What is the Turnitin AI Writing Report?
Table of Contents
- How Does the Turnitin AI Writing Report Detect AI-Generated Content?
- What Do the Different AI Score Percentages Mean in a Turnitin Report?
- Can Students Check Their Own Work With a Turnitin AI Writing Report Before Submitting?
- FAQ
- Sources
- Related articles
Direct Answer — The Turnitin AI Writing Report is a built-in feature within Turnitin's Similarity Report that detects text potentially generated by artificial intelligence tools such as ChatGPT, Claude, Gemini, and other large language models. It displays an overall percentage indicating how much of a submission may have been AI-generated and highlights specific text segments for instructor review. Introduced to help educators uphold academic integrity, the report is a separate, independent indicator from the similarity (plagiarism) score and is designed to provide data for informed human judgment rather than serve as a standalone determination of misconduct. [1]
How Does the Turnitin AI Writing Report Detect AI-Generated Content?
Turnitin's AI detection model works by analyzing text at a granular, sentence-by-sentence level. When a paper is submitted, the system first breaks the document into segments of roughly a few hundred words (approximately five to ten sentences). These segments are then overlapped with each other to ensure each sentence is evaluated in its full surrounding context. Each segment is run through Turnitin's AI detection model, which assigns a score between 0 and 1 to every sentence — 0 indicates the sentence was likely written by a human, while 1 indicates it was likely generated by an AI tool. The average of all segment scores produces the final overall percentage shown in the report. [2]
The technology behind this detection is rooted in the statistical differences between human and AI writing patterns. AI language models like GPT and ChatGPT are trained on massive text corpora and generate sequences by picking the next most probable word in a consistent fashion. Human writing, by contrast, tends to be more inconsistent and idiosyncratic, resulting in lower probability word choices. Turnitin's classifiers are trained to identify these differences in word probability patterns. [2] The first iteration of the model was trained to detect GPT-3, GPT-3.5, and variants including ChatGPT. Since launch, Turnitin has expanded detection to include GPT-4, GPT-4o, GPT-5, Gemini, Claude, LLaMA, and other major large language models, with ongoing updates to keep pace with emerging AI tools. [1]
It is important to note that the model detects qualifying text — prose sentences in a long-form writing format such as essays, dissertations, or articles. Non-prose formats such as poetry, scripts, code, bullet points, and annotated bibliographies are not reliably detected, which can cause a disparity between the percentage shown and the highlighted text in certain submissions. [2] The model was also trained on a representative sample that includes both AI-generated and authentic academic writing across geographies and subject areas, with particular attention to statistically under-represented groups such as second-language learners, to minimize bias. [1]
What Do the Different AI Score Percentages Mean in a Turnitin Report?
The percentage displayed at the top of the AI Writing Report indicates the amount of qualifying text within the submission that Turnitin's model predicts may have been generated by AI, AI-paraphrased, or AI-bypassed. A score of 0% means that after processing, Turnitin's AI detection did not identify any qualifying text as likely AI-generated. For scores between 0% and 20%, the AI indicator displays an asterisk (*%) rather than a numerical percentage. Turnitin implemented this threshold specifically because testing found a higher incidence of false positives in this range, and the asterisk is designed to call attention to the fact that the score is less reliable. [2]
For scores between 20% and 100%, the full numerical percentage is displayed, and the report provides a detailed Submission Breakdown with two detection categories. The first category, "AI-generated only," highlights text in cyan that was likely produced directly by a large language model and may have been further modified by an AI bypasser. The second category, "AI-generated text that was AI-paraphrased," highlights text in purple that was likely AI-generated and then modified by an AI paraphrasing tool or word spinner such as Quillbot. An interactive submission breakdown bar gives instructors a visual overview of where AI-generated and AI-paraphrased content appears across individual pages. [2] The percentage from the AI writing detection model is completely independent of the similarity score; they measure entirely different aspects of a submission. [1]
Turnitin emphasizes that the AI writing indicator should not be used as the sole basis for adverse actions against a student. False positives — incorrectly flagging human-written text as AI-generated — are a recognized possibility in any AI model, and Turnitin has designed its system to keep the false positive rate for entire documents under 1% where possible. [1] Instructors are encouraged to use the report as one data point within a broader assessment process that includes their own professional judgment and institutional academic policies. [2]
Can Students Check Their Own Work With a Turnitin AI Writing Report Before Submitting?
