Universities that turned off Turnitin’s AI detection, and why

Why this list exists

When a department debates whether to keep the AI indicator on, the first question is who else has turned it off and why. The answer is scattered across Brightspace announcements, provost notices and news reports. This page collects the ones we could verify. It is not an argument for or against any institution’s choice; it is the record.

The institutions

Source: Vanderbilt Brightspace statement; Waterloo announcement; Curtin notice; ABC News Australia; GradPilot tracker; all checked 2026-09-08. Public trackers count 60-plus institutions; only entries with a statement or report we checked appear above.

How to read this list

Three things stand out. First, the decisions span three years and three countries, which suggests they were made on local testing and local caseloads rather than on one news cycle. Second, where a reason is on record it is the same reason: the indicator flagged writing that people had written. Waterloo’s internal test is the clearest statement of it. Third, the entries for Northwestern, Yale and Johns Hopkins carry imprecise dates because the reports we found were secondary. We list them because the fact of the decision is well reported, and we flag the dates as approximate rather than invent a month.

What Turnitin says, for balance

Turnitin publishes a page titled “Understanding false positives within our AI writing detection capabilities” and states a document-level false positive rate under 1 percent for documents with more than 20 percent AI writing. Turnitin also reported on February 24, 2026 that 15 percent of submissions between October 2025 and February 2026 were more than 80 percent AI, against 3 percent at launch in 2023, and that 19.4 percent of US higher-education submissions in the second quarter of 2026 were heavy-AI. Both things can be true: the indicator can be broadly accurate and still produce a number of false flags that an institution decides it cannot defend one student at a time. The published research on who gets flagged, collected on the false positive rates page, is the other half of that calculation.

Source: turnitin.com; Turnitin press materials; checked 2026-09-08.

What replaces a score

Our sources do not record what each institution adopted instead, and this list does not guess. The practices in use elsewhere are familiar: assignments that ask for the process as well as the product, writing done in class or in a recording editor, short oral follow-ups, and syllabus language that says which tools are permitted. A keystroke-level record of the writing session shows what a detector cannot: whether a passage was typed or pasted, where the pauses were, and the order in which the argument formed. Its limit is equally plain. It shows how text entered the editor, not whether the ideas were the student’s, and it is evidence for a conversation rather than a verdict. The guide to process-based assessment covers the practice.

Add to this list

If your institution has turned the indicator off, or turned it back on, and there is a statement, notice or news report we can cite, send it through the contact page. We add rows only with a checkable source, and we date every change.