Direct answer
Antiplagio (anti-plagiarism) is the practice and software layer that compares your writing against a database of published work, web pages and previously submitted student papers, then reports what matched and where. In a Turnitin workflow it produces two separate signals: a Similarity Report that highlights matching text and its sources [1], and an AI Writing Report that estimates how much of the text looks machine-generated. Neither number is a verdict on its own — both are evidence an instructor interprets, which is why reading your own report before the deadline is the single most useful antiplagio step you can take.
What does antiplagio (anti-plagiarism) checking actually measure in a Turnitin report?
Antiplagio checking is fundamentally a matching exercise. Your document is broken into text segments and each segment is compared against three broad pools: the public internet, academic and publisher content, and the repository of papers previously submitted to Turnitin by students at institutions worldwide. The Similarity Report then displays every segment that resembles something in those pools, with a link to the source it matched [1]. What you are looking at is therefore a map of overlap — not a judgement about whether you cheated.
The headline percentage is a summary, not a diagnosis. A 22% similarity score can be entirely benign if those 22% are correctly quoted and cited passages plus a reference list, and a 6% score can be serious if that 6% is one uncited paragraph lifted from a key source [2]. This is why Turnitin's own guidance stresses that the report must be interpreted in context by a human who knows the assignment, and why instructors are trained to open the sources behind each match rather than reading the top-line number alone [2].
What the report does not measure matters just as much. It does not detect whether you understood the material, it does not know your citation style, and it does not distinguish between a paraphrase that is too close to the original and one that is genuinely rewritten — both simply appear as matches. That distinction is left to your reader, which is exactly why the source list underneath the percentage is the part worth studying [1]. If you want to use the number at all, use it as a to-do list: every highlighted block is a sentence you should be able to defend.
Settings also change the arithmetic. Instructors can exclude quotations, bibliography and small matches below a word-count threshold, so the percentage you see in one course may differ from the percentage a friend sees in another for the same file [2]. When you compare your score to anyone else's, compare the underlying matches, not the percentage.
Why can a paper you wrote yourself still be flagged for similarity or AI writing?
Because detection systems read text patterns, not authorship. The AI writing indicator is generated by a classifier trained on the statistical texture of machine-generated prose, and it has no record of who typed your document or how long it took [3]. A paragraph you drafted yourself can resemble model output if it is highly formulaic — the same transitional phrases, the same balanced sentence rhythm, the same generic academic register that large language models were trained to produce. Students who write in English as a second language and students who heavily edit AI-assisted drafts are both over-represented in that pattern space, which is the honest reason false positives exist.
Similarity flags have a parallel explanation. Your bibliography, your cover page, your module code, your institution's required declaration, standard definitions of a theory, and the assignment brief you quoted back at the start of your introduction all belong to someone else's text or to a template. They match, legitimately. A paper that looks "flagged" is very often a paper whose matches are concentrated in exactly those structural places [2]. The fix is not to delete your references; it is to know which matches are structural and which are substantive.
Turnitin itself signals uncertainty rather than certainty in this area. When AI detection confidence falls below its 20% threshold, the report displays an asterisk instead of an exact percentage, indicating a low-confidence signal rather than a finding [3]. An indicator of that kind is meant to prompt a conversation between student and instructor, not to function as an automatic penalty. Understanding that framing protects you from panicking over a number that was never designed to be read as a verdict.
The practical consequence is that "I wrote it myself" is not a defence you can rely on after the fact — it is a claim you have to evidence before submission. Evidence means being able to point at each flagged passage and explain the source, the citation, or the reasoning behind the phrasing.
How can you preview your Turnitin similarity and AI report before you submit?
Previewing is the standard pre-submission workflow, and it is the one antiplagio habit that reliably changes outcomes. The cleanest route is a draft assignment: if your instructor has enabled one, you can submit a near-final version, read the report, revise, and submit the final file without the draft counting against you [4]. Ask early in the term whether a draft or originality-check assignment exists in your module, because availability varies by institution and by course settings.
Where no draft assignment exists, students commonly use an independent pre-submission check that runs the file through Turnitin without adding it to the student paper repository. That non-repository detail is the part to verify before you upload anything, since a check that stores your paper can itself create a future similarity match against your own work [4]. A useful preview returns the same two artefacts your instructor sees: a similarity report with highlighted matches and source links, and an AI writing report with the score and flagged segments [1][3]. Reading both together is what lets you tell a benign bibliography match apart from a genuine problem paragraph.
Work through the report in a fixed order and it takes minutes rather than hours. Start with the highest-percentage sources and confirm each one is quoted and cited; then look for long unbroken matches inside your own argument, which are the ones that actually need rewriting; then check the AI indicator and, if it is high, rewrite the flagged sections in your own voice rather than swapping synonyms [3]. Re-check after revision, because rewriting can move a similarity score in either direction [2].
Finally, treat the preview as a diagnostic, not a target to game. The goal is a report you can explain line by line, not a percentage you have optimised into the single digits by paraphrasing your sources into unrecognisable prose.
Reading your own report before the deadline is the difference between guessing and knowing — and turnitin0 exists to give you that read in the same format your instructor will see, with the AI and similarity PDFs delivered together, usually in well under fifteen minutes, and without your file entering the student paper database.
※ Turnitin0.com - Actual Turnitin AI Report Cover, Score, Flag And Similarity Summary
FAQ
Is antiplagio the same thing as plagiarism detection?
It is the broader term: antiplagio covers the whole practice of preventing plagiarism — citation discipline, paraphrasing properly, and checking your draft — while plagiarism detection is just the matching technology inside it [1]. The report is one input; your citation practice is the other.
Does a high similarity score mean I will be penalised?
Not automatically. Instructors read the sources behind each match, and correctly quoted, cited material is expected to match [2]. A high score becomes a problem when the matches are uncited or concentrated in your own argument.
Can Turnitin flag text I wrote myself as AI-generated?
Yes, in some cases. The AI indicator is a pattern-based classifier, so formulaic or heavily edited prose can resemble model output [3]. Below Turnitin's 20% confidence threshold the score is shown as an asterisk rather than an exact figure.
Will a pre-submission check add my paper to the student database?
Only if the service stores submissions. Non-repository checks return the report without filing your paper in the student paper repository, which keeps you from matching against yourself later [4]. Confirm this before uploading.
Do I need to check both the similarity and the AI report?
Yes — they answer different questions. Similarity shows overlap with existing sources [1], while the AI report estimates machine-generated text [3]. A draft can pass one and fail the other.