Direct answer
A Turnitin AI detection image is a screenshot or export of the AI Writing Report: a score panel showing the overall percentage of qualifying text flagged as AI-generated, plus a document view where individual sentences are highlighted. It is a separate report layer from the Similarity Report, and it is generated only for long-form text submissions that meet Turnitin's minimum length requirements [1]. Reading one correctly means treating it as a set of navigation cues rather than a verdict, because the colour overlays point you to passages that still need human judgement [2].
What Does a Turnitin AI Detection Report Image Actually Show?
The image is built from two distinct areas that are easy to confuse when you only see a cropped screenshot. The first is the score panel, which reports the proportion of the submission's qualifying text that the detector flagged, alongside the report type label so you can tell which layer you are looking at [2]. The second is the document view, where the submitted prose is reproduced with coloured overlays running across the sentences that contributed to that score [1]. On a genuine report both areas appear together, which is why a bare percentage floating on a white background is a poor sign of authenticity [4].
The overlays are a navigation aid, not a conclusion. Turnitin's own guidance frames the highlighting as a way to locate passages for review, and the reader is expected to open each flagged passage and judge it in context rather than accept the colour as proof [2]. This matters because the AI layer and the similarity layer are toggled independently, so two screenshots of the same submission can look completely different depending on which layer was active when the image was captured [2]. If someone shows you an image with no visible panel labels, you cannot tell which report you are looking at.
The image also only exists for certain files. Detection runs on long-form text submissions, so a short document, a spreadsheet, or an image-only PDF produces no AI Writing Report at all — there is nothing to screenshot [1]. That single constraint explains a large share of student confusion: a missing report image is usually a file-type or length issue, not evidence that the detector found nothing wrong.
Finally, the score panel has one display quirk worth memorising. When the detector's signal falls below its confidence threshold, the report shows *% instead of an exact figure, which signals a low-confidence result rather than a zero [1]. Students who read *% as "0% AI" misread the image in the most consequential direction possible.
How Do You Read the AI Score and Highlighted Flags in a Turnitin Report Screenshot?
Start with access, because who can see the image determines what you are looking at. Students see the AI Writing Report only when their instructor has enabled student access, so a student-facing screenshot and an instructor-facing screenshot can differ in what is exposed even when the underlying data is identical [3]. When you do have access, read the percentage as the share of qualifying text that was flagged — not as a probability that the author cheated [3]. Those are different claims, and the second one is not what the number means.
Next, work through the flagged sentences one at a time rather than scanning the colour. Turnitin documents false positives on formulaic prose, heavily templated academic phrasing, and text that closely follows quoted or standardised wording, which is precisely the register much of academic writing occupies [3]. A highlighted methods paragraph or a definitions section is therefore weak evidence on its own. The productive question is whether a specific flagged sentence could plausibly have been written by the student in that voice, and whether the surrounding unflagged sentences support or undercut that reading.
Then check the arithmetic of the panel against the document. A high overall percentage with only a handful of highlighted sentences usually indicates a short submission where a few sentences dominate the denominator, whereas a moderate percentage spread across a long document tells a different story [3]. Reading the image this way turns it from a scoreboard into a diagnostic.
The final step is the one students skip: take the image to a conversation. Turnitin's published guidance on detection accuracy is explicit that the score is one signal among several and that flagged passages should be discussed with academic staff rather than treated as a final judgement [4]. An image you have interpreted carefully makes that conversation short and specific.
Where Can Students Get a Real Turnitin AI Detection Report Image for Their Own Draft?
The normal workflow is instructor-led: the report is generated server-side within the institution's system, and the person who opens it is usually the marker, not the author [4]. That is why genuine report images are hard to find in advance — you cannot generate one for your own draft through Turnitin directly, and the images circulating online are other people's submissions with the content redacted or replaced. Because the report is produced by the platform rather than assembled by hand, a real image carries the platform's own panels and labels instead of a standalone percentage graphic [4].
What you can do pre-submission is reproduce the same two-part structure on your own draft: a score panel and a document view with the flagged passages located. Turnitin0 provides exactly that for students in the US, UK, Canada, Australia, New Zealand, and Ireland — you upload a .docx, .pdf, or .txt file (English, 300–30,000 words, under 20 MB) and receive two downloadable PDFs in one checkout: a Turnitin AI detection report and a similarity report, laid out the way your institution's system presents them. The check is non-repository, so your file is not added to Turnitin's student paper database, and turnaround is under 15 minutes in 98% of cases with delivery guaranteed within 30 minutes even during queue spikes.
The practical advantage is timing. Seeing a report image before the deadline lets you act on it — revise the passages that read as machine-generated, or run flagged text through the AI humanizer, which rewrites flagged passages while preserving meaning, citations, headings, and .docx formatting, with a score promise of *% or 0% on ChatGPT, Claude, or Gemini drafts. Turnitin0 has delivered 100,000+ reports to 20,000+ students, holds a 4.9/5.0 satisfaction rating, and is an independent service not affiliated with Turnitin, LLC. Pricing is pay-per-use with no subscription: $3.80 for a single check, or $2.75 per check in a 10-check pack.
Reading the image is only half the job — the other half is having one for your own draft while there is still time to change it. Turnitin0 shows you the report exactly as your marker will see it, so the score panel and the highlighted sentences stop being a mystery you meet after the deadline.
※ Turnitin0.com - Actual Turnitin AI Report Cover, Score, Flag And Similarity Summary
FAQ
Is a Turnitin AI detection image the same as a similarity report?
No. The AI Writing Report is a separate layer with its own score panel and highlighting, and the two layers are toggled independently [1]. A screenshot only shows whichever layer was active when it was captured.
Why does my report image show *% instead of a number?
Turnitin displays *% when the AI signal falls below its confidence threshold, which indicates a low-confidence result rather than a confirmed zero [1]. Treat it as "insufficient signal", not as a clean score.
Can a highlighted sentence prove a student used AI?
No. Turnitin documents false positives on formulaic and heavily templated prose, and its accuracy guidance describes the score as one signal among several for staff to weigh [3][4]. The highlighting locates passages for review; it does not settle authorship.
Why can't I see an AI detection image for my own submission?
Student access to the report depends on the instructor enabling it, and detection only runs on qualifying long-form text files [1][3]. If either condition is missing, there is no report image to view.
What file types produce an AI detection report?
Long-form text submissions such as .docx, .pdf, and .txt in English, within Turnitin's length limits [1]. Files that fall outside those limits produce no AI Writing Report at all.