Does Turnitin Train Its AI on Your Paper? What the Official FAQ Says

Turnitin's FAQ does not say that your submitted paper is used to train its AI detection model — and it does not say it is excluded either. The official wording says the model is trained on "a representative sample" of AI-generated and authentic academic writing, and that Turnitin's models are also "trained on our vast inhouse corpus of academic data."

What you can act on is deletion: the FAQ says institutions can request a full deletion of their submissions, and that partial deletions are not supported. The FAQ frames this as an institutional request and describes no individual path for a student to remove their own paper. Whether you hold a separate deletion right under data-protection law such as GDPR is a legal question this FAQ does not address.

Turnitin's AI detection FAQ answers "How was Turnitin's model trained?" with this:

Our model is trained on a representative sample of data spread over a period of time, that includes both AI generated and authentic academic writing across geographies and subject areas.

A second FAQ answer, on which model the detection is based on, goes further:

More importantly, our models are also trained on our vast inhouse corpus of academic data.

Neither sentence names student submissions as part of the training data, and neither sentence excludes them. "Authentic academic writing" and "inhouse corpus of academic data" are the closest the official text comes to describing where the words come from — and both formulations leave the question of your specific paper open.

That is the whole story of the "training" part: the FAQ does not deny that submitted work could be in the training mix, and it does not confirm it either.

The FAQ's question on data deletion is answered from the customer's — meaning the institution's — point of view:

Yes, customers can request a full deletion of their submissions; we cannot support partial data deletion requests to delete only the AI writing component of the submission data.

Two consequences follow.

First, deletion is possible in principle, but the requester is the institution. A university that wants its submissions scrubbed can ask for a full deletion of the data it owns. Second, you cannot delete just the AI component — there is no "remove the AI score, keep the paper" option, and no individual student deletion path described in the FAQ.

So the "can I request my work to be removed?" question, answered from what this FAQ describes, has a direct answer: not through this FAQ, which sets out only an institutional path. Any right you may have directly is a matter of your jurisdiction's data-protection law, not of this document.

The same FAQ contains the sentence most students are actually looking for, on a closely related worry:

No, it does not. There is no separate repository for AI writing detection, and customers retain the ability to choose whether to add their student papers into the repository or not.

This covers the "AI detection repository" — the idea that Turnitin files your paper into a special AI-checking database. The FAQ says there is no separate repository for AI writing detection, and that the decision to add papers to the similarity repository belongs to the institution.

The word "train" does not appear in this answer. An AI detection repository and training data are different things, and the FAQ keeps them separate: no separate repository for detection, no statement about training.

If you are a student who submitted a paper and now worries about its use, the honest summary is:

• The official FAQ does not confirm that your paper trains the model, and does not promise it does not.
• The closest official statements describe training data as "authentic academic writing" and an "inhouse corpus" — neither of which excludes student submissions.
• Deletion exists, but it is an institution-level full deletion, not a student-level removal, and partial deletions are not supported.
• The "no separate repository" answer is about AI detection storage, not about training.

What you can do with this information: if your worry is about the AI detection results on your own paper, the practical issues are score reading and rechecking, not data ownership. If your worry is institutional, the channel is your university's Turnitin administrator — they are the "customer" the FAQ means when it talks about deletion requests.

Turnitin's FAQ neither confirms nor denies that submitted papers train its AI models. The official text describes training data as a representative sample of authentic academic writing plus a vast in-house corpus, and says nothing that excludes student work. On removal, the only official path is an institution-level request for full deletion — there is no individual student deletion described. If you want to know what your school has agreed to, the person to ask is your university's Turnitin administrator, not Turnitin support directly.