How to Use Turnitin AI Detector: What the Score Means and What to Do With It

What Is Turnitin's AI Writing Indicator?

Turnitin's AI Writing Indicator is a feature built into the Turnitin Feedback Studio submission interface. When a document is submitted through an assignment that has AI detection enabled, the platform runs a statistical analysis to estimate how much of the text was generated by a large language model. The indicator reports a percentage — the proportion of sentences in the submission that the model classified as likely AI-generated — along with color-coded highlighting that marks those sentences directly in the document viewer. Turnitin's underlying model analyzes two main signals: perplexity (how predictable each word choice is given the surrounding context) and burstiness (how much sentence length and structure vary across the document). AI-generated text tends to score low on both metrics because language models produce statistically smooth, consistently structured output. Unlike plagiarism detection, which compares text against a database of existing sources, the AI detection model runs entirely against the statistical properties of the submitted text itself — no match to a specific AI tool or prompt is made. The feature is available on Turnitin accounts that have purchased the AI writing detection add-on. Availability varies by institution — some universities have enabled it across all departments, while others restrict it to specific courses or writing-intensive programs. Instructors typically need to enable the feature per-assignment rather than having it run automatically on every submission, so students should not assume every Turnitin submission is being analyzed for AI writing. The AI Writing Indicator became a significant part of academic integrity workflows from mid-2023 onward, and most faculty who use Turnitin regularly will have encountered it. If you are a student unsure whether your submission was analyzed for AI content, the simplest approach is to ask your instructor directly.

Turnitin reports that its AI detection model has a false positive rate below 1% at the 20% threshold — meaning fewer than 1 in 100 entirely human-written documents should score 20% or higher under typical conditions.

How to Use Turnitin AI Detector as an Instructor

Instructors control whether AI detection is active for each assignment. If the AI detection toggle is not turned on in the assignment settings, no AI report will be generated even if the institution has a Turnitin license. The steps below apply to assignments created through Canvas, Blackboard, Moodle, or the Turnitin.com portal directly.

1. Log in to your LMS (Canvas, Blackboard, or Moodle) or go directly to turnitin.com and open your course dashboard.
2. Open an existing assignment or create a new one, then navigate to the Optional Settings or More Options panel.
3. Enable the AI Writing Detection toggle — this setting is off by default in some institutional configurations, so confirm it is active before the submission deadline.
4. Once submissions arrive, open the assignment inbox and click a student submission to open the Feedback Studio viewer.
5. Look for the AI Writing Indicator badge in the report sidebar — it appears next to the originality (similarity) score and displays the AI percentage.
6. Click the AI indicator badge to expand the full AI detection report, which shows the overall percentage and highlights individual sentences across the document.
7. Review the sentence-level highlighting to see which passages contributed most to the score before initiating any conversation with the student.
8. Download the report using the print or export button if you need a record for your institution's academic integrity file.

How Students Can View Their Own Turnitin AI Score

Whether students can see their own AI detection score depends on the settings their instructor and institution have configured. Some institutions grant students full view access to the AI report; others display only the similarity percentage and restrict the AI indicator to instructors only. This is a deliberate institutional choice — some administrators believe that showing students the score encourages them to game the system, while others treat transparency as the better policy for building trust. If your institution allows it, you can access your own AI score through the following steps.

1. Log in to your LMS account and navigate to the course where you submitted the assignment.
2. Open your submission through the assignment page — in Canvas this is under Submission Details; in Blackboard it is under My Grades.
3. If Turnitin Feedback Studio view is available, click View Submission or the GradeMark icon to open the full report.
4. Look for the AI Writing Indicator badge in the right sidebar of the document viewer.
5. If the badge is visible, click it to see the percentage breakdown and which specific sentences were highlighted as AI-generated.
6. If you do not see the AI indicator at all, your instructor may have restricted student access to that report — contact your instructor or check your institution's Turnitin documentation to confirm.

How to Read the Turnitin AI Detection Report

Once you know how to use Turnitin AI detector to open the report, the next step is understanding what the numbers actually mean. The overall score is a sentence-level proportion, not a document-level probability — which is a crucial distinction for both instructors and students evaluating the results. A common misconception is that a 50% score means half the document was directly copied from an AI tool. In reality, it means half the sentences in the document exhibited statistical patterns that the model associates with AI-generated text. Those patterns can be present in human writing for many reasons. The highlighted sentences in the report are where instructors and students should focus attention first — not the overall percentage.

