What Turnitin AI Percentage Is Safe to Submit? The Numbers People Actually Use

There is no universal "safe" number, and any page that gives you one is simplifying. But the numbers people actually use are not random. Three layers are in play: a Turnitin display rule (the asterisk), a set of commonly repeated reference points (20% / 10% / 0–5%), and the policy of your specific school, which is the only layer that decides anything about your submission.

This page is about the first two layers and how to find the third.

A score shown as *% is the closest thing Turnitin publishes to an official "low" band. In the 1% to 19% range, no score or highlights are attributed at all — the report shows an asterisk instead of a number. Turnitin's stated reason is avoiding false positives in that range.

Two practical consequences follow.

The report carries no highlights in this band, so there is nothing to point at — the low number is a display choice, not a verification.

And the band is wide: everything from 1% to 19% looks identical as an asterisk. A *% report tells you the score is under 20, not what it actually is.

The percentages floating around forums — 20%, 10%, 0–5% — come from two different places, and knowing which is which prevents most confusion.

The first place is Turnitin itself: the asterisk cut-off at 20% is a display rule the company publishes. When people say "under 20% is safe", what they are often actually referencing is the score band the vendor itself refuses to report as a number.

The second place is school and program policy. Universities set their own AI writing expectations, sometimes as a number in a policy document, sometimes as a working rule staff apply informally. There is no global registry of these numbers, and they change.

That is why advice that sounds universal — "below 15% in the UK", "below 20% in the US" — is really a summary of what people commonly report, not a published standard.

Most of the confusion comes from mixing four different kinds of number as if they were the same kind. They are not, and they do not deserve the same weight.

Only the second row decides anything about your submission. The first is real but is a display rule, not a safety line. The third is worth a question. The fourth is worth nothing, however confidently it is said.

The reasons differ so widely because the score itself is not a regulation — it is a number produced by a statistical model, and every institution decides separately what to do with it.

Turnitin's own documentation says the indicator should not be used as the sole basis for action or a definitive grading measure by instructors. The company also runs its false-positive testing with a specific target: under 1% for documents with over 20% of AI writing — in other words, for one fully human-written document out of every hundred, it might still be flagged.

So the same 25% score is read one way at a university that treats the indicator as a conversation starter, and another at one that treats it as a filter. The model is the same; the policy around it is not.

A score above your school's working threshold triggers something — usually a conversation, occasionally an automated review. It does not, by itself, constitute a finding against you. Turnitin's guidance is explicit that the percentage should not be used as the sole basis for action, and that determining whether misconduct occurred takes further scrutiny, human judgment, and the institution's own policies.

That distinction matters when you are reading your own report: a high number triggers a process, it does not settle one.

Two other caveats are worth knowing. For short documents — a few hundred words — the prediction is mostly "all or nothing", because the model is predicting on a single segment without overlap; some text that is a mix of AI and original could be flagged as entirely AI. And for submissions under 300 words, Turnitin's own release notes say the score is likely less accurate. Short flagged documents are the case where the number deserves the least weight.

Instead of asking what percentage is safe, ask who set the number you are worried about.

The three candidate answers map to three different actions:

1. It is the asterisk display rule you read about online — then there is no school policy involved, and the *% you see is Turnitin's own caution indicator, nothing about your school.
2. It is a number a classmate or a forum mentioned — then verify it against your program's actual guidance before acting on it.
3. It is your school's policy — then that is the number that matters, and asking your instructor or the office that handles Turnitin is the way to get the real one.

One more fact to keep straight: student-facing accounts do not show the AI indicator at all — only instructors and administrators see it. If your teacher shared a PDF report with you, you are seeing what the school sees; if you only have a number from a secondary service, you are seeing a different tool's estimate, not Turnitin's.

No single percentage is "safe to submit" everywhere, because the only number that decides anything is your school's own policy. The asterisk means 1–19% and is Turnitin's display rule. The round numbers people trade — 15%, 20%, 10% — have no published source behind them, however confidently they are quoted. And above any threshold, the score triggers a conversation rather than settling one. When someone gives you a safe number, ask which layer it came from — that question is worth more than the number itself.