Turnitin flagged my work and I did not use AI
It happens at a published rate of 4%, it is not random, and the writing it happens to has a specific shape. Here is that shape.
The first thing worth knowing is that this is a documented, quantified event rather than a freak occurrence. Turnitin publishes the figure themselves: 4% of genuine human writing gets flagged.
At a department processing 10,000 submissions a term, that is hundreds of documents. The rate is known. Which specific document it lands on is not, and no version of the tool will say.
Why it happened to yours
Detectors read two things: how predictable your word choice is, and how much your sentence length varies. Generated text is high on the first and low on the second. So is a lot of good human writing.
You wrote formally. Conventional phrasing is predictable phrasing.
Your sentences ran even. Human writing in our corpus varies sentence length with a standard deviation of 8.7 words, generated text with 6.4. Writing at a steady rhythm reads as polished to a person and as machine-made to a classifier.
You edited it hard. Every pass smooths. Smoothing removes exactly the variation that reads as human, so a well-revised piece scores higher than a rough one. See does Grammarly get flagged as AI.
You wrote it in a second language. Published evaluations keep finding elevated wrong-flag rates for non-native English writers, because learned English applies its rules more consistently. See why non-native English writing gets flagged.
You opened sentences with connectives. "Moreover", "Furthermore", "Additionally". These open 2.4% of human sentences and 5.8% of generated ones in our corpus. Taught as good structure, read as a marker.
None of these are mistakes. Several of them are things you were told to do.
What the flag is, precisely
A probability from a classifier that your text has properties characteristic of generated writing. Nothing was matched, nothing was found, and there is no source document to click through to.
That distinction matters because the similarity score next to it on the same report works completely differently, and the two get read as the same kind of thing. See how Turnitin AI detection works.
Useful things to know about the number
Below 20% means less. Turnitin states that documents scoring under 20% show a higher incidence of wrong flags, because a low score is built from a small amount of flagged text and small amounts are where a classifier is thinnest.
Wrong flags cluster. Turnitin found 54% of them sit directly beside a genuinely flagged sentence and 26% two sentences away. If any part of the document was assisted, the human writing around it gets pulled up too.
Detectors disagree. A second tool on the same text frequently returns something far apart, because thresholds and segmentation differ. Wide disagreement means the text sits near a boundary. See why AI detectors disagree.
What is actually in your hands
The report you cannot see. The writing you can.
Run the document through a second detector, look at which paragraphs come back flat, and you have a concrete picture of what the institution's tool was reacting to rather than a guess. See how to show you wrote your essay.
Ours is free and unlimited with no account, takes .pdf and .docx, scores every paragraph separately, and publishes its wrong-flag rate: 0.4%, against Turnitin's 4%.