Falsely Accused of Using AI? What to Do, Step by Step

Don't reply right away or in anger. Read the message twice. Note what is being raised, who raised it, and any deadline for a reply or a meeting.

Then work out whether the message raises a concern or reports a decision. Procedures differ by institution. Under the University of Minnesota Crookston's academic integrity procedure, for example, the instructor first emails the student a notice of the allegation. The instructor then presents the evidence in a meeting, and a decision follows. If the student refuses to meet, the instructor determines an outcome without that meeting. If you don't respond, you may lose your chance to explain.

Before you reply, read the rules your case will be judged by:

• Syllabus and assignment instructions: the course's AI rules, which can differ by assignment.
• Academic integrity procedure: the steps, the deadlines, who decides, and how to appeal.
• AI-use policy: any separate rules from your institution, department, or program.
• Student handbook: your rights in the process, such as bringing someone to a meeting.

Some policies allow certain AI uses, such as brainstorming or a grammar check, and some require you to disclose them. If you used an AI tool for any part of the work, find out exactly what your policy allowed. Our guide to AI and academic integrity explains where help with your writing ends and generated content begins.

You can't answer a concern you can't see. Reply in writing, stay polite, and ask for the details behind it. A written exchange also gives you a dated record.

Example: "Thank you for letting me know. I wrote this essay myself and want to respond properly. Could you tell me which tool and result the concern is based on, which passages were flagged, whether there is other evidence, and which policy and next step apply?"

Good evidence shows how your text developed over time. Hope International University's Darling Library publishes a faculty guide on students who deny using AI. It tells instructors that AI detectors "don't prove anything, they just analyze probability." It also suggests that instructors return to having students turn in rough drafts, notes, and journal articles with highlights and notes. Collect the same kind of material.

Choose quality over volume. An outline, two or three dated drafts, and the notes behind the flagged passages make your case more clearly than a pile of screenshots.

MIT Sloan's Teaching & Learning Technologies group suggests that instructors ask students for a "process statement" that briefly explains how they completed the assignment and whether and how they used AI tools. Your response can work as one. Keep it factual. Don't accuse your instructor of anything, and don't argue about detectors in general. A structure you can adapt:

• Your statement: "I wrote [title] myself between [date] and [date]."
• Any AI use: "I did not use AI tools for this assignment." Or: "I used [tool] only to [purpose], which [the syllabus permits]."
• Evidence: "I have attached the document's version history, my outline, two earlier drafts, and the notes behind the flagged sections."
• Request: "I would welcome a meeting to walk you through the essay and my sources. I am available on [days and times]."

Send it before any deadline, and keep a copy of everything you send.

The Hope International University guide tells instructors, "We must still talk with the student." It suggests asking whether the submitted work is the student's own, the work of an AI, or their work supplemented by AI. Expect questions about your content, and be ready to explain:

• Your argument: your main claim, and why you structured the piece as you did.
• Your sources: where you found them and what each one contributes.
• Your choices: why you used a particular term or phrasing in the flagged passages.
• Your process: what changed between drafts, and why.

Reread the essay and your key sources beforehand. If you're nervous, explain the flagged passages out loud to a friend first.

Detectors measure statistical patterns, such as how predictable the wording is. They do not identify who wrote a text. Our guide to how AI detectors work covers why false positives happen, and signs of AI writing describes the patterns involved. The Hope International University guide names situations that can lead to a wrong flag: translation software, accepted suggestions from grammar and word-processing tools that now have AI built in, very mechanical writing such as heavily formatted lab reports, and traditional five- or seven-paragraph essays. If you used translation or grammar tools, check whether your policy counts that as AI use, and say so in your response.

A 2023 study in the International Journal for Educational Integrity tested 14 detection tools. The authors note that the tools give bare statements such as "This document was likely written by AI," with no way to verify them and no evidence. They concluded that the tools' reports "cannot be used as the only basis for reporting students for cheating." OpenAI launched its own AI-text classifier in January 2023 with the warning that it "should not be used as a primary decision-making tool." As of July 20, 2023, the classifier was no longer available "due to its low rate of accuracy."

Running your essay through more detectors until one says "human" proves nothing either. In the 2023 study, the probability of false positives ranged from 0% to 50%, depending on the tool. The tools also leaned toward labeling text as human-written. The authors estimate that about 20% of the AI-generated texts would likely be misattributed to humans. A low result from a second tool is no more proof that you wrote the text than a high result from the first is proof of AI use.

Depending on your institution, an ombuds office, a student advocacy office, or your academic advisor can explain the process and your options. A writing center tutor who saw your drafts may be able to confirm what they saw and when.

If the decision goes against you and you believe it's wrong, appeal on time through the route your policy sets out. Under the Crookston procedure, a student who disagrees after the meeting with the instructor has five business days to appeal to the academic administrator who oversees the course. The appeal can address only whether the student is responsible, not the outcome the instructor assigns. Other institutions set different deadlines and grounds, so read your own procedure as soon as a decision arrives.

Also check whether you may bring someone with you. Under the same procedure, the student and the instructor may each have an advisor present when the academic administrator interviews them, with 48 hours' advance notice. The advisor may observe and consult but may not speak for them.

The Certentis AI Detector runs the Certentis AI analysis on your text. It shows Human, Mixed, and AI probabilities for the whole document and signals for individual sentences. The sentence signals show which passages this analysis reacts to, so you know which drafts and notes to have ready before a meeting. Your instructor's tool may react to other passages, so start with the ones named in the concern.