Direct answer
A "ChatGPT detector image" is the visual output of an AI-writing detector: a report screenshot showing an overall AI score, a colour-coded overlay of flagged passages, and, in Turnitin's case, a similarity summary alongside it. Turnitin's AI writing report renders the score at the top of the document with highlighted segments you can click into, and it displays *% instead of a number when the score falls below its 20% confidence threshold [1]. Detectors analyse prose, not pictures — the "image" you are looking for is the report, not a scan of a JPEG [1].
What Does a ChatGPT Detector Report Actually Look Like?
A Turnitin AI writing report is a two-layer document. The top layer is a summary panel: the overall AI score as a percentage, plus a breakdown of how much of the submission the system classified as AI-generated versus AI-paraphrased [2]. The second layer is the submission itself, with flagged passages highlighted in colour directly over your text, so you can read the sentence that triggered the flag rather than guess at it [2].
Clicking a highlighted segment in the report reveals the predicted class for that passage and the confidence behind it. That is the detail most students miss when they only look at the headline number: a 34% overall score is rarely spread evenly — it usually concentrates in a handful of paragraphs, and those are the ones an instructor will open first [2]. The report also records which model class the text most resembles, which is why a draft written partly in ChatGPT and partly by hand can show a mixed pattern rather than a single block of red [2].
The report is delivered through the assignment inbox, so what you see when you preview a report is structurally what your instructor sees — same score, same highlights, same layout [2]. That equivalence matters, because a screenshot of an "AI score" pulled from an unaffiliated checker often uses a different scale, a different threshold, and a different visual language, none of which map cleanly onto the Turnitin report your marker will actually open [2].
Why Does Turnitin Show an Asterisk Instead of a Percentage?
The asterisk is a confidence signal, not a score of zero. Turnitin displays *% when the overall AI writing percentage sits below its 20% confidence threshold, meaning the detector did not find a strong enough signal to commit to a number [3]. It is the report's way of saying "low confidence" rather than "clean" — a distinction that matters enormously when you are deciding whether to submit as-is [3].
Students routinely misread *% as a pass. It is better read as an absence of strong evidence in the prose the detector could assess, which is a much narrower claim than "this was written by a human" [3]. The threshold exists because short submissions, heavily quoted material, and text with unusual structure produce weaker signals, and Turnitin would rather flag uncertainty than publish a misleadingly precise figure [3].
In practice, *% is the outcome most students are trying to reach — but it should be reached by revising the flagged passages, not by assuming the report cleared you. If your draft shows *% and you know a large section was drafted by ChatGPT, the honest reading is that the detector's confidence was low, not that the section was undetectable [3].
How Can You See Your Own Turnitin AI Report Image Before You Submit?
Most universities do not give students a preview of the AI writing report before the graded submission — the report appears in the instructor's inbox, and whether you can see it depends on course settings your module lead controls [4]. That is the gap students are trying to close when they search for a detector image: they want the report before the deadline, not after the mark is released [4].
The productive use of a report image is diagnostic, not decorative. Read the highlighted segments, identify why those specific sentences read as machine-generated — uniform sentence length, generic transitions, hedge-free assertions — and rewrite them in your own voice with your own evidence [4]. Resubmitting the same text after cosmetic edits usually reproduces the same pattern, because the underlying rhythm of the prose has not changed [4].
It also helps to know what the report cannot tell you. Turnitin's AI detection assesses prose; it does not evaluate figures, diagrams, or images, so a "ChatGPT detector image" search that is really about AI-generated pictures is asking a question the writing report does not answer [4]. For written work, the report is a revision tool: it points at passages, and the judgement about what to do with them stays with you [4].
If you want to see that report image on your own draft — the score panel, the highlighted flags, the similarity summary — before it lands in your instructor's inbox, turnitin0 lets you preview it on a file you upload yourself.
※ Turnitin0.com - Actual Turnitin AI Report Cover, Score, Flag And Similarity Summary
Get Real Turnitin AI & [Similarity Report](https://www.turnitin0.com/guides/us/ai-tokens)
FAQ
Can a ChatGPT detector tell whether an image was AI-generated?
No — Turnitin's AI writing detection assesses prose, and its report covers text, not figures or images [1]. Detectors built for pictures are a separate class of tool with their own accuracy limits.
What does *% mean on a Turnitin AI report?
It means the overall AI score fell below the 20% confidence threshold, so Turnitin shows an asterisk rather than a precise percentage [3]. It signals low confidence, not a guaranteed zero.
Is the report image I see the same one my instructor sees?
If it comes from Turnitin's own report, yes — the score, highlights, and layout are generated from the same report structure delivered to the assignment inbox [2].
Why is one paragraph highlighted and the rest is not?
Flags track where the detector's signal concentrates, so a draft can show a high overall score driven by a small number of passages [2]. Those passages are the ones worth rewriting first.
Can I lower a flagged score by editing a few words?
Cosmetic edits usually leave the underlying sentence rhythm intact, so the same passages tend to re-flag [4]. Rewriting the flagged sections in your own voice, with your own evidence, is the more reliable route.