Which AI Detector Does Your University Use? Rutgers, UC, WGU & More
If you have found your way here, you probably typed some version of a very specific question into a search box: what does my school actually run when it checks for AI writing? Not colleges in general, not the industry, not what some vendor claims. Your university. The one whose logo is on your tuition bill and whose academic-integrity office could, in theory, decide your fate over a paper you wrote yourself. It is a reasonable thing to want to know, and it is frustratingly hard to get a straight answer, because most schools do not publish a tidy page that says "here is our AI detector and here is how it scores you." This article is an attempt to give you the honest, practical version of the answer, along with a method for figuring out your own school's setup when the internet cannot tell you directly.
Let me start with the single most useful fact, the one that resolves the majority of cases before we get into named examples. For the overwhelming share of universities that check student writing for AI at all, the tool is not some exotic standalone product. It is Turnitin. That name is probably already familiar to you as a plagiarism checker, because Turnitin has been the default similarity-matching service on North American campuses for the better part of two decades. When you submit a paper through Canvas, Blackboard, Moodle, D2L Brightspace, or a similar learning-management system, and your instructor has "similarity check" turned on, that submission is very often flowing through Turnitin behind the scenes. In 2023 Turnitin bolted an AI-writing indicator onto that same pipeline. So for a lot of students, the detector "your university uses" and the plagiarism checker "your university uses" are the same piece of software, and you interact with both through the exact same submission button.
Why the honest answer starts with Turnitin
It helps to understand how this became the near-universal default, because the reason is institutional inertia rather than any judgment that Turnitin is the best AI detector on the market. Universities licensed Turnitin at the campus or system level for plagiarism, signed multi-year contracts, integrated it into their LMS, and trained thousands of faculty to use it. When Turnitin added an AI indicator, it arrived inside a product schools had already paid for and already deployed. There was no procurement battle, no new login, no separate tool for professors to learn. For most administrators, the AI feature simply appeared. That distribution advantage — being already installed everywhere — is why, when a student asks "which detector does my university use," the statistically safest guess is Turnitin, even at schools that have never made an announcement about it. We go deeper into this dynamic in our piece on what AI detector colleges use, and on the specifics of the model inside that product in what AI detector Turnitin uses.
But here is the crucial complication, and it is the reason "your university uses Turnitin" is a starting point rather than a conclusion: having the feature and using the feature are two different things. Turnitin ships the AI indicator to institutions, but whether it is visible to instructors, whether administrators have enabled it, and whether faculty are permitted or encouraged to act on it are all local decisions. A school can have Turnitin fully deployed for plagiarism and have the AI indicator switched off entirely. Which brings us to the most important and least-known part of this whole story.
Many well-known universities turned the AI detector off
Shortly after Turnitin released its AI-writing indicator, a striking thing happened: several prominent universities publicly decided not to use it. This is one of the best-reported facts in the entire AI-detection space, and if you take one thing away from this article, make it this one. Vanderbilt University announced it was disabling Turnitin's AI detection for its instructors. Michigan State University did the same. The University of Texas at Austin likewise moved away from relying on the feature. These were not quiet footnotes; they were deliberate, explained decisions from large, serious institutions, and the reasoning they gave was consistent: the AI detector produced false positives, its accuracy could not be verified independently, and the risk of wrongly accusing an honest student outweighed whatever benefit the tool offered.
Read that reasoning again, because it is remarkable. These schools did not disable AI detection because they were indifferent to academic integrity. They disabled it because they concluded the tool was not reliable enough to justify the harm it could cause. When a university's own teaching-and-learning center looks at a detector and says "we do not trust this to be fair to students," that tells you something no vendor marketing page will. It tells you the "detector your university uses" might, at a thoughtful institution, be no detector at all — a policy of deliberate restraint rather than a policy of surveillance.
I want to be careful and precise here, because the rule for this article is accuracy over drama. What is well-reported is that these named universities disabled Turnitin's AI indicator as of when they announced it. Institutional policy is not frozen. A school that disabled a feature in one academic year could re-enable it later, adopt a different tool, or revise its stance as the technology changes. So treat the Vanderbilt / Michigan State / UT Austin examples as an illustration of a well-documented pattern — respected schools choosing restraint over an unreliable tool — and not as a guarantee of any school's current live configuration. If your own school is one of these, the honest move is still to verify against today's policy rather than assume the announcement from a past year is unchanged. The pattern is the lesson; the specific status of any one campus is something you confirm locally.
How to actually find out what your school does
Since the truth varies campus to campus and even course to course, the reliable path is to investigate your own institution directly rather than trusting a generic list. Here is where to look, roughly in order of how authoritative and how current each source tends to be.
