Direct answer
To check for AI and plagiarism online, you need two separate reports on the same file: an AI writing report that gives a percentage of qualifying prose flagged as AI-generated, and a similarity report that highlights matched text against published sources and previously submitted papers. Turnitin produces both, and its AI writing detection is a distinct report from its similarity score rather than a combined number [1]. A useful pre-submission check returns both PDFs from one upload, so you can read the exact flags your instructor will see before the deadline rather than guessing from a single blended score [1].
What should a pre-submission AI and plagiarism check actually show you?
A pre-submission check is only useful if it reproduces the artefacts your institution will look at. In Turnitin, the AI writing report states an overall percentage for the document, breaks that percentage down paragraph by paragraph so flagged passages are visible, and can be downloaded as a PDF for your own records [2]. The similarity report is generated independently, so a draft can show a clean similarity score and still carry a substantial AI percentage — which is precisely why checking only one of the two leaves you exposed [1].
The AI percentage is not a verdict on authorship. Turnitin analyses only prose that passes a minimum length threshold, so short quotations, headings, reference lists, and bullet fragments are excluded from the calculation [2]. That means the denominator is the qualifying prose in your document, not the word count of the whole file, and a document padded with tables or citations can post a higher percentage than its visible flagged text suggests [2].
What you should be reading, then, is the flagged text itself. Open the paragraph-level view, read each highlighted section, and ask whether that sentence genuinely came from a language model or from your own drafting process [2]. A number you cannot trace back to specific sentences is not actionable; a highlighted paragraph you can rewrite is.
Why can a draft score differently on two different checkers?
Because detectors are not measuring a property of the text — they are running a statistical model and reporting where the text falls relative to a threshold. Different vendors use different models, different training corpora, and different cut-offs, so the same paragraph can land on opposite sides of two systems' boundaries [3]. Turnitin's own documentation is explicit that detection output is a signal to be interpreted alongside context, not a definitive judgement of authorship [3].
The practical consequence is that a "second opinion" from an unrelated free checker tells you very little about what your institution will see. If your university runs Turnitin, a score from a different engine is measuring a different model [3]. What transfers between systems is the underlying writing: heavily templated phrasing, uniform sentence rhythm, and generic transitions are the patterns detectors tend to respond to, and those are fixable regardless of which engine flagged them [3].
There is also a confidence dimension that students routinely misread. When Turnitin's AI detection falls below its confidence threshold, the report displays *% instead of an exact figure — a low-confidence signal rather than a clean bill of health [1]. Treating *% as "0%" is one of the most common misreadings of the report, and it is worth confirming what a given result actually means before you relax [1].
How do you check a draft without adding it to Turnitin's student paper database?
This is the question that decides whether a check is safe to run at all. When a paper is submitted through an institutional Turnitin assignment, whether it is retained in the student paper repository depends on how that assignment is configured by the administrator [4]. Papers held in the repository become part of the comparison set for later submissions, which is why a draft submitted to the wrong place can create a match against itself later [4].
A non-repository check avoids that outcome: the file is compared against existing sources but is not added to the student paper database, and the resulting report is not shared onward to third-party databases [4]. For a draft you intend to revise and resubmit, that distinction matters more than the score itself, because a retained draft can produce a similarity hit against your own final version [4].
So the practical sequence is: confirm the service is non-repository, confirm you receive both the AI and similarity reports rather than one merged figure, and confirm you can delete the uploaded file afterwards [4]. If any of those three is unclear on a checker's site, the check is not worth the risk to your submission [4].
Reading a report you can trace to specific paragraphs is the point at which a check becomes useful — and that is what turnitin0 was built to deliver. Upload a .docx, .pdf, or .txt file (English, 300–30,000 words, under 20 MB) and you receive two downloadable PDFs in one order: the Turnitin AI detection report and the similarity report, in the same form your professor sees in the LMS. Checks run without adding your file to Turnitin's student paper database, most orders finish in 5–15 minutes, and 100,000+ reports have been delivered to 20,000+ students across the US, UK, Canada, Australia, New Zealand, and Ireland.
※ Turnitin0.com - Actual Turnitin AI Report Cover, Score, Flag And Similarity Summary
FAQ
Does an AI check also detect plagiarism?
No — they are separate reports. The AI writing report measures the proportion of qualifying prose that the model attributes to AI generation, while the similarity report matches your text against existing sources [1]. You need both to know where a draft stands.
What does *% mean on a Turnitin AI report?
It means the detected AI writing fell below Turnitin's confidence threshold, so an exact percentage is withheld and the report shows *% instead — a low-confidence signal, not a confirmed zero [1].
Can I check my draft without it being stored in the student paper database?
Yes, if the service runs a non-repository check: your file is compared against existing sources but is not added to the student paper database, and reports are not shared with third-party databases [4].
Why did my draft score high on AI detection when I wrote it myself?
Detectors respond to statistical patterns such as uniform sentence rhythm and templated transitions, and different models apply different thresholds, so false positives occur [3]. Read the paragraph-level flags and revise the specific passages rather than assuming the score is final.
Is a high similarity score automatically plagiarism?
No. Similarity measures matched text, including correctly cited quotations and common phrasing, so the number must be read against the highlighted sources rather than treated as a misconduct finding [1].