How Does Turnitin's AI Rate Detection Work?

Turnitin calculates your AI rate by breaking your paper into sentences, grouping them into overlapping segments, and assigning each segment a probability score between 0 and 1. Those scores are pooled and aggregated into a single percentage for the whole document. The model only analyzes prose, not code or bullet points. Only instructors and administrators can see the score, though they can download and share a PDF with you. If the score falls between 1% and 19%, the report shows an asterisk instead of a percentage.

The process starts when you submit your paper. Turnitin describes it this way: "When a paper is submitted to Turnitin, sentences from the submission are extracted and segmented into overlapping sections for prediction analysis. Each segment is classified by the AI detection model and given a value between 0 and 1, denoting the probability of the text being likely human or AI-generated. Each qualifying sentence within these segments inherits the segment's score. Since segments overlap, some sentences may have multiple scores, which are then pooled into a single score. These sentence scores are further aggregated and used to compute the overall document AI writing score."

There are several things to notice here. First, the unit of analysis is the sentence, not the whole document. Second, segments overlap, which means a single sentence can belong to multiple segments and carry multiple scores that get pooled. Third, the final percentage is an aggregation of all these sentence-level scores. It is not one judgment but many small judgments combined.

This is why two documents with the same percentage can have different distributions of flagged text. One might have a few heavily flagged sentences, while another has many lightly flagged ones. The aggregation smooths these differences into a single number.

Not everything in your document gets analyzed. Turnitin states: "The model does not reliably detect AI-generated text in the form of non-prose, or code, nor does it detect short-form/unconventional writing such as bullet points (short non-sentence structures)."

This means code blocks, bullet points, numbered lists, and other non-prose elements are excluded from the AI detection analysis. Only prose sentences, actual paragraphs of connected text, are fed into the model. Turnitin goes on to explain: "This means that a document containing several different writing types would result in a disparity between the percentage and the highlights."

If your paper mixes long prose paragraphs with bullet-point lists and code snippets, the percentage reflects only the prose portions. The highlights cover only the prose. This can create confusion if you compare the highlighted text to the total document length and the numbers do not match. The percentage is calculated against the qualifying prose, not the full document.

Students cannot see the AI writing detection indicator or report directly. Turnitin states: "The AI writing detection indicator and report are not visible to students." And: "Please note, only instructors and administrators are able to see the indicator."

This does not mean you can never see your own score. Turnitin adds: "However, with the PDF download feature, instructors can download and share the AI report with students." So the pathway to seeing your score goes through your instructor. If your instructor chooses to download the PDF and share it with you, you can see the percentage, the highlights, and any asterisk. If they choose not to share it, you will not see it through the Turnitin interface. Turnitin login for students goes through what the student view does and does not contain.

If you need to see your report for revision purposes, ask your instructor if they can download and share the PDF. Many instructors are willing to do this, especially if you explain that you want to address any flagged passages before the final submission.

If the AI score falls between 1% and 19%, the report does not show a percentage. Instead, it shows an asterisk (%). Turnitin explains: "To avoid potential incidence of false positives, no score or highlights are attributed for AI detection scores in the 1% to 19% range. When AI is detected below the 20% threshold in the report, it is now indicated with an asterisk (%) and no percentage is attributed."

The asterisk means the detector found something, but not enough to commit to a specific number. No highlights appear in this range either. The score is too low for Turnitin to justify pointing to specific passages as AI-generated. This is a conservative approach designed to reduce false positives.

If you see the asterisk, the detector did pick up on some AI-like patterns in your prose. But the confidence was below the threshold where Turnitin is willing to assign a firm percentage. For most students, the asterisk is not a cause for alarm. What the asterisk (*%) means on a Turnitin AI score covers what to say if an instructor asks about it.

If your instructor shares the PDF report with you and you see a percentage above 20%, look at the highlighted passages. Those are the specific segments the model classified as likely AI-generated. Focus your revision on those sections, in the order rewriting only the paragraphs a Turnitin report flagged sets out. Keep your citations, tables, and code blocks intact, since those are not part of the AI analysis.

If you see the asterisk, there are no highlighted passages to address. The score is below the reporting threshold. You can still review your prose for generic or formulaic phrasing, but there is no specific flag to rebut.

If your percentage is high enough to concern you or your instructor, revise the flagged passages and re-submit. Remember that the model updates over time, so a re-submission may produce a different score even with the same text, which is why comparing two reports beats comparing two scores.

Eligible passages can be re-run at no charge, so you can refine flagged sections without additional cost until the percentage reaches an acceptable level.