Direct answer
A Turnitin AI detector reads a thesis sentence by sentence and reports the share of qualifying text it classifies as AI-generated, so the result is a probabilistic signal about your draft rather than a verdict on your authorship [1]. Because the AI Writing Report is only issued through institutional accounts, most thesis writers first meet their score in an instructor's dashboard or in a pre-submission check [1]. Knowing how the percentage is calculated — and what the asterisk display means — is what lets you act on the result instead of panicking at it [3].
How does Turnitin's AI detector actually decide whether a thesis is AI-generated?
Turnitin's detector does not look for "AI words." It evaluates long-form prose in segments and assigns each segment a likelihood that the text was machine-generated, then aggregates those judgements into a single percentage of qualifying text [2]. That design matters for a thesis, because a 12,000-word dissertation gives the model far more continuous prose to analyse than a two-page assignment, and the report exposes the flagged segments so a reader can open them and see exactly which sentences triggered the classification [2].
The output is deliberately framed as an indicator, not proof. Turnitin describes its AI writing detection as probabilistic, which means a high percentage is a prompt to examine the text, not a finding of misconduct, and a low percentage does not certify that every sentence was typed by hand [1]. In practice, supervisors read the report next to the similarity report, because the two answer different questions: one asks "does this match existing sources," the other asks "does this read as machine-generated" [2].
There is also a structural limit worth internalising. The detector only scores text that qualifies for analysis, so quotations, reference lists, and heavily formatted material are excluded from the calculation [2]. A thesis with a long literature review in your own words and a short methods chapter drafted with a chatbot can therefore produce a percentage that feels disconnected from how much of the document you actually wrote — which is precisely why the segment view, not the headline number, is the part worth studying [1][2].
What do the AI score and the *% display in a thesis report actually mean?
The headline number is a percentage of qualifying text, not a percentage of your effort and not a probability that you cheated [2]. If a report shows 34%, the defensible reading is that roughly a third of the analysed prose was classified as likely AI-generated, and the next step is to open the flagged segments and check whether those passages correspond to text you ran through a writing assistant [2]. Treating the number as a grade is the single most common misreading among thesis writers.
The asterisk is the detail that trips people up. When detection confidence falls below Turnitin's 20% threshold, the report displays an asterisk in place of an exact figure, because the signal is too weak to express as a reliable percentage [3]. A *% is therefore a low-confidence indicator — it is not a clean bill of health, and it is not an accusation; it simply means the detector could not make a confident call on that text [3].
Two practical consequences follow. First, an asterisk result on a methods chapter is not something to celebrate or to argue about, because there is no firm number to argue over [3]. Second, a mid-range percentage deserves the same segment-level scrutiny as a high one: the flagged spans tell you which paragraphs to rewrite, and the unflagged spans tell you which chapters are already safe [2]. Reading the report this way turns a frightening number into an editing list.
Can you check a thesis for AI detection before you submit it to your university?
Students cannot generate an official Turnitin AI Writing Report themselves, because the report is delivered through the institution's account and course structure rather than to individual writers [1]. That does not leave you blind, however. Turnitin's own student-facing guidance encourages writers to understand how detection works and to treat the report as a starting point for a conversation with their instructor rather than as a hidden verdict [4]. The practical gap students face is previewing the result before the irreversible upload.
That gap is exactly what a pre-submission check fills. You upload the same file you intend to submit — a.docx,.pdf, or.txt thesis between 300 and 30,000 words — and receive the AI detection report and the similarity report as two downloadable PDFs, matching what a supervisor sees in the institutional system. Because the check is non-repository, the file is not added to Turnitin's student paper database, so a preview does not create the very match it is meant to detect. Turnaround is under 15 minutes in 98% of cases, with delivery guaranteed within 30 minutes even during queue spikes.
For a thesis specifically, the timing of the preview matters more than the price. A 20,000-word dissertation is not something you can rewrite in an evening, so seeing the flagged segments a week before the deadline changes what you can still fix: you can rework the methods chapter, add your own analysis to a literature review, or humanize the passages you drafted with ChatGPT, Claude, or Gemini. Seeing the same report an hour before the portal closes changes nothing except your blood pressure.
If you would rather see your thesis report before your supervisor does, turnitin0 gives you that preview on your own terms — the same AI score, the same flagged segments, the same similarity matches, delivered as two PDFs you can open and act on while there is still time to revise.
※ Turnitin0.com - Actual Turnitin AI Report Cover, Score, Flag And Similarity Summary
FAQ
Is a high Turnitin AI score proof that a thesis was written by AI?
No. Turnitin presents AI writing detection as a probabilistic indicator, so a high percentage means segments were classified as likely machine-generated and should be reviewed, not that misconduct has been established [1].
Why does my thesis report show an asterisk instead of a percentage?
The asterisk appears when detection confidence falls below Turnitin's 20% threshold, which is a low-confidence signal rather than a pass or a fail [3].
Does the AI percentage include my quotations and reference list?
The percentage is calculated over qualifying text only, so quoted material and reference lists are excluded from the calculation [2].
Can I get an official Turnitin AI report as a student?
Official reports are issued through institutional accounts, which is why students typically preview their draft with a pre-submission check before the university upload [1][4].
Will checking my thesis first add it to Turnitin's student paper database?
A non-repository check does not add the file to Turnitin's student paper database, so the preview does not create a future similarity match against your own draft.