Direct answer
Detecting plagiarism and AI-generated text in one pass means running a single pre-submission check that returns two separate reports: a similarity report for matched source text and an AI writing report for machine-patterned prose. Turnitin produces both from the same upload, so you do not need two tools or two submissions [1]. The two scores answer different questions and are read in a fixed order — similarity first, AI writing second [1]. This guide explains how to run the check, what each number actually measures, and how to interpret the two reports together before you submit.
How can I detect plagiarism and AI-generated text in my paper before I submit it?
The practical answer is to run a non-repository pre-submission check on your draft, which returns a similarity report and an AI writing report as two downloadable PDFs from one upload [2]. The similarity report highlights every passage that matches text in Turnitin's comparison repository — billions of web pages, journals, and previously submitted student papers — and links each highlight to its source so you can inspect the overlap yourself [2]. Because the check is non-repository, your draft is not added to the student paper database, so you can iterate on the same file without inflating your own match percentage [2].
Reading similarity correctly matters more than the raw number. Turnitin's similarity percentage is the proportion of your document that matched something in the repository; it is not a plagiarism verdict, and a high score can come entirely from correctly quoted and cited material [2]. Matches are colour-coded by source type, and quotes, bibliography entries, and matches under a small word threshold can be excluded from the calculation, which is why two versions of the same essay can show very different scores [2]. Check the match list before you panic: if every highlight sits inside quotation marks or a reference list, the score is doing its job, not accusing you [2].
For AI-generated text, the same upload also produces an AI writing report that estimates how much of your qualifying prose reads as machine-written [1]. Turnitin only analyses prose segments above a minimum length, so headings, references, and short strings are excluded from the calculation [1]. The report is designed to be read alongside the similarity report in the instructor's LMS view, which means the preview you run is structurally the same artefact your marker will see [1]. That symmetry is the whole point of checking early: you see the flags before they become a conversation.
What is the difference between a Turnitin similarity score and a Turnitin AI writing score?
The similarity score measures text overlap with existing sources, while the AI writing score measures stylistic patterns associated with generative models — they are computed independently and can move in opposite directions [3]. A paper can score 30% similarity with a 0% AI writing indicator if it quotes heavily and cites properly, and a paper can score 2% similarity with a high AI indicator if it is original but written in the even, predictable cadence that detectors associate with large language models [3]. Treating the two numbers as one blended "cheating score" is the most common misreading students make [3].
The AI writing report is also explicitly probabilistic, and Turnitin states that its detector can produce false positives, particularly on highly formulaic academic prose, heavily edited drafts, or text written in a second language [3]. Because of that uncertainty, Turnitin displays an asterisk (*) instead of an exact percentage when its AI detection confidence falls below its internal threshold — a *% is a low-confidence signal, not a clean bill of health [3]. Only prose segments over a minimum length are scored, so a report can show a low percentage simply because most of the document fell outside the analysed segments [3].
The similarity side has its own interpretive traps. The score is a share of matched text, not a judgement, and Turnitin's own guidance stresses that the number must be read together with the match list and the exclusion settings in force [3]. A bibliography-only match, a correctly attributed block quote, and a genuinely copied paragraph can all produce the same percentage on the cover page [3]. The difference between the two scores is therefore not just what they measure but how much human judgement each one requires before it means anything.
How do I read a Turnitin AI writing report and similarity report together?
Read them in sequence, not in parallel: start with the similarity report to clear source overlap, then move to the AI writing report to assess drafting patterns [4]. If the similarity match list is fully explained by quotations and citations, you can set that question aside and focus entirely on the AI indicator [4]. If the similarity report shows unexplained matches in your own prose, fix those first — rewriting matched passages will also change the AI writing signal, so resolving them in the wrong order wastes a check [4].
When the two reports disagree, the disagreement itself is the diagnostic. A high similarity score paired with a 0% AI indicator usually describes a well-sourced paper that leans on the literature, which is exactly what academic writing is supposed to do [4]. A low similarity score paired with a high AI indicator describes original phrasing that still carries machine-like regularity — the pattern you get when a draft is generated and then lightly edited rather than genuinely rewritten [4]. Neither pattern is automatically misconduct, but each points to a different revision task.
Finally, check the version you are reading. Similarity and AI writing results are tied to the specific file submitted, and editing the document after the check invalidates the preview for both reports [4]. If you revise, re-run the check rather than assuming the old numbers still apply, and keep both PDFs so you can compare the before and after [4]. The habit that separates a calm submission from a stressful one is treating the two reports as a revision loop, not a one-time verdict.
Reading both reports is only useful if you are reading the real thing. turnitin0 produces the same Turnitin AI and similarity reports your marker sees in the LMS, delivered as two downloadable PDFs from a single non-repository upload — so the flags, score bands, and match highlights you review are the ones that will appear on your submission. Students who preview both reports before the deadline revise with evidence instead of guesswork.
※ Turnitin0.com - Actual Turnitin AI Report Cover, Score, Flag And Similarity Summary
Get [Real Turnitin](https://www.turnitin0.com/) AI & Similarity Report
FAQ
Does one check really return both a plagiarism and an AI report?
Yes. Turnitin's AI writing detection is integrated into the existing similarity workflow, so a single submission generates both reports [1]. You download them as separate PDFs from the same order.
Can a high similarity score alone mean I plagiarised?
No. Similarity measures matched text, not intent, and correctly quoted or cited passages raise the score legitimately [2]. Always open the match list and check whether the highlighted text is attributed before drawing conclusions [3].
Why does my AI writing score show an asterisk instead of a number?
Turnitin displays *% when its AI detection confidence falls below its internal threshold, which signals a low-confidence result rather than a confirmed finding [3]. It is worth re-checking after you revise the flagged sections.
Should I fix the similarity score or the AI score first?
Clear unexplained similarity matches first, because rewriting those passages also shifts the AI writing signal [4]. Once source overlap is explained, revise for drafting patterns and re-run the check.
If I edit my paper after checking, is the old report still valid?
No. Both reports are tied to the exact file submitted, so any edit invalidates the preview for similarity and AI writing alike [4]. Re-run the check and keep both versions for comparison.