What U of T's Rewritten Academic Integrity Code Actually Says About Using AI
By Muntasir • Published May 05, 2026 • Updated Sep 20, 2026 • Canadian Universities & Colleges, Canada Higher Ed News & Policy
Since July 1, 2025, the University of Toronto treats generative AI as banned by default on marked work. Using it without your instructor's explicit permission counts as an unauthorized aid under the Code of Behaviour on Academic Matters, the same category as bringing in a cheat sheet.
Governing Council approved the updated Code of Behaviour on Academic Matters on June 24, 2025, effective July 1, 2025.
Default rule: AI tools are off-limits on marked assessments unless your instructor says otherwise.
You commit an offence if you use AI knowingly, or if you ought reasonably to have known it was not allowed.
🎓 If a course syllabus is silent on AI, ask your instructor before you use any tool on graded work.
Graduate students need written supervisor approval before using AI on a thesis or other milestone.
The rule that changed
The University of Toronto's Governing Council approved changes to the Code of Behaviour on Academic Matters on June 24, 2025, effective July 1, 2025. The update folds generative AI tools such as ChatGPT into the existing definition of unauthorized aid, the same category that already covered unapproved notes in a closed-book exam or an unapproved calculator on a math test.
The default rule is simple. AI use is off-limits on marked work unless your instructor says it is allowed. There is no university-wide green light. Permission comes course by course, and sometimes assignment by assignment within the same course.
What counts as an offence
The Code defines an offence as using or possessing an unauthorized aid, or getting unauthorized assistance, on work submitted for academic credit. Under U of T's guidance, if your instructor prohibited AI tools for an assignment and you use one anyway, even for part of the work, that counts as using an unauthorized aid. It does not matter whether the AI-generated portion is small.
The university gives one clear example: writing a term paper with ChatGPT after your instructor told the class not to use generative AI tools is treated as knowingly using an unauthorized aid, a formal academic offence under the Code.
Where the standard gets stricter: ought reasonably to have known
You do not need to have read a syllabus line by line to be held responsible. The Code applies if you used AI knowingly or if you ought reasonably to have known it was not permitted. Silence from an instructor is not permission. If a syllabus does not mention AI at all, ask before you use a tool on graded work rather than assuming it is allowed.
When AI use is allowed: what you still need to do
Some instructors permit generative AI for parts of an assignment, such as brainstorming, code debugging or first-draft editing, while restricting it for others. When that is the case, follow the instructor's stated rules exactly and disclose the tool you used. The University of Toronto Libraries maintains a guide on citing generative AI tools for exactly this situation. Treat AI use the way you treat a source: name it, and be ready to explain what it contributed.
Graduate students face an extra layer
If you are a graduate student, the Code is not the only rule that applies to your thesis, comprehensive exam or other milestones. The School of Graduate Studies requires prior written approval from your supervisor and supervisory committee before you use generative AI on any academic milestone. You need to disclose which tools you used, how, and why, cite them under your discipline's normal citation rules, and be ready to defend that use at your final oral exam.
Your supervisor keeps the right to restrict AI use in your writing even after approving it for other parts of your research, such as literature searches or data analysis. AI-generated content on its own does not satisfy the originality requirement for doctoral work. Get approval in writing before you rely on AI for anything connected to a milestone, not after.
Why U of T rewrote the code now
The updated Code follows the university's own AI task force, which produced its final report in June 2025 under the banner Toward an AI-ready University. The academic integrity office reviews its guidance on an ongoing basis as generative AI tools change, so check the current page each term rather than relying on what a friend in an earlier cohort told you the rule was.
How to stay on the right side of the rule
Read the AI statement in every syllabus at the start of term, not only the ones for writing-heavy courses.
If a syllabus does not mention AI, email your instructor and get a written answer before you submit anything.
When AI use is allowed, disclose it using the library's citation guide rather than leaving it out.
Keep drafts and notes for major assignments. They show your own process if a question comes up.
Do not assume a rule from one course, or from last term, applies to your current one.