finished answers and tells you which are wrong. Not permitted — the tutor is supplying assistance on answers, which the line restricts to concepts. Note that the tutor being paid, or being a family member, changes nothing.
3. A generative-AI first draft that you "clean up" and submit. An offence twice over: an unauthorized aid where the instructor has said AI is not permitted, and the submitted work is not yours. "No person wrote it" is not a defence — the policy asks whether the work is yours and whether the aid was permitted, not who or what produced it. Editing the draft does not make it yours any more than editing a friend's essay would.
4. Two friends who "only compared answers" before submitting, then each changed two of theirs. Unauthorized collaboration — the line permits concepts, and answers are what they compared. Comparing is collaborating; the fact that some answers still differ is not a defence, and neither is "we didn't copy, we checked".

Four more to rule on

Under the same instruction line:

• Asking an AI tool to explain what a variance is, then closing it and doing the assignment — permitted use of a concept explanation? The instruction says AI tools are not permitted on this assignment; the safe reading is that any use on the assignment is out, and the honest move is to ask the instructor before, not argue after.
• A shared Google Sheet where three students build the model together, then each submits a copy — unauthorized collaboration for all three.
• Reading the textbook's worked example of the same technique — permitted; the textbook is the course's own material.
• Posting the question to a homework-help site and using the answer — unauthorized aid, and the answer is not yours.

What decides it — and what does not

What decides whether help is permitted is the assignment's instructions. What does not: whether the work is graded (an ungraded practice piece with an "individual" line still has the line), the course level, how much you changed, whether money changed hands, or how the TA feels about it. When the instructions are silent, the default at a university is individual work with the course's own materials — and the professional move is to ask. An email to the instructor asking whether a study group may compare final answers takes two minutes and is itself evidence of integrity.