Can Canvas Detect AI? What Teachers Can and Can't See
By Ashikur Rahman11 min readQuick answer: no, Canvas itself does not detect AI, but your school can connect a tool that does. An Instructure community coach put it plainly in 2025: "Canvas does not provide instructors with access to an AI generated paper detection feature." Detection arrives through add-ons, usually Turnitin, whose report opens from SpeedGrader if your institution enabled it. Canvas logs can show page loads and quiz events, not what else was on your screen. So the real question is which tools your school added and how your teacher set up this assignment.
Why doesn't Canvas detect AI on its own?
Canvas is a learning management system, so it stores and delivers coursework and leaves analysis to connected tools. Plagiarism checks have always worked this way. In an accepted answer on Instructure's community, coach and University of Michigan-Dearborn digital education director Chris wrote: "Canvas itself does not check for plagiarism or provide a similarity score, rather that is done by 3rd party integrations with Canvas (like Turnitin or Copyleaks)."
The clearest statement on AI came from community coach Noah in February 2025: "the short answer is no." Those are community answers rather than formal product statements, and they are not recent. Instructure's AI privacy notice, read on 6 November 2026, gives its current AI tools as examples, including discussion summaries, translation and smart search. It does not list a writing detector among those examples. That supports the same reading, but it is not an explicit denial, and products change, so check with your school's Canvas team.
Which AI detectors can schools add to Canvas?
Turnitin is the common route, and other vendors offer integrations too. Which one you face depends entirely on your institution.
| Tool | How it connects | What the source says |
|---|---|---|
| Turnitin | Canvas Document Processor, or other Turnitin integrations | "Turnitin is currently the only supported document processing tool." |
| Copyleaks | Integration promoted on Instructure's site | Instructure's webinar page describes "transparent, explainable AI detection" |
| GPTZero | Institution-level, sometimes through partners | An Instructure video features one college using GPTZero with K16 Solutions |
| Grammarly Authorship | Student-side process record | Grammarly says it can be enabled in "Canvas LMS submissions" |
| Nothing | Some institutions choose this | Waterloo discontinued Turnitin's AI detection in September 2025 |
Having Turnitin does not mean having its AI check. Turnitin's FAQ says AI writing detection is "only available to customers that license Turnitin Originality" and that admins can disable it. A school can run similarity checks with the AI feature off. The University of Waterloo did exactly that, keeping similarity while discontinuing AI detection after internal testing flagged human text as 100% AI in some cases.
What does a teacher see in Canvas when Turnitin is on?
A teacher sees a link to a report in SpeedGrader, not a warning on the Canvas page. Turnitin's FAQ says "There is no AI writing indicator or score embedded directly in the LMS user interface and users will need to go into the report to see the AI score."
Turnitin's Canvas workflow guide describes the current setup:
- The instructor creates an assignment with the Online submission type and chooses Add Document Processing App.
- They select Turnitin and configure its settings, including whether students can view the report.
- After a student submits, the instructor opens SpeedGrader and finds "Document processor reports".
- They select View Report for Turnitin Similarity and AI writing, "if enabled for your institution".
Students get access only if the instructor ticked "Allow students to view report after submitting", and "This setting is locked after assignment creation." If it was left off, you will not be able to open the report for that assignment, so a missing report tells you nothing about whether one exists.
| In the report | What it means |
|---|---|
| A percentage from 20 to 100 | Share of qualifying prose sentences Turnitin rates as likely AI |
| An asterisk (*%) | Something below 20%; no score or highlights |
| 0% | Nothing identified |
| Could not process | File too short, too long, or in an unsupported type or language |
For the mechanics behind that number, read how Turnitin detects ChatGPT and what its score means. Whether you should trust it is a separate matter: see how accurate Turnitin's AI detector is.
Can Canvas detect AI in discussion posts?
Not by itself, and Turnitin's AI report is a poor fit for most discussion posts. Turnitin's Document Processor guide lists online Text Entry and File Upload for assignments, and does not mention discussion boards. Your school may use another integration, so this is not universal. The bigger limit is length. Turnitin needs "at least 300 words of prose text", so a typical short reply gets no AI score at all.
Teachers still have the oldest method available: reading. A post that ignores the week's discussion or sounds nothing like your earlier ones can raise questions with no detector involved. Instructors can also compare posts across the term.
Can teachers see copy and paste or tab switching in Canvas?
