Why Does Turnitin Show an Asterisk Instead of a Percentage?

Table of Contents

Direct Answer - Turnitin displays an asterisk (*%) instead of a numerical percentage for AI detection scores between 0% and 20% to reduce the risk of misinterpreting low-confidence results. This design choice, implemented across all Turnitin AI Writing Reports, reflects the company's findings that false positives are more frequent in this lower score range. The only explicit low numerical score a student sees is 0%; any score above 0% but below 20% is suppressed to an asterisk, accompanied by sentence-level highlights that still show where AI-like text was detected. [1]

What Does the Asterisk Mean in Turnitin AI Detection Reports?

When you see an asterisk (*%) on your Turnitin AI Writing Report, it means the overall percentage of qualifying text detected as AI-generated falls between 0% and 20%. Turnitin deliberately withholds the exact single-digit or low double-digit percentage because the model's confidence is lower in this band. [2] Instead of displaying a figure like 3% or 12%, which could be misinterpreted or over-relied upon, the system consolidates all sub-20% scores into a single asterisk indicator.

It is important to understand that the asterisk does not mean "no AI was detected." The report still provides granular, sentence-level highlights in cyan (AI-generated) and purple (AI-paraphrased) colors even when the overall score is an asterisk. [2] This means an instructor can look at the highlighted passages and use their professional judgment rather than relying on a numerically uncertain overall score.

Turnitin has been transparent about the rationale: their internal testing confirmed a higher incidence of false positives—incorrectly flagging human-written text as AI-generated—when the detected percentage is between 1% and 19%. [1] By suppressing the exact number and showing an asterisk instead, Turnitin encourages educators and students to focus on the qualitative highlights and contextual review rather than a potentially misleading low percentage.

How Does Turnitin Determine AI Writing Scores and Display Thresholds?

Turnitin's AI writing detection model analyzes two key linguistic features—perplexity and burstiness—to differentiate human-written text from AI-generated prose. Perplexity measures how predictable a sequence of words is, while burstiness evaluates the natural variation in sentence length and structure. Human writing typically exhibits higher burstiness and more varied perplexity patterns compared to the uniform output of large language models. [3]

The detection model processes only "qualifying text," defined as prose sentences in a long-form writing format. Bullet points, poetry, code, annotated bibliographies, and other non-prose elements are excluded from the analysis. [3] This means a submission might contain language from an AI source that, if written outside of qualifying prose, would not contribute to the overall percentage—another reason why the asterisk threshold requires careful interpretation.

The 20% display threshold was set based on Turnitin's performance data. According to the company's documentation, the model has a higher false-positive rate in the 0–19% range, meaning it is more likely to misidentify human-written sentences as AI-generated in this band. [2] Rather than surfacing an unreliable number, Turnitin made the deliberate decision to suppress any percentage below 20% and replace it with an asterisk. This threshold applies to all new submissions processed after July 8, 2024; reports generated before that date may still show a numerical score below 20%. [2]

Can Students Preview Their Turnitin AI Score Before Submitting to Their Institution?

In most university setups, students submit their work directly through an institution-managed Turnitin-enabled learning management system (LMS). While similarity reports are often visible to students after submission, AI writing reports are typically reserved for instructors. This means many students only discover their AI score after their paper has already been submitted—sometimes too late to review or revise flagged sections. [4]

Third-party checking services like Turnitin0.com bridge this gap by allowing students to upload their drafts before the official submission deadline. Students receive both a full similarity report and an AI detection report that mirrors what their professors see in institutional Turnitin. [4] This pre-submission preview is especially valuable when the output shows an asterisk (*%) because the student can still inspect the sentence-level highlights to understand which passages were flagged, even if the overall percentage is suppressed.

Previewing your AI score early also gives you a chance to review your writing process. If you used AI tools for brainstorming or editing, the asterisk may simply indicate that your revisions were thorough enough to stay within the low-confidence band. If you notice highlighted sentences that surprise you, you can revise those sections before final submission. [4] This proactive approach is far better than waiting for an instructor-initiated conversation after the paper has been graded.


Understanding the asterisk is the first step; knowing exactly what your report looks like before it reaches your professor is the next. Turnitin0.com gives students real, institution-matched Turnitin AI and similarity reports within minutes—no subscription needed.

※ Turnitin0.com - Actual Turnitin AI Report Cover, Score, Flag And Similarity Summary

Get Real Turnitin AI & Similarity Report

FAQ

1. Does the asterisk (*%) mean my paper is safe from AI detection?
Not necessarily. The asterisk indicates the overall score is below 20%, but sentence-level highlights may still flag specific passages as AI-generated or AI-paraphrased. Review those highlighted sections carefully rather than relying solely on the asterisk. [2]

2. Can my instructor still see an exact percentage below 20%?
No. Since the July 2024 update, instructors also see the asterisk instead of a numerical percentage for scores between 0% and 20% on new submissions. [2] Both students and instructors view the same suppressed score.

3. What should I do if my report shows an asterisk but has many highlighted sentences?
Use the highlights as a revision guide. The asterisk means the overall confidence is low, but the model still identified specific sentences as potentially AI-generated. Revise or rewrite those highlighted passages to reduce the likelihood of an instructor-initiated discussion. [1]

4. How accurate is Turnitin's AI detection for scores under 20%?
Turnitin has publicly stated that false positives are more frequent in the 0–19% range. This is precisely why the asterisk was introduced—to prevent users from over-interpreting a potentially unreliable low score. [3] Always contextualize the report with your actual writing process.

5. Can I check my Turnitin AI score before submitting my final draft?
Yes. Services like Turnitin0.com provide pre-submission AI and similarity reports that match institutional Turnitin output. You can upload your draft, see the asterisk or any percentage, and review the highlights before your official submission. [4]

Sources

  1. Turnitin AI Writing Detection Frequently Asked Questions — https://guides.turnitin.com/hc/en-us/articles/28477544839821-AI-Writing-Detection-Frequently-Asked-Questions
  2. Using the AI Writing Report — https://guides.turnitin.com/hc/en-us/articles/22774058814093-Using-the-AI-Writing-Report
  3. How Turnitin's AI Writing Detection Works — https://helpcenter.turnitin.com/hc/en-us/articles/27811948436237-How-Turnitin-s-AI-Writing-Detection-Works
  4. Can Students Check Their Turnitin AI Score Before Submitting — https://www.turnitin.com/blog/can-students-check-their-turnitin-ai-score-before-submitting

Contact us

Email us or reach us on WhatsApp. We typically reply within business hours.