Canvas - How to Use and Interpret Turnitin AI Writing Detection
What does the AI Writing Detection feature do?
Turnitin’s AI content checker can help “identify when an AI writing tool such as ChatGPT, text spinners, or AI bypassers (also called humanizers) may have been used in students’ submissions.” It analyzes a student’s submission to determine how much of it is likely “authentic, human-written content”, versus text likely generated or manipulated by AI.
After analyzing the submission, Turnitin displays an AI Writing Indicator score to instructors on the Turnitin Similarity Report page. Instructors can click the score button to open a detailed AI Writing Report.
For more information about how to view the AI Writing Indicator score on a Canvas assignment using Turnitin, visit Turnitin’s How to access the AI Writing Report page.
Can Students See This Score?
No, the AI Writing Detection score is not visible to students.
Only instructors and administrators are able to see the indicator or the AI Writing Report. You can download the AI Writing Report if you’d like to share it with the student.
When utilizing Turnitin in your Canvas course, please do not require students to reference the AI Writing Detection score.
How does the tool work?
When an instructor enables Turnitin plagiarism detection on a Canvas assignment, Turnitin will automatically display an “AI” AI Writing Detection score indicator box on the right side of the Turnitin Feedback Studio screen.
Blue (Score with a percentage between 0 and 100)
If Turnitin can calculate a score, it will display the score in that box. The higher the score, the more likely the submission was generated or manipulated using AI. (This is the opposite of Turnitin’s default Similarity Score, where a lower score is better.)
Low percentage (*%)
If the “AI” box displays an asterisk, Turnitin detected below a 20% threshold in the report: “False positives (incorrectly flagging human-written text as AI-generated) are a possibility in AI models. To avoid potential incidence of false positives, no score or highlights are attributed for AI detection scores in the 1% to 19% range.”
Gray with no percentage displayed (–)
AI writing detection is unavailable for some submissions—these will display an empty “AI” box with two dashes. Reasons for this include:
- Submission was made before the AI Writing Detection feature was enabled (2025).
- Submission file is an unsupported file type
- Submission text is in an unsupported language
- Qualifying text is either fewer than 300 words or more than 30,000 words
Only Analyze Long-form Content
Turnitin allows files with as few as 25 words to be submitted and may generate a score on files over 300 words. However, that doesn’t mean the result is valid or definitive.
Keep in mind that the less text included in the submission, the less likely it will be that either the Similarity Score or AI Writing Detection score will be valid. Also, complex documents with short paragraphs or repeated elements—like filled out forms—can produce invalid results.
Please consider submission length when interpreting Turnitin results.
How does Turnitin calculate this score?
Turnitin uses a computer model trained on large amounts of writing to look for statistical patterns in the writing it analyzes. It breaks each submission into overlapping text segments, to ensure each sentence is analyzed within the context of the surrounding text, and assigns a score indicating how likely it was the text was generated by AI. The score shown to instructors is an average of these segment scores.
A positive score suggests some percentage of the analyzed text was generated by AI. A 50% score does not mean exactly 50% of the text was generated by AI, that the text likely generated by AI is concurrent in the submission, or that an AI tool was intentionally used by the author. (A high score does not definitively prove AI was used, or if it was used, malicious intent.)
For more information about how Turnitin calculates an AI detection percentage, please visit Turnitin’s AI writing detection capabilities FAQs page, which covers the following topics:
- General Information:
- How do Turnitin’s AI writing detection capabilities work?
- AI detection results and interpretation
- Scope of detection
- Access and licensing
- AI paraphrasing:
- Product capabilities
- Detection of results & interpretation
- Access & licensing
- AI Bypasser Detection
- Product capabilities
- Detection of results & interpretation
- Access & licensing
How do I use Turnitin AI Writing Detection?
First, create a Canvas assignment and configure it to use analyze online submissions with Turnitin.
Once students submit to your assignment, you can review the Turnitin analysis for their submission by grading it in the Canvas SpeedGrader. Click the “Similarity Score” flag next to the Assessment grade entry box, which will open the Turnitin Similarity Report page. Look for the “AI” box on the bottom-right side of the screen in the vertical toolbar. Click it to open the AI Writing Report.
What do I do if I see a High AI Detection Score?
As an instructor, you are not required to use Turnitin to analyze your assignments or to act on the analysis shown in the tool.
If you’re concerned about the AI Detection Score shown on a student’s submission, consider meeting with the student to share Turnitin’s analysis, discuss how the student created their submission, and confirm whether further action is necessary. The student may be able to show you change tracking information from their word processing tool, or reveal their writing system includes AI functionality they weren’t aware of.
After discussing your concerns with the student, if you feel it’s necessary and helpful, you may begin the academic misconduct process. Please visit our campus policies page for a link to our Incident Report Form and a Academic Misconduct Guide for Instructors.
For more information about how to discuss this topic with your students, please contact the UW-Superior AI Faculty Coordinator.
Use Cases: When and When Not to use Turnitin
Turnitin is a useful tool, but is not appropriate for every assignment or compatible with every activity in Canvas.
When Not to Use Turnitin
- Short Writing Assignments: Turnitin requires submissions have a minimum number of words and was designed to analyze long-form text. Analysis of short submissions may not be valid.
- Discussions: Turnitin is only compatible with “Online” submission Canvas assignments.
- Poetry, Code, and Heavily Formatted Text: Turnitin should only be used to analyze what it calls “qualifying text”.
Qualifying text (prose sentences contained in long-form writing format) means individual sentences contained in paragraphs that make up a longer piece of written work, such as an essay, a dissertation, or an article, etc. The model does not reliably detect AI-generated text in the form of non-prose, such as poetry, scripts, or code, nor does it detect short-form/unconventional writing such as bullet points, tables, or annotated bibliographies. This means that a document containing several different writing types would result in a disparity between the percentage and the highlights.
When to Use Turnitin
The most common use for Turnitin is to analyze essays and thesis papers. Turnitin will be most useful for reviewing the main article, versus titles, annotations, bibliographies, and included graphics.
The Turnitin Similarity Report and AI Writing Report both include annotations and will allow you to review specific passages of text in more detail.
Need Help?
Please contact the UW-Superior Canvas Team.