1. The overall percentage at the top of the AI report represents the share of sentences the model flagged as likely AI-generated — a 35% score means roughly 35% of the document's sentences triggered the classification.
2. Yellow and orange highlighted sentences in the document viewer are those Turnitin classified as AI-generated; unhighlighted sentences were not flagged.
3. Hover over a highlighted sentence to see any additional context Turnitin provides about that passage.
4. A score below 20% is treated as inconclusive by Turnitin's own guidelines — most institutions consider this range low-risk and do not escalate it.
5. Scores between 20% and 40% typically warrant a conversation between instructor and student, not immediate disciplinary action.
6. Scores above 40% may trigger a formal review under some institutional policies, though exact thresholds vary — check your institution's academic integrity policy for the specific numbers they use.
7. The AI report does not identify which AI tool generated the text, nor does it link the submission to any specific prompt or session.

A 25% AI score means roughly one sentence in four showed the statistical patterns Turnitin associates with AI output — not that a quarter of the document was copied from ChatGPT or any other tool.

What Turnitin AI Detector Cannot Tell You

Knowing how to use Turnitin AI detector also means understanding where the tool's analysis ends. The report describes statistical patterns in a final submitted document, not the process that produced it. Several categories of entirely human-written text regularly produce high AI scores without any AI involvement, and this is important context before any decision is made based on the score. Turnitin itself acknowledges a non-zero false positive rate and recommends that instructors never rely on the score alone. The tool cannot determine authorship intent, cannot distinguish between a student who accidentally adopted an AI-like writing style and one who submitted machine-generated text, and cannot see any part of the writing process that happened before the final document was submitted. When scores are used as evidence in an academic integrity case, the burden of interpretation falls on the institution, not the algorithm.

1. Non-native English speakers who write in careful, grammatically simple sentences often score high because their prose lacks the idiosyncratic variation that marks native fluency.
2. Technical and scientific writing in constrained fields — lab reports, case studies, legal analysis — uses narrow vocabulary and repetitive sentence structures that can resemble AI output patterns.
3. Heavily revised text that has been edited for clarity and concision may have had its natural variation smoothed out, producing a more uniform statistical profile than the original draft.
4. Short submissions under 300 words have reduced detection accuracy, which Turnitin explicitly acknowledges in its product documentation.
5. Text translated from another language may score high if the translation is very literal and consistent in structure.
6. Formal academic prose in any discipline — where style conventions constrain vocabulary and sentence form — tends to score higher than informal writing regardless of origin.

Turnitin's own documentation states: 'The AI writing detection capabilities are not designed to be the sole basis for academic misconduct cases and should be used as part of a holistic review process.'

What to Do When a Submission Gets Flagged

A high AI score from Turnitin is a signal to investigate further, not a verdict. Both instructors and students benefit from knowing the appropriate next steps when the Turnitin AI detector returns a notable result. Turnitin itself recommends that educators treat the indicator as a starting point for conversation rather than a standalone finding. The most important thing to know is that the AI report does not automatically trigger any institutional process — what happens next depends entirely on the instructor and the institution's academic integrity policy. Many institutions have published internal guidelines specifying at what score level an inquiry begins, what evidence a student may present, and how the appeal process works. If you are a student facing a flag, locating that policy document before your first conversation with your instructor puts you in a much better position. If you are an instructor reviewing a score, Turnitin's own training materials recommend asking the student to describe their writing process before framing the conversation around misconduct.

1. If you are an instructor, request a meeting with the student before taking any formal action — ask them to walk you through their writing process, their sources, and the hardest part of the assignment.
2. Ask the student for version history or draft evidence: Google Docs version history, Word AutoSave records, or timestamped cloud backup files are all useful.
3. If you are a student who has been flagged, gather documentation of your writing process immediately — draft versions, outline notes, research materials, and browser history showing your source reading.
4. Review whether the flagged sentences correspond to sections written in a particularly uniform or formal register — this context helps explain the score without requiring accusation or denial.
5. Check your institution's published policy on AI detection thresholds — many institutions specify at what score level a formal review is triggered and what evidence can be presented.
6. If the matter escalates, consult your institution's academic integrity office for the formal appeal process and what supporting materials are accepted.

Turnitin's guidance for instructors states that the AI Writing Indicator 'should be used alongside other evidence and not as the sole determinant' in any academic integrity case.

Check Your Writing Before It Reaches Turnitin

Running your text through a second AI detector before submission lets you see which sentences are most likely to trigger a flag, giving you time to revise or prepare context. This is especially useful for students who write in formal registers, non-native English speakers, and anyone submitting technical or scientific work — writing styles that face higher false positive rates with any AI detection tool, including Turnitin. NotGPT's AI Text Detection tool shows a sentence-by-sentence probability score with highlighted passages — similar to how Turnitin presents results — so you can identify which parts of your own writing read as AI-generated and address them before formal review. If specific sections consistently score high, the Humanize feature can adjust phrasing to sound more naturally human without changing your core argument. A pre-submission check does not guarantee a low Turnitin score, but it gives you a concrete picture of where your text sits statistically and enough time to act on it.