Does Canvas SpeedGrader Show Turnitin AI Scores When Hidden from Students?

Yes. In Canvas, your professor can see your Turnitin AI detection score in SpeedGrader even if the originality report is completely hidden on your student dashboard.

The core reason lies in how Canvas structures its permission architecture. Assignment visibility settings govern what is released to students, not what is generated for instructors. When a course instructor chooses to withhold similarity or originality reports from students, that toggle alters the student view exclusively. It does not pause Turnitin processing, nor does it suppress the data sent to the instructor's grading view.

Unless your institution has disabled Turnitin's AI writing detection feature at the institutional account level, every submitted document that meets Turnitin's technical requirements generates a report for the grader. In SpeedGrader, the instructor sees a Turnitin badge on the sidebar and can launch the complete Feedback Studio workspace with a single click, bringing up both the text-matching similarity breakdown and the AI writing indicator.

When an instructor sets up an assignment using the Canvas Plagiarism Framework or Turnitin LTI integration, Canvas provides a dedicated configuration field labeled "Show report to students." This dropdown menu determines when, or if, students can inspect their Turnitin results.

Canvas offers four distinct timing options for this setting:

• Immediately: Students can view the similarity report as soon as Turnitin completes processing the uploaded file.
• After the assignment is graded: The report remains hidden from students until the instructor enters and posts a grade.
• After the due date: Students cannot view the report until the formal assignment deadline has passed.
• Never: The report is never made accessible to students through their Canvas portal.

Students frequently encounter the "Never" setting—or an empty score badge—and assume that Turnitin was either disabled for the assignment or that reports were withheld across the board. That assumption misreads the function of the control.

According to Canvas documentation and administrator guides, this setting operates purely as a student release filter. It defines the visibility permissions for the student role without altering the grader role. For instructors and teaching assistants opening SpeedGrader, report generation occurs automatically in the background upon document submission, regardless of which release option was selected.

To understand what your instructor actually sees, it helps to examine the layout of SpeedGrader. When a grader opens your submission, the left and center portions of the screen render your document inside Canvas DocViewer, allowing the instructor to highlight text, draw annotations, and attach point comments.

The right sidebar contains the grading tools: the student's submission timestamp, the file download link, rubrics, and the Turnitin badge. This badge—often referred to as the similarity pill—displays a percentage along with a color code indicating the extent of text matching (blue for 0%, green for 1% to 24%, yellow for 25% to 49%, orange for 50% to 74%, and red for 75% to 100%).

Clicking this colored pill immediately launches Turnitin Feedback Studio in an overlay window or a new browser tab. Feedback Studio is the full inspection interface where all Turnitin analysis takes place.

Inside Feedback Studio, the right-hand navigation bar divides Turnitin's tools into distinct functional layers. The red layer houses the Similarity Report, detailing matched sources across web repositories and published journals. At the bottom of the toolbar sits the blue AI Writing indicator. Selecting this icon opens the complete AI Writing Report, displaying the overall estimated AI percentage (or an asterisk for low-percentage detections between 1% and 19%) along with sentence-by-sentence cyan highlights across your manuscript.

A common source of confusion among students is equating the colored percentage on Canvas with their AI score. Seeing a low green number, such as 8%, often leads writers to believe their submission cleared both plagiarism and AI checks.

In reality, the similarity index and the AI indicator are two independent readings generated by separate models operating on fundamentally different principles. The similarity score measures text matching—direct string overlap against billions of web pages, student papers, and academic articles. The AI detector, by contrast, evaluates statistical characteristics of prose, examining sentence-level perplexity and burstiness to identify patterns typical of large language models.

Because these tools evaluate different properties, their outputs routinely diverge. An essay generated entirely by an AI model in response to an original prompt will often return a 0% similarity score because the phrasing does not match existing indexed texts, while returning an AI indicator of 90% or higher. Conversely, a comprehensive literature review containing properly attributed quotations may display a 30% similarity score with an AI indicator of 0%.

