Which universities disabled Turnitin's AI detector?
Vanderbilt disabled it in 2023 and published its reasoning; Yale, Georgetown, the University of Pittsburgh, Johns Hopkins, the University of Alabama and Curtin have since turned it off, and Washington State University cancelled its contract outright in February 2026.
Last reviewed 15 August 2026 · The HumanFlow team
Vanderbilt is the case worth reading rather than just citing, because it explained itself publicly rather than quietly changing a setting. Its Brightspace post set out why the university did not consider the indicator reliable enough to use, at a point when almost nobody in the sector was saying so on the record.
Washington State is the strongest signal, because cancelling a contract costs money and required a provost's office memo to instructors. That is an institution deciding the tool was not merely imperfect but not worth having.
Others stopped short of removal and still declined to rely on it. UC Berkeley piloted the tool and opted out; NYU's provost's office has said it does not believe any current detector works well enough to license.
The list matters in an appeal for a narrow, useful reason: it establishes that treating these scores cautiously is a mainstream institutional position rather than a student's excuse. It does not bind your institution, and citing it as though it does will not help you.
When this answer changes
It changes as institutions revisit the decision in both directions, and this page will lag. Check the current position published by the specific institution rather than relying on a list — including this one.
It also changes nothing about whether your own work was flagged correctly. An institution elsewhere switching the tool off is context, not evidence about your document.
Where to go next
- If you have been falsely accused — What to do today, and what to say — where this list is one supporting argument among several.
- Appealing an AI detection finding — Once a written decision exists: stages, evidence and the deadline that matters most.
- What Turnitin's AI detection measures — The tool itself — the 20% threshold, the asterisk band, and its own stated limits.
- Writing the policy instead — The three decisions that determine whether a policy is usable, with published prior art.
Sources
One of our direct answers.