Generally not reliably, and the two sources read here need careful handling. Canvas keeps records of what reaches its servers. The most detailed explanation found is a January 2026 accepted answer from an English department head active on Instructure's forum, who describes logs as "event-based, server-side records" and "diagnostic breadcrumbs, not behavioral telemetry." That person is an experienced community member, not an Instructure employee.
| Activity | Forum answer's account of Canvas logs | What else applies |
|---|---|---|
| Page loads, navigation within Canvas | Reliably recorded | Timestamps do not show attention |
| Quiz start, auto-save, answers, submission | Reliably recorded | Quiz logs show these events |
| Switching tabs or window focus | "do not reliably record" | Classic quiz events include page_blurred and page_focused |
| Keyboard and mouse activity | Not reliably recorded | A proctoring tool might |
| Other apps or a second device | Not recorded | Canvas cannot see them |
| Copy and paste | Not listed as recorded | Paste tracking lives in process tools |
The tab-switching row is where many articles go wrong in both directions. Instructure's Quiz Submission Events API shows page_blurred and page_focused events in its examples, so Canvas can receive a signal when a quiz page loses focus. But that does not make it proof. As that answer notes, logs "cannot distinguish intent", and a gap may be a network drop or a browser event. It warns that "Using logs as standalone proof almost always leads to false positives." Details also vary by quiz type: Classic Quizzes and New Quizzes keep different logs, and Instructure's own knowledge-base pages on them did not render when this article was researched.
New Quizzes also has a "Detect Multiple Sessions" setting. Instructure's permission documentation says educators turn it on to collect multi-session data, viewable by those with the permission on the activity log and moderate page. So a quiz can be set up to record more than an essay assignment would.
Transparency note: CriticNest has no affiliate or commercial relationship with Instructure, Turnitin, Copyleaks, GPTZero, Grammarly or Respondus. Nothing here is a paid placement. No link in this article earns a commission. Every quotation was read from the linked pages on 6 November 2026. No Canvas course or quiz was set up or tested for this article. If that changes, the disclosure will move up.
What can exam proctoring tools see that Canvas can't?
Much more, because they run on your device instead of on Canvas's servers. If your exam opens in a separate locked browser, normal assumptions no longer apply. Respondus LockDown Browser, which integrates with Canvas, states what it does:
- "Assessments are displayed full-screen and cannot be minimized".
- "Copying and pasting anything to or from an assessment is prevented".
- "Right-click menu options, function keys, keyboard shortcuts and task switching are disabled".
- It blocks "Virtual Machines, remote desktops, screensharing, instant messaging, screen recording" and more.
Respondus also sells a companion product, Respondus Monitor. Which tools apply is usually stated in the exam instructions. So "Can Canvas detect AI extensions?", "split screen", "screenshots" and "Google Lens" have the same honest answer: plain Canvas logs have no view of your screen, but a locked browser, a proctoring tool or your school's device-management software might, and those rules are set locally. This article does not advise on getting around any of them. Using banned help in a closed exam is misconduct whether or not software sees it.
Can teachers tell you used AI without any detector?
Often they can raise a reasonable question, and for coursework that may matter more than software. These are ordinary academic checks, available in any course:
| What a teacher compares | What might raise a question |
|---|---|
| Your earlier work | A sudden change in voice, vocabulary or structure |
| Citations | Sources that do not exist. Pittsburgh's Teaching Center says hallucinated sources can be a violation by themselves. |
| The assignment brief | General answers that miss class-specific material |
| A short conversation | Being unable to explain your own argument |
| Drafts and history | A document that appears all at once |
None of these is proof either. Writing improves, students get help from tutors, and nervous people explain themselves badly. They are reasons for a conversation, which is what both Turnitin and GPTZero say a detector should lead to as well.
Does Canvas show if you used Grammarly or QuillBot?
Canvas does not identify the tools you used. A connected detector may still react to what those tools did to your sentences. Turnitin says grammar corrections were "in most cases" not flagged, while generated or paraphrased output is a different matter. The details are in whether Grammarly shows up as AI and whether QuillBot shows up as AI.
How do you find out what your course actually uses?
Ask, and read what is already in front of you. Most of this is findable in a few minutes.
- Read the syllabus and assignment page. Look for an AI policy, plus any mention of Turnitin, Copyleaks or similar.
- Check the submission page. An assignment that sends work to Turnitin may mention it there. Some schools also state it in their student handbook.
- Search your school's IT or academic integrity site. Schools often publish whether AI detection is on. Waterloo and Pittsburgh both do.
