If a detector flags work you wrote, keep the submitted file and genuine drafts intact, check the assignment and school policies, and ask to see the report and the specific passages at issue. Respond calmly through the institution’s process. A detector result is a reason to review the work, not conclusive proof of who wrote it.
What should you do first?
- Preserve your existing records. Keep a copy of the submitted file, genuine drafts, notes, outline, source list and any version history already available. Preserve original files and timestamps. Do not create or backdate evidence, or edit the submitted work to try to change a score.
- Read the relevant rules. Check the assignment instructions, course AI-use statement and academic-integrity policy, along with any notice you received. Rules about permitted AI use, disclosure and review procedures can differ by course and institution.
- Ask what prompted the concern. If you cannot see the report, ask the instructor or appropriate office to share it and identify the passages or conduct under review. Keep your request factual and professional.
- Prepare to explain how you wrote the work. Be ready to discuss your argument, how it developed, the sources you used or rejected, and the decisions behind your conclusions. Offer relevant existing records through the approved process.
- Follow the stated procedure. Ask whether you are having an informal conversation or responding to a formal allegation, what evidence is being considered, how and when to respond, and what support is available.
There is no single piece of evidence that guarantees a particular outcome. A clear account of your process and authentic records can help an instructor assess the work alongside the detector report.
As an Amazon Associate I earn from qualifying purchases.
What does a detector flag establish?
An AI-writing detector estimates whether text has patterns associated with AI-generated writing; it does not directly observe who typed the words. A score or highlighted passage is the tool’s output, not a finding that a student committed misconduct. The Australian regulator TEQSA describes detectors as examining linguistic or structural patterns and cautions that results can be less reliable for short, edited or mixed human-and-AI text. A score should not be read as the probability that a particular person used AI.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Different tools use different methods, and a result depends on the detector, the text and the report date. A conversation prompted by a flag is not necessarily a formal allegation. Institutions also set their own rules for how detector output may be considered. For example, the University of Melbourne’s guidance says an AI detection report alone is insufficient evidence for an allegation under its process; that is not a universal rule for every school.
#1 Best Overall
How should you understand a Turnitin AI report?
Turnitin’s AI-writing report applies to “qualifying text,” which its current support guidance defines as prose sentences in long-form writing. The guidance lists a minimum of 300 prose words, a maximum of 30,000 words, English, Spanish, Japanese and Arabic, and .docx, .pdf, .txt and .rtf files. These are Turnitin product requirements, not general standards for other detectors.
Turnitin says its report may misidentify human-written, AI-generated and AI-paraphrased text, and that it should not be the sole basis for adverse action against a student. Its guide says scores from 1% to 19% display as an asterisk rather than a numerical percentage, without highlights, to reduce possible misinterpretation. Older reports may look different. Turnitin also documents detector updates in February and May 2026; it says those updates did not retroactively change earlier reports unless the submission was sent again. Check the report date and current product guidance before drawing conclusions from an old screenshot or score convention.
Turnitin has published its own performance figures, but they do not settle an individual case or describe all detectors. In a 2023 statement, its chief product officer said the company used 800,000 pre-ChatGPT academic-writing samples for further testing. The same source reported a document false-positive rate below 1% for documents with 20% or more AI writing, and a sentence-level false-positive rate of around 4%; it also said 54% of false-positive sentences were adjacent to actual AI writing. These are vendor-reported figures for different measures and specified conditions. An August 2026 Turnitin explainer repeated a claim of less than 1% false-positive risk for documents with over 20% likely AI-generated content. None of these figures is a guarantee about a particular report, and none should be applied to another company’s detector. TEQSA specifically cautions that even a low false-positive rate does not mean a high score proves that a particular paper was AI-written.
How can you explain your writing process?
Focus on the choices and work you actually did, rather than trying to guess what answer an instructor wants. Washington University’s student guidance suggests being prepared to discuss when you worked, notes and earlier versions, resources consulted, obstacles, and your understanding of the submitted work. You might explain how you arrived at your thesis, why you used particular sources, what changed between drafts, and how you reached the conclusion.
Rank #3
Share only relevant, authentic materials, and use the school’s approved channel. TEQSA describes verifiable version history as one way to show how an assessment was compiled. Drafts and history can provide context, but they should be considered with the content and the institution’s process rather than treated as automatic proof.
How do you protect your work while seeking help?
Before uploading assessed work to a public detector or other third-party service, check both your institution’s rules and the service’s terms about storage and reuse. TEQSA warns that unlicensed third-party detector services may raise IT, privacy, intellectual-property or copyright issues. The University of Melbourne also warns students about intellectual-property and possible data-use concerns with free online tools. A second detector score does not establish authorship, and a “humanizer” or bypass service is not a sound way to respond to a flag.
If you need help preparing for a meeting, ask whether your institution has a student advocacy office, academic skills service or writing support service. Use the actual service and procedure available at your school rather than assuming another institution’s guidance applies.
Recommended Free Tools
Why is “watermark detector” ambiguous?
In everyday discussion, people often use “watermark detector” to mean an AI-writing detector. This guidance addresses that common situation: a tool flags text as likely AI-written. Technical text watermarking is a different subject; it concerns detectable signals embedded in generated text. The evidence discussed here does not establish how any specific watermarking scheme works or how reliable one is.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