Furthermore, Canvas SpeedGrader displays only the similarity score on its outer sidebar pill. The AI detection percentage is never displayed directly on the Canvas interface without opening Turnitin Feedback Studio. Relying on a low similarity pill on your submission page provides no information about what the AI indicator recorded.

While instructors can access the AI report in most standard Canvas setups, there are specific technical and administrative conditions under which an instructor genuinely will not see an AI score.

Before drawing conclusions about what was evaluated, consider the following four scenarios:

• Institutional opt-out: Several universities have formally disabled Turnitin's AI writing detection feature at the institutional administrator level due to documented concerns regarding false positive rates. If your university opted out, the AI indicator icon is removed from Feedback Studio for all instructors across the entire campus.
• Unsupported document format or word count: Turnitin's AI detection model processes only qualifying English prose. It requires a minimum of 300 words of standard grammatical text. If a submission consists primarily of code, mathematical notation, bulleted slides, or image-based scans, the AI engine will return a blank reading or a double dash (`--`) rather than a percentage.
• Assignment configured without Turnitin: Instructors can set up Canvas assignments using standard online file uploads without enabling the Turnitin Plagiarism Review framework, meaning no Turnitin processing takes place.
• Processing timeouts during peak traffic: During heavy submission periods, such as midterm or finals week, Turnitin reports can experience processing backlogs, displaying a pending clock or error icon in SpeedGrader until generation completes.

Unless your institution maintains a published policy against AI detection or has disabled the tool institution-wide, students should assume that any standard paper submitted to a Turnitin-enabled Canvas portal gives the grader full access to the AI writing report.

If an instructor contacts you regarding an AI flag in Canvas SpeedGrader, approaching the conversation methodically is essential. The first step is to request the official Turnitin AI Writing Report PDF rather than discussing an isolated number mentioned in grading comments. You need to examine the exact sentences highlighted in cyan, the document word count counted as qualifying text, and the generation timestamp.

Once you have the report, compare the flagged passages against your documented writing workflow. Timestamped Google Docs revision histories, Word version logs, research notes, and initial outlines provide concrete evidence of authentic drafting that can clarify how specific passages took shape.

If your instructor permits or requests a revision of flagged sections, avoid the mistake of rewriting the entire manuscript from scratch. Complete rewrites often introduce new stylistic inconsistencies, damage carefully structured bibliographies, and scramble technical terminology.

Import the Turnitin or iThenticate report and rewrite only the passages it flags, returned as the same editable file.

Targeted revising addresses the specific sentences that triggered the statistical flag while leaving approved arguments, empirical tables, citation keys, and document formatting completely undisturbed.

Why do instructors choose to hide Turnitin reports from students in Canvas? Instructors typically set report visibility to "Never" or "After grading" to prevent students from repeatedly adjusting phrasing to game similarity thresholds before submitting. Others hide reports to prevent unnecessary student anxiety over routine bibliographic matching before the instructor has had an opportunity to review the context of the paper.

Does Canvas automatically notify an instructor if a Turnitin AI score is high? No. Canvas does not send automated alerts, flag submissions with pop-up warnings, or automatically deduct points based on Turnitin AI percentages. The instructor or teaching assistant must manually open SpeedGrader and launch Feedback Studio to inspect the reading.

Can teaching assistants (TAs) see the Turnitin AI report in SpeedGrader? Yes. Anyone assigned a grading role in Canvas—including co-instructors and teaching assistants—generally possesses the permissions required to view SpeedGrader and launch Turnitin Feedback Studio, unless an institutional administrator has customized sub-account role permissions to restrict external tools.

If I resubmit a paper before the deadline, can the professor still view my previous AI score? Yes. While SpeedGrader displays the most recent submission attempt by default, a dropdown menu on the sidebar allows the grader to toggle between all previous submission attempts. If Turnitin was active for earlier uploads, the instructor can review the corresponding reports for each attempt.

Can students view their Turnitin AI score if the teacher makes the similarity report visible? No. Turnitin's current integration architecture restricts the AI writing detection indicator exclusively to instructors and administrators. Even when an assignment is set to release originality reports immediately, the student view typically provides access only to the text-matching similarity report, not the AI detection panel.