Start with the syllabus for each class. This sounds almost too simple, but the syllabus is frequently the single most binding document about how AI is handled in a specific course. Many instructors now include an explicit AI-use section: some ban it, some allow it with citation, some permit it for brainstorming but not drafting. Critically, the course-level rule can override or refine any campus-wide default. Two professors at the same university can hold opposite policies, and both are legitimate. If your syllabus says how AI is treated and whether submissions are checked, that is the rule that governs your grade — read it before you read anything else.
Read the university's academic-integrity or academic-honesty policy. Most schools publish this centrally, often through the dean of students, the provost's office, or an "academic integrity" microsite. Search your school's name plus "academic integrity policy" or "generative AI policy." This is where institution-wide stances live: whether AI writing is categorized as misconduct, what the process is if you are accused, and sometimes — though not always — whether detection tools are used at all. Even when it does not name a specific detector, it tells you how seriously the school treats the issue and what your rights are in a dispute, which matters more than the brand name of the software.
Ask the library or the writing center. This is an underused move and often the fastest way to a real answer. Librarians and writing-center staff are usually neutral, well-informed, and not the people who would be accusing you of anything, which makes them safe to ask. They frequently know exactly which tools are integrated into the LMS, whether the AI indicator is turned on, and how faculty have been advised to use it, because they help train faculty and support students on both sides. A polite email — "I want to understand how AI writing is checked here so I can make sure my process is above board" — will often get you a clearer answer than any policy document.
Look at which LMS you use, and check whether similarity checking is on. If your submissions run through Turnitin, you can sometimes see it directly. When an assignment has similarity checking enabled, students can often view their own similarity report after submitting, and where the AI indicator is enabled and visible to students, an AI-writing percentage may appear alongside it. If you submit a paper and see a Turnitin report, you have your answer about the plagiarism side; whether the AI number is shown to you specifically depends on how your school configured it. Absence of a visible AI score does not prove the tool is off — instructors often see indicators students never do — but its presence is a strong signal.
When in doubt, ask the instructor plainly. Framed the right way, asking is not suspicious; it is responsible. "Do you use an AI-detection tool on our submissions, and how do you interpret the results?" is a fair question that a reasonable instructor should be willing to answer. Their reply also tells you something about how they would handle a flag — whether they treat a score as gospel or as a prompt for a conversation. That posture matters as much as the tool.
Named examples, honestly framed
Let me walk through the categories of schools people ask about most, because the type of institution shapes the likely setup more than any single brand does. I will be deliberate about not inventing a current per-school policy for any of these; I am describing patterns and how to reason about them.
Large public university systems
People often ask about big public systems — the University of California campuses, large state universities in general, big multi-campus systems — as if the whole system runs one detector by decree. Usually it does not work that cleanly. Large systems tend to license enterprise tools centrally (Turnitin is common at this scale precisely because it is sold that way), but the decision to enable the AI indicator, and the guidance given to faculty, frequently sits with individual campuses, colleges within a campus, or even departments. A sprawling public system may have Turnitin available everywhere while one campus's teaching center urges caution and another leaves the feature on. The UCs and similarly large public institutions are exactly the kind of place where the honest answer is "the tool is probably available, but whether and how it is used depends on your specific campus and course — verify locally." Do not assume uniformity just because it is one system on paper.
Traditional state schools
Mid-size and regional state universities are the heartland of the Turnitin default. Many adopted it for plagiarism years ago and inherited the AI indicator when it shipped. Some of these schools have thoughtful, cautious guidance echoing the reasoning behind the high-profile disablings; others simply left the feature on without a public statement. If you attend a state school and have no other information, "Turnitin is likely present, its AI indicator may or may not be active, and my syllabus and integrity policy are the documents that actually govern me" is the accurate mental model. It is unglamorous, but it is true.
Online-first and competency-based schools like WGU and SNHU
Online-first institutions — Western Governors University, Southern New Hampshire University's online programs, and the broader category of large online and competency-based schools — deserve their own note, because the environment is genuinely different. Fully online programs often lean harder on a combination of tools than a traditional campus does. They tend to use plagiarism and similarity checking (which may include Turnitin or comparable services), and because coursework is remote, many pair that with proctoring technology for exams — software that monitors a webcam, locks the browser, or flags unusual behavior during a timed assessment. Detection of AI writing, where it happens, sits inside this broader monitoring posture rather than standing alone.
It is worth understanding why online programs behave this way without assuming the worst about any specific school. When you cannot see a student in a room, institutions substitute technological verification for physical presence, and that verification stack is usually more layered than what a lecture hall requires. That does not mean online schools are running some uniquely aggressive AI detector; it means the AI question at an online school is entangled with the general question of how remote work is verified. If you are at WGU, SNHU, or a similar program, your investigation should include not just "is there an AI detector" but "what proctoring and integrity tools does this program use overall," because those pieces are related and your student handbook usually addresses them together. As always, the handbook and the program-specific policy are the authoritative sources, not a generic assumption.
