Direct answer
Turnitin detects two main things in a submitted document: text similarity (matches against its repository of internet content, publications, and previously submitted student papers) and AI-generated writing (patterns typical of large language models). Its similarity report highlights matching text and assigns a percentage, while its AI writing indicator estimates how much of the text looks machine-generated [1]. A high match or AI percentage is a signal for review, not automatic proof of misconduct.
What does Turnitin's similarity report actually flag in a paper?
Turnitin's similarity report compares your submission against a large repository of internet pages, academic publications, and student papers already in its database [2]. When it finds overlapping text, it highlights the passage and lists the matched source, breaking the overall percentage down by source type so you can see where the matches came from [2].
Crucially, a match is not the same as plagiarism. Properly quoted and cited material still registers as a match unless your instructor chooses to exclude quotes or the bibliography, and small matches can be filtered out entirely [2]. A high similarity score can simply mean you used a lot of direct quotations or a common phrase.
That is why the report is designed to be read, not just scored. Instructors are expected to review each highlighted passage in context before drawing any conclusion [2]. If you understand which passages matched and why, you can usually tell whether the score reflects a real problem or just heavy citation.
Can Turnitin detect AI-generated writing, and how does the AI score work?
Yes. Turnitin's AI writing detection analyzes your text for statistical patterns that are typical of large language models, rather than comparing it against a source database [3]. It returns an overall AI percentage and a breakdown showing which segments look AI-generated versus human-written [3].
The indicator works on probability, not certainty. Turnitin itself notes the detector can produce false positives and false negatives, and that the result should be treated as one signal among several rather than definitive proof [3]. Short passages, heavily edited text, and technical or formulaic writing can all skew the estimate.
Because of that uncertainty, the AI score is best read alongside the similarity report and your own writing process. If you drafted with ChatGPT, Claude, or Gemini and then edited, the flagged segments are the ones worth reviewing first [3]. Knowing this before you submit gives you a chance to revise the specific passages that triggered the flag.
How can you see your own Turnitin report before you submit?
Many students only see a Turnitin report after the graded deadline, which is the worst possible time to discover a flagged passage. A pre-submission check lets you preview both the similarity score and the AI writing indicator on your own draft first [4]. That way you can fix accidental matches or AI-flagged sections while you still have time to revise.
Reviewing your report early also teaches you how the tool reads your writing. You can see which sources matched, how much of the text was flagged as AI, and whether your citations are being counted as matches [4]. That feedback loop is far more useful than guessing what an instructor will see.
The practical takeaway is simple: treat the report as a diagnostic, not a verdict. Check it, read the highlighted passages, and revise anything that genuinely needs work before the final submission [4].
Understanding what Turnitin detects is only half the battle — the real advantage is seeing your own report before your professor does. turnitin0 lets you preview the exact Turnitin AI and similarity reports on your draft, so you can catch flagged passages while there is still time to fix them.
※ Turnitin0.com - Actual Turnitin AI Report Cover, Score, Flag And Similarity Summary
FAQ
Does Turnitin detect AI even if I edited the text?
Yes. The AI detector analyzes statistical patterns in the final text, so editing does not automatically remove the signal [3]. If you drafted with an AI model and then revised, the remaining flagged segments are the ones worth checking.
Is a high similarity score automatically plagiarism?
No. Matches include properly quoted and cited material, and instructors are expected to review each highlighted passage in context [2]. A high score often just reflects heavy use of quotations or common phrasing.
What sources does Turnitin compare my paper against?
Turnitin checks your text against internet content, academic publications, and student papers already in its database, and the report breaks the percentage down by source type [2].
Can I see my Turnitin report before I submit?
Yes — a pre-submission check lets you preview both the similarity score and the AI writing indicator on your draft before the graded deadline [4].
Does Turnitin detect AI writing with 100% accuracy?
No. Turnitin states the AI indicator can produce false positives and false negatives and should be read as one signal rather than definitive proof [3].