- Ask your instructor directly. "Is AI detection on for this assignment, and what counts as allowed help?" is a reasonable question, and the answer protects you both.
How can you protect yourself if you did nothing wrong?
Create a record while you work, because the evidence is hardest to produce afterward.
- Draft in a tool with version history, such as Google Docs, rather than typing the whole thing into a Canvas text box.
- Save notes, outlines and sources as you go.
- If you use a process recorder such as Grammarly Authorship, switch it on before the first word. It cannot record what happened while it was off.
- Note any allowed AI help the way your course asks.
- If you are questioned, stay specific and calm, ask to see the report, and refer to the instruction that Turnitin itself gives: its score "should not be used as the sole basis for adverse actions against a student".
How this was researched
Read on 6 November 2026: Turnitin's FAQ and Canvas workflow guide through its help-centre API; Instructure's developer documentation on quiz events and multi-session permissions, its AI privacy notice and three community threads; the Respondus LockDown Browser page; and pages from the University of Waterloo and University of Pittsburgh.
Not verified: Instructure's knowledge-base articles on quiz logs, which rendered empty; any school's specific configuration; whether a third-party integration handles discussion boards at your institution. Community forum answers are labelled as such and are not Instructure policy. No Canvas course or quiz was tested. The author's role is careful reader of the fine print, not tester.
The verdict: Canvas doesn't detect AI, but your course setup might
Canvas is the container. Detection comes from whatever your institution connected to it, most often Turnitin, and only if the AI feature is licensed and on. Teachers reach that report through SpeedGrader. Plain Canvas logs show server-side activity, with limits their own users warn about, while locked browsers and proctoring tools can see far more.
So don't plan around what Canvas can or can't see. Find out your course's rules, write in a way you can show, and be ready to explain your work. That holds whichever tools your school chose.
Frequently asked questions
Does Canvas automatically check for AI?
No. Canvas has no built-in AI writing detector, according to Instructure community coaches. An AI check runs only when your school connects a tool such as Turnitin, enables its AI feature, and the assignment sends work through it. Your instructor or school IT site can confirm this.
Can Canvas detect ChatGPT?
Canvas alone cannot. If your assignment uses Turnitin with AI writing detection on, Turnitin's model can flag text it rates as likely from ChatGPT, for submissions with at least 300 words of prose. Your teacher then sees that report through SpeedGrader rather than on the Canvas page itself.
Does Canvas detect AI without Turnitin?
Not by itself. A school might use a different integration, such as Copyleaks or GPTZero, which Instructure's own site features in its resources. If none is connected, no automatic AI score exists, although your teacher can still compare your work, check your citations and ask about it.
Can Canvas detect AI on a PDF?
Canvas does not scan it, but Turnitin accepts .pdf files for its AI report when that feature is enabled. The PDF needs at least 300 words of long-form prose. A scanned image of a page may not offer readable text; Turnitin's pages read did not settle that case.
Can Canvas detect AI images?
No source read describes Canvas detecting AI images. Turnitin's AI writing report covers prose text and says it does not process other formats reliably. Some vendors sell separate image checks, and your school's policy on AI images still applies whether or not a tool checks them.
Can Canvas see if you switch tabs?
Sometimes a quiz log may show a signal. Canvas's Classic quiz events API includes page_blurred and page_focused events, but a forum explanation says logs "do not reliably record" tab changes and cannot show intent. Locked browsers go further by disabling task switching during the exam.
Can teachers see what you do on Canvas?
They can see activity inside Canvas, such as page views, quiz events and submission times, depending on tools and permissions. They cannot see other apps, other tabs' contents or another device through those logs. A school proctoring tool or a managed device may be different.
Can Canvas see copy and paste?
Plain Canvas logs are not described as recording it. Paste tracking comes from other tools: Turnitin Clarity reports pasted text in its own writing environment, and GPTZero's Writing Replay shows pastes in Google Docs. Respondus LockDown Browser prevents copying and pasting during exams.
Can students see the Turnitin AI report in Canvas?
Only if the instructor allowed it. Turnitin's Canvas guide has an "Allow students to view report after submitting" setting that locks once the assignment is created. Without it, students cannot view the report. Instructors can also download an AI report as a PDF and share it with you.
What happens if my teacher thinks I used AI on Canvas?
Your school's integrity policy sets the steps, so read it first. Bring your drafts, version history and notes, and ask to see any report. Turnitin says it "does not make a determination of misconduct", so the decision rests with people, under local rules.
