Direct answer
An AI checker does not read images the way it reads text. Turnitin's AI writing detection works on the extractable text of a submitted document, so a photo, a diagram, or a screenshot of writing is not scored as AI-generated prose [1]. What the detector evaluates is the language in your document — sentence rhythm, word choice, and statistical predictability — not pixels [2].
Can an AI Checker Detect AI in Images or Screenshots?
No — mainstream academic AI checkers do not analyze the visual content of an image. Turnitin's AI writing indicator is built on a language model that examines the text of a submission and estimates how much of it was generated by AI [1]. A screenshot is a picture file; there is no prose for the model to score, so no AI percentage is produced from the image itself [2].
Where students get confused is the difference between "AI image detection" and "AI writing detection." Tools that claim to spot AI-generated artwork use entirely different signals — pixel artifacts, frequency patterns, and generator fingerprints — and those tools are not the ones your university runs on essays. Turnitin's report answers one question: how much of this written text looks machine-generated [1].
There is also a practical trap. If you paste a screenshot of ChatGPT output into your essay, the checker does not "see" the screenshot, but the surrounding text you typed or pasted still counts. Detection follows text, not file type [2]. So the honest answer is narrower than most students expect: images are not scored, but the writing around them is.
Does Turnitin Detect AI-Generated Text Inside Images or Scanned PDFs?
Generally, no. Turnitin processes submissions by extracting text from supported document formats, and a scanned or image-only PDF contains no machine-readable text layer to extract [3]. When the text cannot be pulled out, there is nothing for the AI writing model to evaluate, and the AI indicator is not meaningfully applied to that content [3].
This matters for two very different groups of students. If you submit a photo-heavy portfolio, a lab report with figure captions, or a design document, the prose you typed is what gets analyzed — the images themselves are outside the AI writing model's scope [3]. If you submit a scanned PDF of a handwritten or printed assignment, the detector may return little or no AI signal simply because it could not read the page, not because the writing was cleared [1].
That second case is where students misread their own reports. A missing or low AI score on a scanned file is not proof of originality; it is a sign the tool had nothing to work with [3]. The same logic applies to text baked into infographics, tables saved as pictures, and slides exported as images. If your assessor needs the text checked, it has to exist as text.
How Can Students Verify Their Own AI Score Before Submitting?
The reliable path is to check the actual document you plan to submit, in a text-based format, before the deadline. Turnitin's own guidance for students is to understand what the similarity and AI reports show and to review flagged passages while there is still time to revise [4]. That means submitting a .docx, .pdf with a real text layer, or .txt — not a screenshot of your draft.
Reading the report correctly is half the skill. Turnitin displays an asterisk (*%) instead of an exact figure when its AI detection falls below the 20% confidence threshold, which signals a low-confidence result rather than a guaranteed clean score [1]. Students who expect a precise number often panic at the asterisk or, worse, assume it means zero [4].
The second half is knowing what to do with a flag. A high AI indicator does not automatically mean misconduct — heavily templated academic phrasing, translated text, and grammar-tool edits can all shift the signal [2]. The productive response is to rewrite the flagged passages in your own voice, keep your citations and headings intact, and re-check the revised file rather than guessing [4]. Previewing the report turns an invisible risk into a concrete editing task.
If you would rather see the real report than guess at it, turnitin0 lets you upload your draft and preview the same Turnitin AI and similarity reports your instructor will open — so you know exactly which paragraphs need work before the deadline does.
※ Turnitin0.com - Actual Turnitin AI Report Cover, Score, Flag And Similarity Summary
FAQ
Can I upload a JPG or PNG to an AI checker?
Not usefully for AI writing detection. Turnitin extracts text from supported document formats, and an image file has no text layer to analyze [3]. Convert the content to a text document first if you want a score.
Does Turnitin detect AI in images I include in my essay?
No. The AI writing model scores the prose in your submission, not the pictures, charts, or screenshots inside it [1]. Your written analysis around those images is what gets evaluated [2].
Why does my report show *% instead of a number?
Turnitin shows an asterisk when AI detection falls below its 20% confidence threshold, meaning the signal is too weak to report as an exact percentage [1]. Treat it as a low-confidence result, not a certified clean score [4].
Can a scanned PDF hide AI-generated text from detection?
It can prevent extraction, which means the text is not properly evaluated rather than proven original [3]. Submitting a text-based file gives you a real, interpretable result instead of a blank one.
If my AI score is high, does that mean I cheated?
No. Flags indicate statistical patterns in the writing, and legitimate editing, translation, or templated phrasing can raise them [2]. Review the flagged passages, revise in your own words, and re-check the file [4].