Turnitin0

Check AI Cards

Direct answer

An "AI card" is the AI-writing indicator that Turnitin attaches to a submitted document: an overall percentage plus inline highlights on the passages the detector believes were machine-generated [1]. To check it, you need a report generated from your own draft before it reaches your instructor's assignment inbox, then you read the overall figure together with the highlighted segments rather than the headline number alone [1]. The score is a probability estimate, not a verdict, so it should be read alongside your drafting history and your institution's AI policy [2].

What Does an AI Card Actually Show, and How Do You Read It?

When a document is processed, the detector splits the text into segments, estimates for each segment how likely it is to have been generated by a large language model, and then aggregates those estimates into a single overall figure shown at the top of the report [2]. That overall figure is the number students usually quote as "my AI score," but it is only a summary of many sentence-level judgements. The more useful layer sits underneath it: the inline highlights that mark exactly which sentences drove the score up [1].

Reading the card well means reading it in three passes. First, note the overall percentage and whether it is a plain number or an asterisk. Second, open the highlighted text and ask whether those specific sentences are the ones you actually wrote with AI assistance, or whether they are formulaic passages — definitions, standard methodology phrasing, quoted policy language — that a detector can misread [2]. Third, compare the flagged passages against your own version history: an essay you drafted in a document with tracked edits has a recoverable trail that explains the pattern far better than the score does [4].

Two interpretive traps are worth naming. The first is treating a high score as proof of misconduct; detection output is an estimate produced by a statistical model, and Turnitin itself frames the report as a signal for human academic judgement rather than a finding [2]. The second is treating a low score as a clean bill of health, which ignores the possibility that paraphrased AI text slipped under the threshold. Neither number replaces the question your marker will actually ask: can you account for how this text was produced [1][4]?

Why Does the Report Sometimes Show an Asterisk Instead of a Percentage?

An asterisk appears when the detected AI share falls below Turnitin's 20% confidence threshold, which is the point at which the system stops publishing a precise figure [3]. It is a deliberate design choice: below that level the model's estimate is not reliable enough to be presented as a number, and showing "4%" would imply a precision the underlying signal does not support [3]. So the asterisk is best read as "low-confidence signal detected," not as "zero AI" and not as "clean" [3].

This matters practically because the asterisk is the single most misread element of the whole report. Students who see it sometimes assume the document passed cleanly and submit without further thought, while others panic because a symbol appeared at all. Neither reaction matches what the threshold means [3]. If you want to know what sits behind the asterisk, the highlighted segments are still the place to look — they show whether anything was flagged and how much text was involved [1].

The asterisk also interacts with how your institution reads reports. Because the number is withheld rather than displayed, a marker reviewing your submission sees the same low-confidence signal you do, and the conversation shifts from "what is your percentage" to "what does this flag mean for this passage" [3][2]. Preparing for that conversation is more useful than chasing a decimal point.

How Can You Check Your Own Draft's AI Card Before You Submit?

The only way to see your own AI card before submission is to run the draft through a detector and read the report it returns. Turnitin's own guidance to students centres on transparency and process: keep your drafts, keep your revision history, know your institution's policy on AI assistance, and be able to explain your writing choices if a flag appears [4]. A pre-submission check simply gives you that explanation material earlier, while you can still revise.

When you run a check, match the conditions your instructor will use as closely as you can. Upload the same file type you intend to submit, keep the word count inside the range the checker supports, and read the AI report and the similarity report together — a passage that is flagged for AI and also matches a source needs a different fix than a passage flagged for AI alone [1]. Then work through the highlighted sentences one at a time: rewrite the ones that genuinely came from a model, and leave the ones that are your own phrasing even if they were flagged, because over-editing your authentic prose to satisfy a detector damages the writing you were graded on [2].

Finally, distinguish between the two things a check can tell you. It can tell you what a detector currently outputs for this exact file. It cannot tell you what your marker will conclude, because the report is one input into a human judgement that also weighs your process, your citations, and your institution's rules [2][4]. Students who hold that distinction tend to use the report as a revision tool rather than a verdict, which is exactly what it is designed to support [1].


Reading the highlights is only useful if you can act on them before the deadline — and the report you generate yourself is what makes that possible. turnitin0 produces the same Turnitin AI detection report and similarity report your instructor sees, delivered as two downloadable PDFs from a single check, so you can inspect the overall figure, the flagged passages, and the match summary while there is still time to revise. Documents are checked without being added to Turnitin's student paper database, turnaround averages under 15 minutes, and 100,000+ reports have been delivered to 20,000+ students. A single check is $3.80, with no subscription required.

※ Turnitin0.com - Actual Turnitin AI Report Cover, Score, Flag And Similarity Summary

Get Real Turnitin AI & Similarity Report

FAQ

Is an AI card the same thing as a plagiarism report?
No. The AI writing report estimates whether text was machine-generated, while the similarity report matches your text against sources in a database [1]. They are separate reports, and a document can be flagged on one and clean on the other.

Does an asterisk mean my document has no AI content?
No. An asterisk means the detected share fell below the 20% confidence threshold, so the precise number is withheld as low-confidence [3]. It signals a weak AI signal, not a confirmed absence of AI text.

Can a detector flag writing I did entirely myself?
Yes. Highly formulaic prose — standard definitions, templated methodology, quoted policy wording — can be misread by a statistical model, which is why reports are meant to be interpreted with human judgement [2].

Should I rewrite every highlighted sentence?
Only the ones that genuinely came from a model. Rewriting your own flagged phrasing to satisfy a detector can flatten your voice, and your drafting history is the stronger evidence of authorship [4].

Can I see the report before my instructor does?
Yes, by running your draft through a pre-submission check and reading the AI and similarity reports it returns [1]. The report reflects the detector's output for that file; your marker's conclusion remains a separate human judgement [2].

Sources

  1. Using the AI Writing Report — https://guides.turnitin.com/hc/en-us/articles/22774058814093-Using-the-AI-Writing-Report
  2. The Turnitin AI Writing Detection Feature — https://guides.turnitin.com/hc/en-us/articles/28477544839821-The-Turnitin-AI-Writing-Detection-Feature
  3. Why does the AI writing report show an asterisk? — https://helpcenter.turnitin.com/hc/en-us/articles/27811948436237-Why-does-the-AI-writing-report-show-an-asterisk
  4. How to cite AI tools and ChatGPT — https://www.turnitin.com/blog/how-to-cite-ai-tools-and-chatgpt

Related articles

Contact us

Email us or reach us on WhatsApp. We typically reply within business hours.