Direct answer
The Turnitin AI writing report is a separate indicator inside Turnitin Feedback Studio that estimates the percentage of a submitted document likely written by generative AI tools, and it highlights the specific sentences the detection model suspects [1]. It is not the same as the Similarity Report: similarity checking looks for copied or unreferenced text, while AI writing detection evaluates how a document was generated [1]. This guide explains what the report actually shows, how to read the AI percentage, and how students can see the same report their instructor would receive before they ever hit submit.
What Does the Turnitin AI Writing Report Show and How Does It Detect AI-Generated Text?
The report lives in its own AI writing tab that appears next to the Similarity Report when a submission is processed with AI writing detection enabled. It opens with an overall indicator: a percentage that reflects the share of the document Turnitin believes was produced by an AI text generator rather than written by a person [2]. In our experience checking drafts for thousands of students, this overall figure is what most people remember, but it is only the top layer of the report.
Below the overall score, the report provides sentence-level evidence. Turnitin's model analyzes the text in segments and marks the sentences it attributes to AI, so an instructor can click a highlighted area and see exactly which passages were flagged [2]. This granular view matters because a paper can carry a low overall percentage yet still contain clearly AI-written paragraphs that deserve attention before submission [2].
Detection works by comparing a submission against patterns found in academic writing and in text produced by modern large language models. The model is trained specifically on academic prose, and Turnitin designed it to recognize text generated by tools such as ChatGPT rather than to judge grammar or detect plagiarism [2]. That distinction explains why the AI writing report and the similarity report regularly disagree: one measures copied text, and the other measures machine-generated text.
No detector is perfect, and knowing the limitations is part of reading the report correctly. Turnitin's own guidance notes that results are less reliable for very short documents, heavily edited or paraphrased AI text, and writing in languages other than English [2]. For students, the practical takeaway is that a clean score does not replace honest drafting, and a flagged score should be treated as a signal to review, not as a verdict on its own.
What AI Percentage on a Turnitin Writing Report Means a Paper Is Flagged as AI-Written?
Turnitin does not publish a single percentage that automatically means a paper is "flagged" or penalized, and instructors are explicitly told not to treat the score as proof of misconduct on its own [3]. When a student opens the AI writing indicator in their submission feedback, they see the same elements their instructor sees: the overall percentage, the sentence highlights, and the asterisk-style display used for very low scores [3].
The most common source of confusion is the asterisk bucket. Any AI writing score below 20% is displayed as % rather than as a specific single-digit number, because the detection model has lower confidence in that range; the only explicit low numeric outcome students typically see is 0%, which means no AI-written text was detected [3]. A score of % therefore tells you the model found little or no AI text, but it deliberately avoids implying a false level of precision.
For higher scores, the percentage reflects the estimated share of the document written by AI, and Turnitin reports a very low false-positive rate on entirely human-written academic documents in its published testing [1]. Even so, official guidance stresses that the report should be a starting point for discussion between student and instructor rather than an automated accusation [1]. In short, treat percentages below 20% as essentially clean, and treat higher percentages as an invitation to rewrite the flagged passages in your own voice.
How Can Students Check Their Own Turnitin AI Writing Report Before Submitting an Assignment?
Turnitin only generates the AI writing report after a file has been submitted to an assignment whose instructor has enabled AI writing detection; there is no official Turnitin self-check box where students can test a draft beforehand [4]. That means, inside Turnitin itself, the first time you see your own AI writing report is only after your paper is already sitting in the instructor's queue [4]. In practice, this is exactly why students who care about their score check before submission through a service that runs the real Turnitin engine.
When AI writing detection is enabled in an assignment, every submission is processed and the report appears automatically next to the similarity results [4]. Because the same engine is used regardless of who triggers the check, a pre-submission scan can reproduce the indicator you will later see in Feedback Studio: the overall percentage, the *% bucket for sub-20% results, and the sentence-level flags [4]. Uploading a.docx,.pdf, or.txt file and receiving that report in minutes gives you a realistic preview of what your instructor will see.
The practical workflow we recommend to students is simple: run a real Turnitin AI and similarity check on your near-final draft, review the flagged percentage and highlighted sentences, revise the AI-written sections in your own words, and re-check once more before submitting. Reports are generated per submission, so a fresh scan reflects your edits rather than an old version of the file [4]. Checking early also protects your academic record, since a flagged report is far easier to address before it is attached to your name than after a misconduct conversation has started.
Want to see the exact report your professor would receive before you press submit? Turnitin0 runs the genuine Turnitin AI writing engine and returns the real AI writing report and similarity summary on your own draft, so you can read your percentage, review every highlighted sentence, and decide whether to revise before your paper reaches your institution. Thousands of students worldwide use turnitin0 precisely for this pre-submission confidence check.
※ Turnitin0.com - Actual Turnitin AI Report Cover, Score, Flag And Similarity Summary
FAQ
Is the Turnitin AI writing report the same as the similarity report? No. The similarity report checks for copied or unreferenced text, while the AI writing report estimates how much of the document was likely generated by AI and highlights the specific sentences it suspects [1]. A document can score low on one report and high on the other [1].
What does *% mean on a Turnitin AI writing report? Any AI writing score below 20% is displayed as *% because the detection model has lower confidence in that range; 0% is the only explicit low numeric outcome students typically see [1]. An asterisk result should be read as "little or no AI text detected," not as a specific single-digit percentage [3].
Can instructors see which sentences were flagged as AI-written? Yes. The AI writing report provides sentence-level highlights, so an instructor can click a flagged region and see the exact passages the model attributes to AI [2]. This is why reviewing your own report before submission is so valuable: you can see precisely which paragraphs to rewrite [2].
Can students check their AI writing score inside Turnitin before submitting? No. Turnitin generates the report only after a file is submitted to an assignment with AI writing detection enabled by the instructor, and there is no student self-check feature [4]. To preview your score beforehand, you need a pre-submission check that runs the same Turnitin engine [4].
Do all submissions automatically get an AI writing report? No. The report appears only when the instructor has enabled AI writing detection for the assignment [1]. Results are also less reliable for very short documents, so an indicator may not be generated for every submission [2].