Canadian Application Feedback Rules — Where Help Stops

Canadian Application Feedback Rules — Where Help Stops

There is no single Canadian rule for application feedback or AI use. The boundary can change between two programs at the same university.

Queen's Medicine prohibits generative AI writing tools in any application component. Queen's Faculty of Education says its Personal Statement of Experience must be prepared without professional assistance. McGill Law, by contrast, says candidates “may even consider” having a friend, family member, or academic advisor check a personal statement for errors, while cautioning against reliance on generative AI.

Those are different rules. Treating them as one university-wide policy can put an otherwise careful applicant on the wrong side of a program's instructions.

This comparison uses official pages checked on July 29, 2026. Policies can change between cycles or move into an application portal. Recheck the exact program and application component before asking a person or a tool to review your work.

Start With Four Different Kinds of Assistance

Application pages use familiar words without always defining them. Before reading a policy, separate four actions:

• Proofreading identifies spelling, grammar, or typographical errors without changing the applicant's ideas or authorship.
• Feedback points out what is unclear, unsupported, or missing so the applicant can decide whether and how to revise.
• Editing or rewriting changes the expression itself, sometimes substantially.
• Generation produces ideas, examples, structure, or submitted language for the applicant.

A program that permits proofreading has not necessarily approved substantive feedback. A warning about AI-generated false information is not necessarily a complete list of allowed AI uses. Silence is not permission.

We quote each rule as published rather than reading it to its outermost stretch. Where a program bars professional assistance or collaborative preparation, that sentence is the one you are deciding against — GradPilot reviews writing you already wrote, and how you apply the rule is your call.

The Rules at a Glance

The table is a reading aid, not a substitute for the current application instructions.

Waterloo AIF — Independent Work and AI-Assisted Scoring

The University of Waterloo tells applicants that they must complete and submit the Admission Information Form on their own. Its AIF guidance also says Waterloo uses an AI system alongside trained human scorers. When the scores differ significantly, a second human reviews the submission.

That is a disclosure about how Waterloo evaluates an AIF. It is not evidence that an applicant may use any particular kind of AI assistance.

The same page warns about answers produced with AI that are untrue, incorrect, or plagiarized. That warning should not be reversed into an assumption that all truthful AI-assisted work is allowed. The independent-work instruction still applies.

For applicants, the defensible reading is simple: write and submit the AIF yourself, and do not treat Waterloo's use of AI in scoring as permission for AI in authorship.

Queen's Medicine — No Generative AI in Any Component

Queen's Medicine publishes one of the clearest restrictions in this group. Its MD eligibility and application page says applicants may not use generative AI writing tools to obtain responses at any stage of assessment. It also says discovered use in any application component results in disqualification.

The rule is program-specific and broad across the MD application, and that published sentence is the one that governs a Queen's MD response. GradPilot reviews writing you already wrote; the determination is yours, from Queen's own words.

Do not project that rule onto every Queen's program. The Faculty of Education publishes a different boundary, discussed below, and another faculty may publish another one.

McMaster BHSc — Responses Must Be Entirely Your Own

McMaster's Honours Health Sciences program answers the help question directly. Its supplementary application FAQs, under the heading “Can I get help from others when writing my Supplementary Application?”, say:

“Your responses to the questions must be entirely your own. Discussion with others, including current program students, is not permitted. Posting answers to third party websites or online forums is strictly prohibited. Use of admissions consultants is prohibited.”

The same page adds:

“Posting your answers on the internet, whether draft or final, use of paid admissions consultants, plagiarism, use of generative artificial intelligence, or otherwise working with others to prepare your responses is not permitted and may result in removal from the applicant pool or withdrawal of an offer of admission.”

That page publishes no proofreading or feedback exception. The answer to the literal question is an unqualified no, and McMaster calls a violation a breach of academic integrity that is grounds for removal from the applicant pool.

This is the strictest boundary in the comparison, and it makes the product decision simple. Do not send a BHSc supplementary response to GradPilot, to any other tool, or to another person for review. Write every idea and sentence yourself.

