Which Universities Stopped Using Turnitin's AI Detector

The list of universities that stopped using Turnitin's AI-detection feature is shorter, and better documented, than search results make it look. Three US institutions, Vanderbilt University, Michigan State University and the University of Texas at Austin, have each published their own page confirming they switched the feature off, stating their reasons in their own words. None of the three dropped Turnitin itself.

That last sentence is the part most roundups skip. Turnitin offers two products in one interface: a similarity checker that compares student papers to a text database, a service that has been available since long before generative AI was invented, and an AI writing indicator, a classifier that scores a paper's prose on how much it resembles machine output. This new feature went live in April 2023. Only the second product was turned off by every institution below.

What follows is the reasoning three universities gave in their own words, two more names that press coverage repeats without an institution-owned page to confirm them, and what an administrator is actually changing when they make this call.

Did Any University Stop Using Turnitin Completely?

No confirmed case of that exists. Every documented example is a university disabling one specific setting inside a product it otherwise keeps using. Turnitin's own guidance explains why that distinction is even possible: administrators "have the option to enable/disable the AI writing feature from their admin settings page," and doing so "will remove the AI writing indicator and report from the Similarity report," leaving the similarity report itself running.

The Similarity Score and the AI Score are two separate, independent reports, in Turnitin's own words, built on different models and shown on different tabs of the same interface. TextPulse has covered how the similarity score and the AI score actually differ in more detail elsewhere. A university can turn off the second without touching the first, and every case below did exactly that.

Vanderbilt University: The Most Detailed Public Explanation

Vanderbilt disabled Turnitin's AI detector on August 16, 2023, and published its reasoning the same day rather than leaving journalists to piece it together. The university's own arithmetic is the clearest version of the argument any institution has made public: Turnitin had claimed a 1 percent false positive rate, and against the roughly 75,000 papers Vanderbilt submitted to Turnitin in 2022, the university calculated that rate meant "around 750 student papers could have been incorrectly labeled as having some of it written by AI."

Vanderbilt gave two further reasons alongside that arithmetic: documented bias against non-native English writers, and insufficient transparency into how the underlying model actually reaches its score. Those three threads, a false positive count too large to wave off, a bias problem, and an opaque model, reappear in some form at nearly every institution that has explained an opt-out in public since.

Institution	Action	Date	Stated reason, in the institution's own words
Vanderbilt University	Disabled the AI writing indicator account-wide	August 16, 2023	A claimed 1% false positive rate applied to its own submission volume; bias against non-native English writers; lack of model transparency
Michigan State University	AI writing detection not made available to instructors	Guidance dated August 2023	AI detection tools should not be used as the sole basis for adverse action against a student
University of Texas at Austin	Bans use of AI detection software on student work	No date stated on UT Austin's own page; listed among opted-out institutions by February 2024	Classifies AI detection software as High Risk; cites copyright, intellectual property and FERPA concerns

Michigan State and UT Austin: Two More Confirmed Opt-Outs

Michigan State's own help documentation states plainly: "Turnitin AI Detection is not currently available at MSU." Its guidance adds a caution that applies regardless of which detector a course uses: AI detection tools "should not be used as the sole basis for adverse actions against a student." The page sits inside MSU's broader generative-AI guidance, dated from August 2023, the same month Vanderbilt made its own decision public.

UT Austin took the more sweeping position of the three. The university's Office of the Provost states outright that it "does not endorse the use of AI Detection Software and does not have any central contracts or purchase orders with active AI Detection Software features." The page goes further than a simple opt-out: it classifies AI detection software as "High Risk," meaning a department cannot buy one even with a procurement card, and it warns that running a student's work through such software "may be a violation of that student's copyright and intellectual property rights," alongside separate FERPA concerns. UT Austin's own page states no specific decision date; education-press coverage already had the university on its list of opted-out institutions by February 2024.

Other Universities That Stopped Using Turnitin, According to the Press

There're two others that show up often when rounding up coverage of this story: Northwestern University and Montclair State University. Both exist, but neither is cited by the same level of credible sources as the other three. In February 2024, Inside Higher Ed reported that Montclair "announced in November [2023]... that academics should not use the AI-detector feature in a tool from Turnitin," following "similar moves from institutions including Vanderbilt University, the University of Texas at Austin and Northwestern University.".

That is solid trade-press sourcing, but it is a different standard from an institution's own page stating a decision in its own words, and no such page turned up for either university while researching this piece. The gap is worth naming rather than smoothing over. A list built from every press mention would run longer than three names; a list built from what an institution has actually published for itself currently stops at three, which is the stricter standard this piece keeps to throughout.

Why Are Universities Turning the AI Detector Off?

The same handful of reasons recur across every public statement above: a false positive rate that stops looking small once it is multiplied by an actual student population, documented bias against non-native English writers, and a classifier whose internal workings the vendor has not fully disclosed. The reasoning lines up with the wider evidence on whether AI detectors are accurate at all, not just at Turnitin. None of the three institutions above disputes that Turnitin's similarity checker still does its original job; the objection is specific to one added feature.

None of this means a university that keeps the feature switched on is acting carelessly, and none of it means a university that switched it off has solved the underlying problem. It means two groups of administrators looked at the same published numbers and reached different judgment calls, and a student's experience of this technology now depends partly on which university they attend rather than on anything about the paper they submitted.

A writer wondering whether their own draft carries the statistical profile Vanderbilt's bias objection describes can check it directly with a free perplexity checker rather than wait to find out which policy their own university happens to have chosen.

The evidence behind those decisions is set out separately: what a false positive actually is, and the false positive rate Turnitin itself publishes.

TextPulse's free tools cover that same ground before any of it reaches a detector at all. Turnitin will keep adjusting its own thresholds as it retrains the model; whether a given university is currently running the feature is a setting, not a fact about the paper itself, and it is worth checking a draft on its own terms regardless of which way that setting happens to be flipped this semester.