# What does a Turnitin AI writing score actually mean?

Source: https://humaniseai.ai/forum/read/what-does-a-turnitin-ai-writing-score-actually-mean/
Site: humanise.ai Forum, category Detector flags and appeals
Published: 2026-11-11. Updated: 2026-11-11. Replies: 0.

## Question by Bronwyn Ashby (staff), 2026-11-11

A Turnitin AI writing score is the share of eligible prose sentences in your submission that Turnitin's model predicts were written by AI, or written by AI and then reworked by a paraphraser. It is a prediction, not proof. Turnitin itself says the score should not be the sole basis for action against a student.

## What the percentage counts

Turnitin's own [AI writing detection FAQs](https://guides.turnitin.com/hc/en-us/articles/28477544839821-Turnitin-s-AI-writing-detection-capabilities-FAQs) explain the number in some detail, and they are worth reading in full if you have been flagged. The short version:

- The model only looks at "qualifying text": prose written in standard grammatical sentences. Bullet points, tables, code and reference lists are largely left out.
- Each qualifying sentence gets a prediction. Sentences above Turnitin's threshold are labelled AI and the rest human. The document score is the percentage of sentences above the threshold.
- Depending on the language, Turnitin does not process a document when the evaluated sentences total under 300 words.
- The score covers text the model thinks was "likely generated by AI or likely generated and modified by an AI paraphraser or bypasser". Turnitin says the model is built to pick up text run through so-called humanisers as well.

So 40% does not mean "40% of the ideas came from ChatGPT". It means the model scored roughly 40% of the eligible sentences above its threshold.

## Why low scores show as an asterisk

Turnitin says that to reduce false positives it no longer shows a number between 1% and 19%. If it detects something in that range, the report shows an asterisk (*%) with no percentage. A grey indicator with dashes means the submission could not be processed, for example because it was too short or in an unsupported format.

## How accurate is it

Turnitin states that it aims to keep its false positive rate, meaning fully human text wrongly identified as AI, under 1% for documents with over 20% AI writing. It describes that as possibly flagging one in every 100 fully human documents, and says it tests each model update on 700,000 academic papers written before ChatGPT existed.

Two things follow from that, and both matter if you are the student:

1. Under 1% is not zero. Across a large cohort, some human-written essays will be flagged.
2. That figure is about whole documents. Turnitin's FAQ also notes that the percentage does not always match the amount of highlighted text.

Independent research adds a further caution. Researchers at Stanford found that several popular detectors (not Turnitin specifically) misclassified a large share of essays written by non-native English writers as AI, which [Stanford HAI summarised here](https://hai.stanford.edu/news/ai-detectors-biased-against-non-native-english-writers). Plain, formulaic or heavily edited writing can look "predictable" to a detector whoever wrote it.

## Who sees the score

Turnitin says the AI writing indicator and report are not visible to students. Your marker can download the report as a PDF and share it with you. If your university uses the score in a misconduct case, ask for that report; you are entitled to see the evidence used against you under most university procedures.

## What a marker is supposed to do with it

Turnitin's position is that it does not make a determination of misconduct. It "provides data for the educators to make an informed decision", and the percentage "should not be used as the sole basis for action or a definitive grading measure". In practice, a well-run process treats a high score as a reason to look closer: compare the writing with your earlier work, ask you about your drafting, and then decide.

## If your score worries you

| Situation | Sensible next step |
|---|---|
| You used no AI and the score is high | Gather drafts and version history, then read the thread below on what happens next |
| You used AI within the rules and declared it | Point to your declaration and keep your prompts; the rules themselves are covered in [am I allowed to use ChatGPT for university assignments](/forum/read/am-i-allowed-to-use-chatgpt-for-university-assignments-in-the-uk/) |
| You used AI outside the rules | Read your policy carefully and seek advice from your students' union before any meeting |
| The score shows *% or dashes | There is nothing to act on; no percentage was attributed |

Keep calm and keep evidence. A score is the start of a conversation, not the end of one.

---
Cite as: humanise.ai Forum, "What does a Turnitin AI writing score actually mean?", https://humaniseai.ai/forum/read/what-does-a-turnitin-ai-writing-score-actually-mean/, accessed 2026-11-11.