In most institutional Turnitin setups, students cannot directly access the AI Writing Report for a draft before submitting it to an official assignment. The AI indicator and report are visible only to instructors and administrators by default, meaning students typically see their AI detection results only after their instructor has reviewed them. [1] However, there are some circumstances in which students may gain visibility. If an institution has enabled Turnitin Draft Coach, students can run similarity checks (and in some cases citation and grammar checks) directly within Google Docs or Microsoft Word. [4] Additionally, if an instructor configures an assignment to allow resubmissions, a student may submit drafts multiple times and see updated reports, though this depends entirely on the instructor's specific settings.
The ability to check for AI writing before submission is particularly valuable because it gives students the opportunity to understand how their writing may be perceived by detection systems and to make informed revisions. For students whose institutions do not provide Draft Coach or assignment resubmission access, third-party pre-submission checking services offer a practical alternative. These services allow students to upload their work and receive both an AI detection score and a similarity report before ever submitting to their instructor's Turnitin assignment, providing the same preview that educators see in institutional systems. [4] This pre-submission insight can help students identify unintended AI flags in their work — especially when they have used AI for research assistance, brainstorming, or light editing — and address those concerns proactively.
Understanding what the AI Writing Report shows and how scores are interpreted ultimately helps students take greater ownership of their academic work. Rather than being a punitive tool, the AI Writing Report, when used transparently, can serve as a learning mechanism that encourages students to engage more thoughtfully with their writing process and the ethical use of AI tools in their research and drafting workflows. [1]
※ Turnitin0.com - Actual Turnitin AI Report Cover, Score, Flag And Similarity Summary
FAQ
1. Is the Turnitin AI Writing Report the same as the Similarity Report?
No, they are completely separate and independent. The Similarity Report checks text against a database of existing sources to detect plagiarism, while the AI Writing Report analyzes sentence-level word probability patterns to identify text that may have been generated by AI tools. [1]
2. What does an asterisk (*%) score mean in the AI Writing Report?
An asterisk (*%) indicates an AI detection score between 0% and 20%. Turnitin hides the exact numerical percentage in this range because testing found a higher rate of false positives below 20%. The asterisk signals instructors that the score is less reliable and should not be taken at face value. [2]
3. Can the AI Writing Report detect content rewritten by paraphrasing tools like Quillbot?
Yes. Turnitin's AI detection capabilities now include AI paraphrasing detection, which identifies text that was likely AI-generated and then modified by paraphrasing tools. These segments are highlighted in purple in the Submission Breakdown, distinguishing them from directly AI-generated text shown in cyan. [2]
4. Should instructors penalize students based solely on the AI Writing Report score?
No. Turnitin explicitly advises that the AI indicator should not be used as the sole basis for adverse actions. The report is designed to provide data that educators can combine with their own judgment, knowledge of the student, and institutional academic policies before reaching any conclusion about misconduct. [1]
5. Can I check my own paper with a Turnitin AI Writing Report before my instructor sees it?
If your institution provides Draft Coach or if your instructor has enabled resubmissions on an assignment, you may be able to see results before final submission. Otherwise, students typically cannot view the AI Writing Report within institutional Turnitin. Third-party services like Turnitin0 offer pre-submission AI and similarity checking as an alternative. [4]
Sources
- Turnitin's AI Writing Detection Capabilities FAQs — https://guides.turnitin.com/hc/en-us/articles/28477544839821-Turnitin-s-AI-writing-detection-capabilities-FAQs
- Using the AI Writing Report — https://guides.turnitin.com/hc/en-us/articles/22774058814093-Using-the-AI-Writing-Report
- AI Writing Detection Guide for Educators — https://www.turnitin.com/resources/ai-writing-detection-educators-guide
- Can Students Check a Paper in Turnitin for Similarity Before Submitting? — https://helpcenter.turnitin.com/hc/en-us/articles/27811948436237-Can-students-check-a-paper-in-Turnitin-for-Similarity-before-submitting-it-to-an-assignment
Related articles
- Can Turnitin Detect Chatgpt Humanizer Tools? What Students Should Know
- Canada Essay Deadlines: Running Similarity + AI Checks the Night Before (Safely)
- How to Test Turnitin Responsibly: Similarity and AI Reports Before You Submit
- Turnitin Humanizer Review: What Reddit Claims vs What Pre-Submit Checks Show
- Why Does Turnitin Show an Asterisk Instead of a Percentage?