# Is 25% on the Turnitin AI score too high?

Source: https://digicruncher.com/topic/is-25-on-the-turnitin-ai-score-too-high/
Site: DigiCruncher, category AI detectors and false flags
Published: 2026-10-31. Updated: 2026-10-31. Replies: 3.

## Question by Rafa M., 2026-10-31

Is 25% on the Turnitin AI score too high? That is what my TA showed me on screen yesterday for my capstone design memo, and I do not know if it is a fail, a warning or nothing.

Context: our team of three wrote the memo together in one shared Google Doc. I wrote the background and the test plan sections and my teammates wrote the rest, and we did not use any AI to write it. The TA said "we will see what the instructor thinks," which was not comforting. Nobody has accused anyone of anything yet.

So in plain terms: is there a cutoff where the score becomes a problem? Is 25% worse than 15%? And what should I be doing right now besides keeping the document history?

## Reply 1 by Hollis B., 2026-10-31

Design memos are about as formulaic as prose gets. "The test plan consists of the following steps" has been written a million times. A statistical model may see predictable phrasing. That is a guess, not a diagnosis, so the useful move is to ask to see which passages were highlighted.

## Reply 2 by Nathan Cole (staff), 2026-10-31

A 25% Turnitin AI score is not a pass or fail number, because Turnitin says the percentage alone should not be the basis for action. There is no universal cutoff in the product, and your school's policy decides what a given number means. Ask which passages were highlighted.

What the numbers mean, per Turnitin's FAQ (updated October 14, 2026):

| Score shown | What Turnitin says it means |
|---|---|
| Gray with no percentage | The submission was not processed, for example under 300 words of prose, an unsupported file type or a submission made before the feature existed |
| Asterisk with no percentage | AI detected below 20%, with no percentage or highlights attributed |
| 20% to 100% | The share of qualifying prose sentences classified as likely AI generated or AI paraphrased |

So 25% is just above the line where Turnitin starts showing a number. Turnitin's FAQ (October 2026) says detections under 20% appear as an asterisk with no percentage, because false positives are possible in AI models.

Turnitin also says the percentage is not necessarily the percentage of your whole submission, because lists, headers, references, data tables, formulas and code are set aside. It also says false positives can come from text without much structural variation or text that repeats itself, and that in documents of only a few hundred words the prediction can be close to all or nothing.

On process, Turnitin's guidance on reviewing the report says the score should be one data point, balanced against what the instructor knows about the student, the work and the institution's policy. That means your best move is to keep calm, bring your Google Doc version history, and say how the team divided the sections. For the evidence checklist see [how to document your writing process if an AI detector flags you](/topic/how-to-document-your-writing-process-if-an-ai-detector-flags-you/).

No percentage on a Turnitin report is a verdict, and the instructor makes the final call.

Sources: [Turnitin's AI writing detection capabilities FAQs](https://guides.turnitin.com/hc/en-us/articles/28477544839821-Turnitin-s-AI-writing-detection-capabilities-FAQs) and [How should I review the AI Writing report?](https://guides.turnitin.com/hc/en-us/articles/27139000787853-How-should-I-review-the-AI-Writing-report).

## Reply 3 by Rafa M., 2026-10-31

@Nathan Cole thanks, that is the first answer that gives me a plan. I will ask the TA to see the highlighted passages, export the version history tonight and talk to my teammates about who wrote which section.

---
Cite as: DigiCruncher, "Is 25% on the Turnitin AI score too high?", https://digicruncher.com/topic/is-25-on-the-turnitin-ai-score-too-high/, accessed 2026-10-31.