Turnitin Publishes Two Different False Positive Rates. The One for Sentences Is 4%.

If you are looking at a report where a couple of sentences are flagged, the 4% is the chance that any one of those highlighted sentences is actually human-written. The flag on that particular sentence might be wrong, and "around 4%" is how often that happens. It is a per-sentence number, not a per-document number, and the two cannot be derived from each other.

Turnitin's Chief Product Officer Annie Chechitelli described it this way in June 2023:

"Our sentence-level false positive rate is around 4%. This means that there is a 4% likelihood that a specific sentence highlighted as AI-written might be human-written. The incidence for this is more common in documents that contain a mix of human- and AI-written content, particularly in the transitions between human- and AI-written content."

The 4% attaches to a sentence that is already highlighted. It tells you something about the accuracy of an individual flag, not about how many flags you will see across a whole document.

Turnitin publishes a separate, lower rate for documents:

"Our document false positive rate - incorrectly identifying fully human-written text as AI-generated within a document- is less than 1% for documents with 20% or more AI writing."

Two things stand out. The rate describes a whole document being wrongly flagged as containing AI text when it is fully human-written. And the <1% figure is specifically conditioned on documents that contain 20% or more AI writing, meaning mixed documents where some AI text is genuinely present.

Here is the part that is easy to get wrong, and worth being precise about: neither published figure is scoped to a paper you wrote entirely yourself.

The 4% is described as a characteristic of mixed documents, "particularly in the transitions between human- and AI-written content." A paper with no AI text in it has no such transitions, so the scenario Turnitin attaches the 4% to does not describe your situation.

But the <1% does not step in to cover you either. Read its own wording again: it is "less than 1% for documents with 20% or more AI writing." That figure carries a condition, and a document with no AI writing in it does not meet the condition.

So both numbers come with a scope, and a fully human paper sits outside both. That is worth knowing before someone quotes either one at you as though it settled the question.

Turnitin gave a specific data point about where these wrongly flagged sentences show up:

"As explained in my earlier article, there is a correlation between these sentences and their proximity in the document to actual AI writing. 54% of the time, these sentences are located right next to actual AI writing."

"These sentences" means the false positives: the human-written sentences that got flagged anyway. Over half of them sit immediately next to genuine AI-written text. This lines up with the earlier statement about transitions between human and AI writing being where false positives concentrate.

Which is why the statistic cuts the other way for you. If the flags are scattered through a paper with no AI text anywhere in it, the mechanism Turnitin is describing here is not the one producing them. It is an account of where its errors cluster in mixed documents, and nothing more.

The same post is explicit that a highlight is a starting point, not a finding — and it says so to instructors, which is who you want reading it:

"Consider highlighted sentences as areas of interest because they're predicted to be close to where AI writing is present. But a small percentage of times, the AI model could get it wrong. So, use the information to initiate a conversation, not to draw a conclusion."

It also rules out the thing students most often assume they are being measured against. There is no number you are supposed to hit:

"Remember that there is no 'right' or 'target' score with the AI writing indicator, just like with a Similarity Score."

None of the above tells you whether you are in trouble, because no published number does. What it does give you is a way to keep the conversation on the report itself rather than on a percentage.

• Ask which passages, and get the report rather than a screenshot. Two highlighted sentences and a document-level number are different objects. The highlights are the only part with a location you can point at.
• Ask which number is being quoted at you, and read its condition aloud. The under-1% figure carries "for documents with 20% or more AI writing" in its own sentence. If that condition does not describe your paper, say so.
• Take Turnitin's own instruction with you. It tells instructors to use a highlight to start a conversation, not to reach a conclusion. That sentence is addressed to the person holding your report, and it is quoted above word for word.
• Keep what shows how the paper was written — drafts, version history, notes, the reading you did. That is evidence about authorship. A percentage is not.