Do colleges use AI detectors? It depends on what they bought

Colleges do not all have AI detection, and the reason is not that some have a policy against it. The reason is that Turnitin's AI writing detection is only available to customers that license Turnitin Originality. A school running Turnitin Feedback Studio (TFS), Similarity, or the legacy Originality Check does not have AI detection unless somebody went and negotiated access. So the question "does my college check for AI" resolves to three smaller questions: did the institution buy the licence that includes it, did the instructor turn it on, and what does that number actually count for once it shows up.

We are going to walk through those three in order, because the answer people usually want is a yes or a no, and the real answer is a dependency chain. Each link in that chain is documented on Turnitin's own pages, and none of them involve guessing.

The single sentence that answers most of this is buried in the "Access & licensing" section of Turnitin's detection capabilities FAQ:

"AI writing detection is only available to customers that license Turnitin Originality."

We cleaned up the grammar. The page itself reads "is only be available," which is a typo we are not carrying into this article. What the sentence says is plain: AI detection is a feature of one specific product tier. It is not a switch that exists on every Turnitin installation.

The sentence directly after it closes the gap:

"If you are a Turnitin Similarity, Turnitin Feedback Studio (TFS) or Originality Check customer, please speak to your Turnitin account manager regarding access to AI writing detection."

Read those two together and the picture is complete. A college that bought TFS for grading and rubrics, or Similarity for plagiarism checking, does not have an AI detector sitting in the menu waiting to be toggled on. The capability is not there. Getting it means going back to the account manager and changing what the institution pays for, which is a procurement decision, not a settings decision.

This is why two students at two different schools can submit to Turnitin and get completely different reporting surfaces. One sees an AI indicator on the assignment inbox. The other sees similarity scores and nothing else, because the second school's licence does not include the tool that produces the indicator. If you want to understand the report you actually have, how to read a Turnitin AI writing report walks through the indicators and what each one means.

For postgraduate research and journal submissions, the tool in play is often iThenticate, not Turnitin Feedback Studio. The same licensing story applies, with an extra step.

Turnitin's iThenticate product page states: "iThenticate customers can access AI writing detection as a paid add-on." It sits under Add-ons, not under core features. The default iThenticate licence checks for duplicated text. AI detection is an additional line item.

What this means in practice is that even an institution paying for iThenticate is not necessarily paying for AI detection. A researcher uploading a manuscript to a journal that runs iThenticate might assume an AI score is being generated on the other end. It might not be. Whether that score exists depends on whether the journal's licence includes the add-on, and that is not something the submitter can see from their side.

The same iThenticate page also makes clear that its core check is for duplicated text, not AI content. We are not going to name specific publishers' licensing arrangements, because those change and we cannot verify them from the outside. But the structural point holds: AI detection on the iThenticate side is opt-in and costs extra, so assuming it is there by default is assuming a purchase nobody told you about.

Now suppose the licence is there. The institution bought Originality. The instructor enabled AI detection on the assignment. The paper goes in, the model runs, and a percentage appears somewhere.

That somewhere is not the student's screen.

"The AI writing detection indicator and report are not visible to students."

That is one of two sentences on the same FAQ page that say the same thing. The other reads: "only instructors and administrators are able to see the indicator." What closes the loop is the sentence that comes right after the first one:

"However, with the PDF download feature, instructors can download and share the AI report with students."

So the path from the score to you runs through a person. Your instructor sees the number. You see it only if they decide to download the PDF and hand it to you. Many instructors do, which is why plenty of students have seen their own reports. But that is an act of sharing, not a feature of your account. You cannot log in and check it the way you check a similarity score, and the absence of a number in your inbox does not mean the number does not exist.

This matters for the "does my college check" question in a specific way. A student can spend a semester being checked without ever knowing it happened, because the indicator lives on a screen they do not have access to. If you are trying to figure out whether AI detection is running on your submissions, asking classmates is not going to give you a reliable answer. The person who knows is the instructor, or the administrator who configured the assignment. And if the conversation about a flag has already started, the conversation after an AI flag covers what that exchange can and cannot settle.

So far we have established that the tool might not be there, and that even if it is, you might not see what it says. The third link in the chain is about what the number means once someone does see it.

Turnitin's help documentation says this in three places. The version on the AI Writing Report guide reads:

"Our AI writing detection model may not always be accurate (it may misidentify human-written, AI-generated, and AI-paraphrased text), so it should not be used as the sole basis for adverse actions against a student."

