Turnitin's AI Checker Has No Student Version; Turnitin0 Is Another Firm

No. The switch sits on the instructor's side, and that settles it. Turnitin's help center says AI writing detection reports are not visible to students; an instructor who wants a student to see one downloads the report and shares the saved file ⁸. The feature is only live on accounts whose institution holds a Turnitin Originality license ⁸.

EssayDone sells a detector of its own, so it has every reason to say otherwise, and its Turnitin guide still concedes the point: "Only instructors or admins can enable AI detection." ²

The license explains the rest.

The Markup pulled purchasing records from around 60 California institutions and the CUNY system, and the price came out per student for 2021-22. CUNY paid $1.79. UC Berkeley paid $2.11. Cal State Long Beach and Antelope Valley College each paid $3.03, while the UC Irvine Division of Continuing Education paid $6.50 ¹.

The AI detector is an upgrade negotiated on top of that. Sixteen more Cal State campuses got the AI detector at $3.12 per student instead of $3.19 after the chancellor's office put about $100,200 toward it ¹. South Orange County Community College District was paying $3.57 in the year The Markup reported, June 2025 ¹.

There is no checkout button for one person in that chain. The one thing Turnitin sells an individual is iThenticate credit for similarity checking, and its purchasing guide points there ⁹. The same guide states that subscriptions are designed for educational institutions, not for individual purchase, and that Turnitin does not offer subscriptions directly to individuals for Feedback Studio and Similarity ⁹.

A student's only contact with the tool is the submission. The paper goes into an assignment somebody else configured and the report is generated on that side. Whether an AI indicator shows up at all was decided by an instructor or an administrator when the assignment was set up ². EssayDone puts the student route in one short sentence: "There's no standalone student version." ²

A similarity percentage, or nothing at all. An instructor can release the similarity report to a class. The AI writing report is a separate panel. Turnitin's help center is flat about it: AI writing detection reports are not visible to students, and the only route to one runs through an instructor who downloads the report and hands over the file ⁸.

Draft Coach is the Turnitin tool a student can run alone, where an institution has switched it on ¹⁰. It runs similarity, citation and grammar checks, and Turnitin's student help page lists no AI writing check in it ¹⁰.

On some campuses there is no panel to share.

Vanderbilt disabled Turnitin's AI detector in August 2023 and published its arithmetic. The university had put 75,000 papers through Turnitin in 2022, so a 1% error rate could have wrongly labeled around 750 of them, had the tool existed that year ⁷. That is a lot of papers to get wrong.

A student on a campus like that can spend an evening hunting a score no instructor will ever open.

Where the panel is on, two of Turnitin's own rules decide whether any number appears at all.

The 300-word floor went up from 150 in the same May 2023 update, and Turnitin's reasoning was that accuracy improves with more text ⁵. A short discussion post or a one-page reflection can fall under it. What the percentage counts, and what it can't settle, is worked through in what a Turnitin AI percentage means.

Other companies, and Turnitin0's own FAQ says so ³. The site, at turnitin0.com, calls itself an independent service that helps university students preview Turnitin results before final submission, and two sentences later the same FAQ adds: "Turnitin0 is not affiliated with Turnitin, LLC." ³ The homepage above that FAQ still tells a visitor the reports are identical to what a professor sees ³.

Same page. Both claims.

The second product on those pages is the one worth naming, because Turnitin0 sells rewriting at $2.00 per 1,000 words and promises to bring an AI score under 20%, with a refund if it doesn't ³. EssayDone's guide tells a student, correctly, that no standalone version exists, then closes by pointing at its own detector and its own rewriting tool ².

A pre-check and an evasion product sold through one checkout make a single funnel, not two offers. The pre-check is the half that gets the visitor through the door.

What that rewriting does to a real paragraph, and the traces it leaves in the prose, is set out in what a humanizer does to a student's paragraph. The teacher's version of this search, which surfaces resale sites rather than lookalikes, is the free Turnitin AI checker question.

Because two detectors reading one paragraph return two numbers, and Grammarly admits as much about its own ⁶. Grammarly's user guide warns a writer not to read a Grammarly AI score as an indication that a professor's preferred detector will report the same thing ⁶. That is a company describing the limits of its own product, and it is worth taking at its word.

A percentage from a site that publishes no false-positive rate at all carries even less.

In 2023 Turnitin published two rates. They are not interchangeable. The document-level false-positive rate is under 1%, and that figure holds only for documents the model reads at 20% or more AI writing ⁴. The sentence-level rate is around 4%, which comes to about one highlighted sentence in every twenty-five being human-written ⁴.

A student squinting at a handful of highlighted lines is reading the noisier number.

Turnitin's own instruction to instructors is the shortest statement of what a score is for: "So, use the information to initiate a conversation, not to draw a conclusion." ⁴

A flag is never a finding, and that cuts both ways. A low reading bought the night before submission proves nothing about what the instructor's copy will show. The file was scored on a different day, through what Turnitin0 itself calls a non-repository path ³, and whether Turnitin's own model read it at all is the site's claim; nothing in this article verifies it. The money buys a number. It doesn't buy a prediction.

Open the document's version history and then leave it alone. Where the word processor keeps one, it records when text appeared and in what order, and that record costs nothing to keep. It is the clearest account of how the paper was written, as opposed to how it reads; notes, outlines and earlier drafts do the same job.

Nine drafting sessions look different from one paste. Either can be explained.

Then two questions. Both go to the instructor, not to a search engine. Does this course use Turnitin's AI writing indicator at all? And what does the department do with a reading when one appears? At a campus that switched the panel off, a paid pre-check answers a question nobody on the faculty is asking ⁷.

Where it was left on, the vendor's own instruction is that a reading starts a conversation ⁴. The second question is about what happens after that.

The patterns a revision history actually shows, including the ones that look worse than they are, are in Google Docs version history and AI. What the free student-facing checkers can and cannot do, and what each one publishes about its own errors, is in free AI detectors for students.