Turnitin in Moodle: Setting Up the Plagiarism Plugin and Reading the AI Report

If your school runs Moodle, the Turnitin AI writing report doesn’t appear by magic. Moodle has no detector of its own. What it has is a plugin slot, and someone with admin access has to drop Turnitin into that slot and flip the right switches before a single AI percentage shows up anywhere.

This guide covers both halves of that: how the Turnitin plagiarism plugin gets set up in Moodle, and how to actually read the AI report once it’s running. We’ll keep the admin steps practical and the score interpretation honest.

Key takeaways

• Moodle has no built-in AI detector. Turnitin runs through the Plagiarism plugin (the plugin framework plus the Turnitin sub-plugin), and the AI writing report only appears once that’s installed and configured.
• Setup is two layers: a site admin enables the plagiarism API and the Turnitin plugin with your account credentials, then AI detection has to be switched on at the integration or instructor level.
• The AI writing report is separate from the Similarity (plagiarism) report, even though both surface from the same Turnitin viewer inside the Moodle assignment.
• Turnitin only scores qualifying submissions: at least 300 words of long-form prose in a supported language and file format. Short, list-heavy, or unsupported work often shows no score.
• The AI percentage is a probability estimate, not proof. A responsible workflow uses it as a prompt to look closer, not as a verdict.

Two systems, one assignment

Here’s the thing that trips people up. Moodle and Turnitin are different products from different companies. Moodle provides a generic “Plagiarism API,” basically a hook where any compatible detector can plug in. Turnitin builds a plugin that fits that hook. When you see an AI report on a Moodle assignment, you’re looking at Turnitin’s analysis displayed inside Moodle’s interface.

That matters because the setup lives in two places. The plugin has to be installed and authenticated at the Moodle site level. Then AI detection, which is a Turnitin account feature, has to be turned on too. Either one being off means no report. A school can have the plagiarism plugin happily checking for copied text while the AI writing indicator stays completely dark, because those are separate toggles.

There are also two flavors of the integration floating around. The older one is the standalone Turnitin plugin (sometimes called Turnitin Direct or the v1 plugin). The newer, recommended one is the Turnitin Plagiarism Plugin that works through Moodle’s Plagiarism API, often paired with an LTI launch for the viewer. New installs should use the current plugin; the legacy versions are being wound down.

Setting up the plugin (admin view)

This part is for whoever holds the site administrator role. If you’re an instructor or student, you can skim it, but you won’t have the access to do it yourself.

1. Turn on the Plagiarism framework

In Moodle, go to Site administration > Advanced features and tick Enable plagiarism plugins. Nothing detector-specific happens yet. This just unlocks the plugin slot so Moodle will accept a detector.

2. Install the Turnitin plugin

Grab the current Turnitin Plagiarism Plugin (plagiarism_turnitin) that matches your Moodle version from the Moodle plugins directory or your Turnitin account, and install it the usual way: either by uploading the ZIP through Site administration > Plugins > Install plugins or by dropping it into the plagiarism/turnitin directory and running the upgrade. Moodle walks you through the database upgrade step.

3. Authenticate with your Turnitin account

Under Site administration > Plugins > Plagiarism > Turnitin, you’ll enter the credentials Turnitin gave your institution: the API URL (your region’s endpoint) and the account ID and shared key, or the API key, depending on which integration version you’re on. Save, and the plugin tries to handshake with Turnitin. If the credentials are wrong, this is where it fails quietly, so confirm the connection before you move on.

4. Enable AI detection

This is the step people forget. Authenticating the plugin gets you plagiarism similarity checking. AI writing detection is a separate capability tied to your Turnitin license. Depending on your setup, it’s enabled in the Turnitin administrator console or exposed as an option in the plugin settings. If your license includes AI detection and it’s switched on, qualifying submissions start generating AI reports. If it isn’t, you’ll get similarity scores and no AI indicator.

5. Set the defaults instructors inherit

On the same plugin settings page you can set site-wide defaults: whether the plugin is on for assignments by default, whether students can see reports, when reports generate (immediately or on the due date), and how the repository is handled. Instructors can usually override some of these per assignment, but sane defaults save everyone confusion later.

Where the AI report actually appears

Once it’s all running, an instructor enables the plugin on a specific assignment (in the assignment’s settings, under the Turnitin plagiarism options) and waits for submissions.

On the assignment’s grading or submissions page, each student’s work shows a colored similarity icon in the Turnitin column. Clicking that icon opens the Turnitin viewer, Feedback Studio. The AI writing report is a distinct panel inside that viewer, usually its own icon or tab, showing a percentage of the qualifying text that the model flagged as likely AI-generated. The plagiarism similarity percentage and the AI percentage are two different numbers that happen to live behind the same icon.

Picture a common situation. An instructor opens a Moodle assignment after the deadline. One submission shows a similarity score in the low single digits, which is nothing, but the AI report comes back high. The two numbers aren’t in conflict; they measure different things. A low similarity score means almost none of the text matches existing sources. A high AI number means Turnitin’s model thinks the prose reads as machine-generated. The sensible next move isn’t to file an academic-integrity case on the spot. It’s to look at the student’s earlier drafts and have a short conversation, because the number is a flag, not a finding.

Reading the number without overreading it

The AI percentage is an estimate of how much of the qualifying text reads as AI-written. It is not a confidence level, not a probability that the student cheated, and not evidence on its own.

Two facts worth holding onto. First, detectors are probabilistic by design; the underlying models score how “predictable” the text is, which is why detectors output a probability, not proof. Second, false positives are real and unevenly distributed. A 2023 Stanford study by Weixin Liang and colleagues, published in the journal *Patterns*, found that several GPT detectors misclassified more than half of a set of TOEFL essays by non-native English writers as AI-generated, while flagging native-speaker essays at close to zero. The authors point to low perplexity, plainer and more predictable phrasing reads “machine-like” to these models. OpenAI itself retired its own AI text classifier in July 2023, citing low accuracy. None of that means the tools are useless; it means the output is a starting point.

So when you read a Moodle AI report, treat a high number as a reason to look closer, not as a conclusion. Look at version history, the student’s prior voice, and whether the writing is just clean and ESL-plain rather than generated.

What the blank states mean

A missing AI report usually isn’t a 0%. Turnitin needs at least 300 words of long-form prose in a supported file format and language. Short answers, bullet lists, code, heavy quotation, and unsupported languages often produce no score. Reports also take a few minutes after submission. So “no number” most often means “not enough qualifying text,” not “definitely human.”

A note for students and writers

If you’re a student, you generally can’t see your own AI report inside Moodle; it’s instructor-facing by default. The best protection against a false flag isn’t gaming the tool, it’s keeping a paper trail: drafts, comments, outline notes, version history. If you write with AI assistance and want the final work to genuinely read as your own voice rather than flat, predictable model output, the fix is editing for specificity and rhythm, not tricks. We’ve written more about that across more guides on detection and writing, and if you check your work often, you can see which plan fits how often you check.

Bottom line

Turnitin AI detection in Moodle is a two-part affair: the plugin has to be installed and authenticated, and AI detection has to be separately switched on. Once it’s live, the AI report sits inside the Turnitin viewer next to the familiar similarity score, and it tells you a probability, not a verdict. Read it that way and it’s useful. Read it as proof and you’ll eventually be wrong about a real student.

Want to see how your own writing reads before anyone else does? Run it through PaperBleach and check for yourself.