The next sentence is the one that gets dropped:

"It takes further scrutiny and human judgment in conjunction with an organization's application of its specific academic policies to determine whether academic misconduct has occurred."

That is the full statement. The model can be wrong. It should not be the only basis for action. And determining misconduct takes further review plus the institution's own policies. The FAQ says the same thing in its own words: the percentage "should not be used as the sole basis for action or a definitive grading measure."

There is a separate guide page, "How should I review the AI Writing report?", that puts the same point in positive terms:

"It is not meant to provide definitive answers in isolation. More important than any tool is the educator who sees the score and makes decisions balancing this information with their personal knowledge of their students, their work, and institutional policy."

And the sentence after that one: "When educators look at the AI writing score and utilize it as a single data point rather than a definitive response, then it is being used as intended."

A single data point. That is the company's own description of what its tool produces. Not a verdict, not a ruling, not a percentage that mechanically triggers a consequence. A data point that gets weighed alongside other things a person knows.

This is reinforced by the false positive rate the company publishes. The FAQ states: "We strive to maximize the effectiveness of our detector while keeping our false positive rate - incorrectly identifying fully human-written text as AI-generated - under 1% for documents with over 20% of AI writing." The next sentence translates that: "we might flag a human-written document as AI-written for one out of every 100 fully-human written documents." One in a hundred is low. It is not zero.

The company also says it accepts that the low false positive rate comes with a trade-off: "In order to maintain this low rate of 1% for false positives, there is a chance that we might miss some AI written text in a document." The direction the tool leans is toward not calling human text AI, even at the cost of missing some AI text. That design choice is good for students who did not use AI. It also means a low score is not a certificate of purity. What a 20% or 30% score means in practice, and why the threshold sits where it does, is laid out in is 20 percent AI too high, and what a 1% false positive rate actually looks like in volume is unpacked in what a one percent false positive rate means.

So the third answer is: even where AI detection exists and the number is visible, the company that makes the tool says it should not stand alone. If your college is using it, the number is input to a conversation, not the conversation itself. And if you are on the wrong end of a flag, flagged but you wrote it goes through what to bring to that conversation.

Put together, the chain looks like this.

1. Did the institution licence Turnitin Originality, or is it running TFS / Similarity / Originality Check? That determines whether AI detection exists on the assignment at all. If the licence does not include it, there is no toggle to flip, no setting to enable, and no AI indicator that could appear on your submission.
2. Is the instructor using it on this assignment? Even with the right licence, AI detection is configured per assignment. An instructor can choose to leave it off. Whether it ran on a specific submission is something only the instructor or administrator knows.
3. If it ran, what does the number count for? Turnitin says the score is not a definitive measure, should not be the sole basis for action, and is meant to be used as a single data point. The institution's academic integrity policy, not the tool, decides what follows from a given percentage.

Those three questions map onto a licence, a setting, and a policy. None of them is answered by "colleges use AI detectors" or "colleges don't." The tool is not a uniform layer across higher education. It is a product feature that some institutions have paid for, some instructors have turned on, and some policies treat as more conclusive than the maker intends.

Does every college that uses Turnitin have AI detection? No. AI detection is only available to customers that license Turnitin Originality. Institutions using Turnitin Feedback Studio, Similarity, or the legacy Originality Check do not have AI detection unless they have arranged access separately. It is a licensing question, not a feature that ships with every Turnitin product.

Can I check whether my school has AI detection? Not from your student account. The AI indicator is visible only to instructors and administrators. You can ask your instructor or your institution's instructional technology office whether AI detection is enabled on your assignments, and they will know.

If my instructor showed me an AI report, does that mean my college has the tool? Yes. The report you saw was produced by the AI detection feature, which means the licence includes it. Instructors can download the AI report as a PDF and share it with students, which is how most students see their own scores.

Does an AI score mean I'm in trouble? Not by itself. Turnitin says the score "should not be used as the sole basis for adverse actions against a student" and describes it as "a single data point." What happens after a flag depends on your institution's academic integrity policy and your instructor's judgment. The company explicitly says the tool does not make a determination of misconduct.

My friend at another school said their Turnitin doesn't show AI scores. Is that possible? Yes, and it is probably not a policy difference. Their school likely licences a Turnitin product that does not include AI detection. The absence of an AI indicator on their inbox is consistent with running TFS or Similarity without Originality.