Does Turnitin Detect AI? How the AI Writing Report Actually Works

Yes — Turnitin detects AI writing, when the feature is enabled. Its AI writing report scores how much of a submission reads as AI-generated, separately from similarity checking. But Turnitin AI detection only runs if your university has switched it on — 22 of the 95 UK universities we track have the indicator explicitly off — and what the report shows is more limited than most students assume.

This guide explains, from the submitting student's side, what Turnitin's AI writing report is, how it behaves, what it cannot show, and how to find out whether your university even uses it.

The whole page in one table:

The short answer

Turnitin's AI writing detection exists, is separate from its similarity (plagiarism) checking, and reports the percentage of qualifying long-form prose that its model classifies as likely AI-generated or AI-generated-then-paraphrased. Documents need at least 300 words of prose to be scored, scores from 1–19% are shown as an asterisk rather than a number because false positives are more common in that range, and Turnitin itself says the result should never be the sole basis for an adverse decision. Crucially, the AI indicator is an institutional setting: many UK universities have chosen not to run it.

The rest of this cluster

This page is the general guide. Three companion pages answer the narrower questions:

• Can Turnitin detect ChatGPT? — the ChatGPT-specific case, including copy-paste, Copilot and Gemini.
• Can Turnitin detect paraphrased AI writing? — what the AI-paraphrase category does, and why humanisers backfire.
• How accurate is Turnitin's AI detection? — vendor claims against independent evidence.

How Turnitin's AI detection works

Turnitin's published model documentation describes a classifier trained to distinguish AI-generated prose from human writing. The workflow, per Turnitin's own guides:

• The submission is broken into overlapping segments of text.
• Each segment is scored by the detection model.
• Qualifying segments' results are aggregated into a document-level percentage.

Only "qualifying" text counts: long-form prose in supported languages. Bullet lists, tables, code, references and very short passages are excluded from the calculation, which is one reason the AI percentage is not a percentage of your whole file. Turnitin's AI writing detection model page and its file requirements for an AI writing report set out the current rules; check them for the live details, because they change.

The model also has a category for AI-generated text that was later paraphrased — which is why paraphrasing ChatGPT output is not the workaround some students assume.

What the AI report shows — and hides

According to Turnitin's guide to using the AI writing report:

• Staff see the report; students usually do not. The AI indicator is shown to instructors. Several universities state that students never see the AI percentage for their own submission.
• Scores of 1–19% display as an asterisk. Turnitin suppresses the number in this band because low scores carry a higher risk of false positives.
• Highlighting shows which passages drove the score. The report distinguishes text classed as AI-generated from text classed as AI-generated-then-paraphrased.
• The percentage covers qualifying prose only — not the whole document.

What the report cannot show is equally important: it cannot identify which tool produced the text, when it was written, or whether the writer had permission to use AI assistance. Those are questions for a human process, and Turnitin's own myth-busting post says the score should start a conversation, not end one.

AI score vs similarity score

The AI writing percentage and the similarity score are different systems answering different questions:

A document can score 2% similarity and 90% AI, or the reverse. Our guide to AI detection vs plagiarism detection unpacks the difference in detail, and Turnitin's similarity score guide covers the plagiarism side.

Your university may have it switched off

The AI indicator is licensed and enabled at institution level. Genutext's UK university AI-detection policy tracker records, from universities' own published statements, that a substantial share of UK institutions with a public position have opted out of or disabled Turnitin's AI indicator, while a smaller group confirms it is in routine use. Newcastle University's library FAQ, for example, states the university does not use Turnitin's AI feature and describes detection tools as highly unreliable, while other universities confirm the AI report is part of their workflow.

That means the practical answer to "will Turnitin flag my essay for AI?" starts with a different question: does your university run the indicator at all? Check your university's own guidance first — the tracker links each statement.

What Turnitin says about its own accuracy

Turnitin publishes a document-level false positive rate below 1% for documents with 20% or more AI writing, and a sentence-level rate of roughly 4% — figures measured on its own benchmarks. Independent testing reports weaker real-world accuracy, with errors concentrated on non-native English writers.

Because that evidence deserves more than a paragraph, it has its own page: how accurate is Turnitin's AI detection? sets the vendor's claims beside the independent studies and UK universities' own conclusions, every figure dated and linked.

How to see your AI percentage as a student

Usually, you cannot — the AI report is an instructor-side feature, and no third-party tool can tell you the number Turnitin would produce. What you can do before submitting:

1. Check whether your university uses the indicator in the policy tracker and your course handbook.
2. Keep your drafts and version history. They are stronger evidence of your writing process than any score. Our guide to proving you didn't use AI explains what actually persuades a panel.
3. Understand your own writing's signals — the honest way. Our pre-submission check page explains what an independent scan can and cannot tell you. An independent check such as Genutext's free AI detector shows whether your prose carries AI-like patterns and which passages carry them — as a way of understanding your own work, not as a prediction of Turnitin's number. The two systems will disagree, and that disagreement is normal.

Frequently asked questions

Does Turnitin detect AI writing?

Yes. Turnitin offers AI writing detection alongside its similarity checking, reporting the share of qualifying prose its model classes as likely AI-generated. Whether it runs on your submission depends on whether your institution has enabled it.

Does Turnitin check for AI automatically?

Only where the institution has the AI indicator enabled. It is an institutional licensing and configuration choice, and many UK universities have publicly opted out.

How does Turnitin detect AI?

Its classifier splits the submission into overlapping segments, scores each for AI-like statistical patterns, and aggregates qualifying segments into a document percentage. It also attempts to identify AI text that was subsequently paraphrased.

Can students see the Turnitin AI score?

Usually not. The AI report is shown to staff; several universities state students cannot see the AI percentage for their own work.

Is a Turnitin AI score proof of cheating?

No. Turnitin itself says the score should not be the sole basis for action, and UK guidance requires evidence beyond a detector result. See can an AI detector score be used as evidence?

Can Turnitin detect ChatGPT specifically?

It detects AI-like writing patterns; it does not name the tool. Our dedicated guide covers what Turnitin can and cannot show about ChatGPT.

Sources and further reading

• Turnitin: Using the AI Writing Report
• Turnitin: AI writing detection model
• Turnitin: File requirements for an AI writing report
• Turnitin: Understanding the similarity score
• Turnitin: Does Turnitin detect AI writing? Debunking common myths
• Turnitin: Understanding the false positive rate
• Turnitin: AI writing detection model architecture and testing protocol
• HEPI: Catching the wrong students (20 July 2026)
• Genutext: UK University AI-Detection Policy Tracker

Turnitin is a trademark of its owner. Genutext is independent and not affiliated with, endorsed by, or connected to Turnitin. Product behaviour described here reflects Turnitin's public documentation as of 23 August 2026; check the linked pages for current details.