Why "which detector" matters less than you think
Here is a reframing that I think genuinely helps, even though it may not be what you came looking for. Students fixate on the identity of the detector — Turnitin versus GPTZero versus Copyleaks versus Originality.ai — as if knowing the exact brand were the key to safety. It rarely is, for three reasons.
First, every one of these tools is fallible in the same fundamental way. They are all probabilistic. None of them can actually "see" whether you used AI; they estimate a likelihood based on statistical properties of the text, and every one of them produces false positives on human writing. The failure mode that got real students in trouble — clean, original prose flagged as machine-written — is not unique to one vendor. It is inherent to how the whole category works. Knowing your school uses detector A rather than detector B does not make you safe from a wrong flag, because detector B has the same weakness. Our explainer on why AI detectors produce false positives goes through the mechanism in detail, and it applies across brands.
Second, as the disabling stories show, the "which detector" question sometimes has the answer "none, deliberately." A school that turned off Turnitin's indicator has effectively decided the brand does not matter because the whole approach is too unreliable to use. You cannot reverse-engineer safety from a tool that may not even be running.
Third, and most practically: you have no control over which detector your school uses, and you have no realistic way to write specifically to beat a particular one. Students sometimes ask which third-party checker is "closest to" the institutional tool so they can pre-screen their own work — a question we address candidly in which AI detector is closest to Turnitin — but even a good proxy only estimates a probability, and chasing a lower score can push honest writers into contorting their natural style, which is its own trap. The energy is better spent elsewhere, which brings us to the part that actually protects you.
The privacy and appeal angle
Two things about your rights are worth holding onto, because they shift the ground from "how do I beat the machine" to "how am I treated as a person."
On privacy: when your writing goes through a detector, it is being processed by third-party software, and depending on the tool and the school's configuration, your text may be stored, added to a comparison database, or retained under the vendor's terms. Reasonable students want to know what happens to their words. Your school's data-handling and integrity policies, and sometimes the vendor's own documentation, are where this is spelled out. If it matters to you — and it is legitimate for it to matter — the library or the office that administers the tool can usually point you to the specifics. This is a fair question to ask, and asking it is not an admission of anything.
On appeals: this is the part that should reassure you and that too few students know. At any institution operating in good faith, a detector score is not a verdict. It is, at most, a flag that triggers a human process. You are generally entitled to be told what you are accused of, to respond, to present evidence, and in most systems to appeal an adverse finding. The high-profile disablings happened precisely because administrators took seriously that a number from a tool should never be allowed to function as an automatic conviction. If you are ever flagged, the score is the beginning of a conversation, not the end of one — and you are allowed to participate in that conversation with evidence of your own. If you have already been flagged over writing you did yourself, our guide on what to do when Turnitin flagged your essay but you did not use AI walks through the practical steps.
The one piece of honest advice that actually helps
So what should you actually do with all of this? Not try to identify the exact detector and reverse-engineer it. That is a losing game: you cannot reliably know the tool, you cannot control it, and writing to defeat a probability estimator tends to make honest work worse, not safer. The move that protects you regardless of which tool your school runs — or whether it runs one at all — is to keep evidence of your own process.
Process evidence is the record that a real human did real work over real time. It is the thing a fabricated score cannot argue with. Draft in an environment that preserves version history, so your document shows an evolution rather than appearing fully formed. Keep your notes, outlines, and the messy intermediate drafts. Save the sources you consulted and when. If your writing tool logs revision history, leave it on. None of this is about proving a negative to a machine; it is about being able to walk into a conversation with a human and show, concretely, how the work came to exist. A student who can produce a trail of drafts, notes, and timestamps is in a fundamentally stronger position than one who can only insist "but I really wrote it," no matter how sophisticated or absent the detector turns out to be.
That habit is the great equalizer across all the uncertainty this article has described. You may never learn the precise brand your university uses. It may be Turnitin's indicator, quietly enabled. It may be Turnitin with the indicator deliberately switched off. It may be a course-level judgment call that varies from one professor to the next. It may be, at a school that read the research and chose caution, effectively nothing at all. In every one of those worlds, the student who wrote honestly and kept the receipts is protected, and the student who tried to game an invisible tool is exposed. Figure out your school's setup as best you can — read the syllabus, read the integrity policy, ask the library, ask the instructor — but do not let the hunt for the detector's name distract you from the one form of insurance that works no matter what the answer is. Write your own work, keep the trail that proves it, and treat any score you encounter as a claim to be examined rather than a truth to be feared.