Direct answer
The Turnitin AI detector is the AI writing detection layer built into Turnitin's similarity report. It does not search for copied text; it analyses the statistical patterns of your prose and returns a document-level percentage of qualifying text it believes was AI-generated, plus sentence-level highlights showing exactly which passages drove that number [1]. The percentage is a signal for instructor judgement, not a verdict, and Turnitin withholds a precise figure when its confidence falls below its own threshold [2].
How Does the Turnitin AI Detector Actually Work?
Turnitin's AI detector is a language model trained to recognise the statistical fingerprint of machine-generated prose rather than to match strings against a database. It looks at properties such as how predictable each next word is and how much sentence rhythm varies across a passage, because human academic writing tends to be less uniformly predictable than text produced by ChatGPT, Claude, or Gemini [2]. That is why the detector can flag a paragraph you wrote yourself in a very even, formulaic register, and why it can miss lightly edited AI text that a student has deliberately roughened.
Mechanically, the detector segments your document into sentences and produces a prediction for each one, then rolls those predictions up into a single document-level figure [1]. The overall percentage is calculated against the qualifying prose in the document, which means reference lists, bibliographies, and other excluded sections do not inflate or dilute the number [1]. The sentence-level layer is the part instructors actually read, because it tells them where the signal is concentrated instead of only how large it is.
Turnitin is explicit that the output is probabilistic and should be treated as one piece of evidence alongside a conversation with the student, draft history, and the assignment's own rules [2]. The model was also built with deliberate caution: when the system is not confident, it declines to present a hard number rather than guessing [2]. For a student, the practical takeaway is that the detector is a measurement of style, not a plagiarism database lookup, and it can be wrong in both directions.
What Does the AI Score Percentage in a Turnitin Report Mean?
The headline percentage is the share of your document's qualifying prose that the detector classifies as AI-generated — it is not the share of the whole file, and it is not a probability that you cheated [3]. Because reference lists and other excluded material are stripped out before the calculation, a 40% score on a 3,000-word essay means roughly two-fifths of the analysed prose carried the AI signal, which is usually a smaller absolute word count than students assume [3]. Reading the number without reading the highlights is the most common way students misjudge their own report.
The asterisk matters as much as the number. When Turnitin displays *% instead of an exact figure, the score sits below the detector's 20% confidence threshold, which means the signal is too weak or too ambiguous to report precisely [3]. Many students panic at an asterisk when it actually indicates a low-confidence, low-signal result rather than a high one. Conversely, a clean-looking low percentage with three heavily highlighted paragraphs deserves more attention than the summary line suggests.
The report is also advisory by design: Turnitin positions the AI indicator as information for educators, and institutions apply their own academic-integrity policies on top of it [3]. Two universities can receive identical reports and reach different conclusions, because the report is an input to a human process rather than an automated sanction. That framing is why the sentence highlights, not the percentage, are the best guide to what a marker will actually question.
Can You Check Your Own Draft With Turnitin Before You Submit?
In most cases, no — a student account cannot run a full Turnitin check on demand. Self-check and draft-submission folders only exist when an institution's administrator deliberately enables them for a module, so whether you can preview your own report depends entirely on your course setup rather than on your Turnitin login [4]. If no such assignment has been created, there is no button that produces an AI writing report for you.
There is a second reason to be careful with institutional drafts. Work submitted to a Turnitin assignment may be retained in the student paper repository depending on the settings your department chose, which can affect similarity results for later submissions of overlapping material [4]. A preview that quietly enrols your draft in a database is a poor trade for peace of mind, and it is not something most students can verify from the student side.
That gap is why independent preview services exist: they let you see the same style of AI and similarity report before the graded submission, so you can catch a false positive, rewrite a flagged paragraph, or confirm that a low-confidence asterisk really is low risk [4]. The value is in the timing — a flagged paragraph is a fixable paragraph a day before the deadline and an academic-integrity meeting afterwards. Previewing turns an unknown into a decision you can still act on.
If you would rather see the report before your marker does, turnitin0 gives you exactly that view: upload a .docx, .pdf, or .txt file over 300 words, and within roughly 15 minutes you receive the Turnitin AI detection report and the similarity report as two downloadable PDFs in one checkout — the same reports your professor sees in the LMS, generated without adding your file to Turnitin's student paper database. Over 100,000 reports have been delivered to more than 20,000 students across the US, UK, Canada, Australia, New Zealand, and Ireland, with no subscription and a single check at $3.80.
※ Turnitin0.com - Actual Turnitin AI Report Cover, Score, Flag And Similarity Summary
FAQ
Is the Turnitin AI detector the same as the similarity checker?
No. The similarity report matches your text against source databases, while the AI writing indicator is a separate layer that analyses writing patterns and appears alongside it in the same report [1]. A document can score low on similarity and still carry an AI flag, or the reverse.
Why does my report show an asterisk instead of a percentage?
An asterisk means the AI score falls below Turnitin's 20% confidence threshold, so the system reports *% rather than an exact figure [3]. It signals a low-confidence result, not a severe one.
Can the detector flag human writing as AI?
Yes. Because the model judges statistical predictability rather than authorship, very even or formulaic human prose can be flagged, which is why Turnitin frames the score as a signal for human review rather than proof [2]. Checking before submission is how you find out in advance.
Does checking my draft add it to Turnitin's student paper database?
Not with an independent preview service — turnitin0 checks your file without adding it to Turnitin's student paper database and does not share reports with third-party databases. Institutional draft folders are a different matter, since retention depends on your department's settings [4].
Do I need to fix a flagged paragraph, or just understand it?
Read the sentence highlights first. If the flagged passages are your own, formulaic writing, a rewrite in your natural voice usually resolves the signal; if they came from AI drafting, rewriting is the only reliable fix [3].