Note that McMaster's general supplementary application overview page carries no AI or collaboration rule at all — the rule lives on the FAQ page linked above, which was re-verified on July 30, 2026. Check it again in the current portal before you submit.

University of Toronto Medicine — Plagiarism Checks Are Not an AI Rule

The University of Toronto MD non-academic requirements say the program performs random plagiarism checks on brief personal essays and adds submitted essays to the tool's reference database.

That tells applicants about similarity screening. The checked page does not publish a rule permitting or prohibiting specific uses of generative AI.

Do not collapse the two questions:

1. Does the program compare text for overlap? The published page says yes.
2. What AI assistance may an applicant use? The checked page does not answer that question.

Absence of a published AI rule is not approval. Use the current OMSAS and program instructions, and keep the submitted ideas and language your own. The sketch is also the part of the OMSAS application most often left too late — the OMSAS deadline work-back plan counts backwards from the October 1 close so verifier confirmations are not the last thing outstanding. UBC applicants face the same timing question a term later, and Sauder applicants answer additional faculty-specific Personal Profile questions on top of the general set.

McGill Law — Error Review Is Named, Rewriting Is Not

McGill Law says the personal statement should be the candidate's own reflection. Its supporting-document guidance cautions applicants against relying on generative AI. On the same page, after telling candidates to review and edit the statement themselves, it adds that they “may even consider having a friend, family member or academic advisor review the document to ensure it contains no errors.”

That is a permissive aside about error-checking, not a recommendation to seek substantive review. The useful distinction is between checking and replacing. A reviewer can flag a typo or an unclear sentence within that stated boundary. The applicant should not hand over the selection of experiences, the reflection, or the writing.

McGill's wording is a caution rather than a detailed AI permission matrix. Keep the human-review carve-out narrow and recheck the live instructions for the cycle in which you apply.

Western Law — Proofreading Yes, AI and Rewriting No

Western Law's personal-statement guidance places the permission and prohibition side by side. Applicants are encouraged to ask someone to proofread. They are told not to use AI and not to permit rewriting or excessive editing by others.

That makes the practical boundary clearer than a generic “get feedback” instruction:

• another person may identify mechanical errors;
• that person should not recast the statement;
• an AI tool should not be used for the application statement.

Western's rule permits a human proofread and rules out AI, rewriting, and excessive editing by others — that is the boundary as Western published it, and it is the one that governs this statement. A permitted human proofread should preserve the applicant's wording and decisions.

Western is one of eight Ontario law schools whose limits and authorship rules we verified together — Ontario's OLSAS statement limits, school by school sets Western's 5,000-character box beside Toronto's, Queen's and Osgoode's.

Queen's Education — No Professional Assistance

The Queen's Faculty of Education says its Personal Statement of Experience must be prepared in the applicant's own words without professional assistance.

That instruction reaches professional assistance even though it never names generative AI — a wider rule than an AI-specific one, and the sentence a PSE applicant is deciding against.

This is also why a university name alone is not enough. Queen's MD has an AI-specific prohibition; Queen's Education has a professional-assistance rule. Both are restrictive, for different published reasons. American applicants meet the same pattern, which we mapped in same school, different AI rules.

A Compliance-First Checklist

Before sharing any Canadian application response:

1. Find the instruction for the exact program, cycle, and component.
2. Check the live application or portal as well as public guidance.
3. Identify whether the rule addresses proofreading, feedback, editing, generation, or all assistance.
4. Treat “your own words” and “complete on your own” as real boundaries, not decorative language.
5. Do not infer permission from silence or from a university's use of AI.
6. If two official sources conflict, follow the current application interface and ask the program for clarification.
7. Recheck shortly before submission and keep a dated copy of the instruction you followed.

GradPilot's application AI policy directory records program-specific rules when a university publishes them and makes uncertainty visible when it does not. It does not certify compliance. Queen's Medicine, McMaster BHSc, Western Law, and the Queen's Education PSE publish the most restrictive sentences on this page — quoted verbatim above — and each one governs its own application component. GradPilot reviews writing you already wrote; the determination under each rule is yours.

For the broader distinction between policy language and detection claims, read what college AI checks can actually establish.