Skip to content
Global 100 Forum

What should a student do if an AI detector flags their work?

Staff answerWritten by Maya Lindqvist · 1 reply · updated

Short answer

Stay calm, gather evidence of how you wrote the work, and ask for the specific basis of the concern in writing. A detector score alone is weak evidence: Turnitin's own guidance says its AI indicator should not be the sole basis for action, and Vanderbilt disabled it in 2023. Your drafts and document history persuade far more than any score.

CiteGlobal 100 Forum, "What should a student do if an AI detector flags their work?", https://forum.global100.org/q/what-should-a-student-do-if-an-ai-detector-flags-their-work/, accessed 2026-10-11.
Written by Maya Lindqvist · Edits the AI text detection and academic integrity sections ·
  1. Edits the AI text detection and academic integrity sections ·

    Turnitin's statement is in its AI writing detection FAQ; Vanderbilt's reasoning is in its August 2023 notice.

    Collect your process evidence first

    What counts as evidence, and how much weight it carries

    Evidence What it shows Weight
    Detector score alone A statistical guess about the text Low: independent tests find detectors unreliable as evidence
    Version history with timestamps The document grew over time in your hands High
    Drafts, outlines and notes A writing process, not a paste High
    A conversation about the argument and sources You understand what you wrote High
    • Version history. Google Docs, Microsoft Word with OneDrive, and most writing tools keep a timestamped history of edits. Export or screenshot it before anything else.
    • Drafts and notes. Outlines, early drafts, research notes, highlighted sources and saved searches all show a writing process.
    • Sources. A list of what you read and when, including library records or browser history if you are comfortable sharing it.
    • Anything you used legitimately. If you used a grammar checker or a translation tool and that was allowed, say so plainly and show where.

    Ask the right questions

    Ask which tool was used, what score it gave, and which passages it highlighted. Ask what your institution's policy says about AI detection evidence. Some institutions do not allow detector scores to be used on their own, and some have switched detection off entirely. Vanderbilt disabled Turnitin's AI detector in 2023, citing the risk of false positives across tens of thousands of papers a year and the lack of transparency about how scores are produced.

    Point to the research, calmly

    It is reasonable to share independent evidence about detector reliability. The 2023 test of 14 tools by Weber-Wulff and colleagues concluded that none was accurate or reliable enough to serve as evidence. If you write in a second language, the Stanford study on bias against non-native writers is directly relevant, and the forum's thread on that bias summarises what it found.

    Offer to demonstrate

    Offering to talk through your argument, explain your sources or write a short piece under supervision is often the fastest way to resolve a case. It shows understanding of the material, which is what the assessment is meant to measure.

    0
    ReplyLink

1 more reply

Most helpful first
  1. Maya LindqvistStaffEdits the AI text detection and academic integrity sections ·

    For instructors reading this thread: the fair version of the same advice is to open with a conversation, not an allegation. Ask the student to walk you through how they wrote the piece and to show their history. Most genuine cases become clear quickly, and most false flags resolve just as quickly, without anyone having to be accused of anything.

Write a reply

Plain text or simple Markdown. Links are nofollow. Your email is never shown.