Ohio State AI Accusation Withdrawn After Handwritten Notes

A student accused of using AI on a class assignment produced handwritten notes from the drafting period. The professor withdrew the accusation. The pattern is familiar: a detector score triggers the allegation, and process evidence ends it. If you are facing something similar at Ohio State or a comparable institution, the mechanics of what turned this case around are worth understanding in detail.

The pattern: detector flag, handwritten evidence, retraction

The sequence in cases like this is consistent. A written assignment is submitted through Turnitin or a similar platform. The AI indicator returns a high percentage. The instructor emails the student and opens an academic misconduct inquiry. The student produces something the detector cannot see: a physical notebook of pre-writing, an outline drafted longhand, marginalia in a printed source. The instructor reviews the artifacts, compares them to the submitted essay, and closes the case without a formal finding.

No detector output survives contact with genuine process evidence. That is not because handwriting is magical; it is because AI detectors have no way of knowing what a student did before they typed the first sentence. A dated notebook page with the thesis crossed out and rewritten is exactly the kind of artifact a probabilistic classifier cannot produce or fake retroactively. Related patterns are documented in our coverage of a UC Davis handwritten draft case and the broader UCSB retraction reporting.

Why handwritten notes are decisive

A detector score is a statistical guess about a finished text. Handwritten notes are a physical record of the thinking behind that text. When they match, the argument that an AI wrote the paper collapses under its own weight: any model that could reverse-engineer a student's notebook, in that handwriting, on that paper, with those specific false starts, does not currently exist.

Instructors know this. Even ones inclined to trust the detector will usually reverse course when the alternative is arguing that the student manufactured a fake notebook to cover their tracks. The effort required to fabricate that evidence exceeds the effort of writing the essay honestly.

Why the detector flagged the work in the first place

AI detectors classify text using signals like perplexity (how predictable word choices are) and burstiness (how much sentence structure varies). Careful academic prose, formal argumentation, and writing that avoids idiom all tend to produce the same statistical fingerprint the detectors associate with AI output. In the peer-reviewed 2023 evaluation by Weber-Wulff and colleagues in the International Journal of Educational Integrity, no tested detector was found reliable enough for high-stakes decisions.

The Stanford study by Liang and colleagues, published in Patterns in 2023, found that non-native English writing was misclassified as AI-generated at strikingly higher rates than native writing. The mechanism is not fraud on the part of the writer; it is the way the detector reads clarity and simplification as a statistical anomaly. Students who write carefully, revise heavily, or write in a second language are structurally more likely to be flagged.

What Ohio State policy requires

Ohio State's Code of Student Conduct governs academic misconduct allegations through the Committee on Academic Misconduct (COAM). Instructors who suspect misconduct are directed to report the matter to COAM rather than resolve it unilaterally, and findings must be supported by evidence sufficient under the standard set in the code. A detector percentage, on its own, is not that evidence. It is a signal that prompts an inquiry, not a finding.

Ohio State students accused of academic misconduct have specific procedural rights: written notice of the allegation, access to the evidence being used against them, and a hearing before COAM if the case is not resolved informally. Our procedural rights FAQ covers what to request in writing before your hearing.

Evidence that tends to close these cases

The retractions that happen quickly tend to involve a specific mix of artifacts, not any single item. What tends to work:

• Handwritten notes or outlines from the pre-writing stage, photographed with metadata intact
• Google Docs or Word version history showing the essay being built over hours or days, with revisions and deletions visible
• Browser history showing research on the specific sources cited in the essay, in the days before submission
• Library database access logs or citation manager entries that predate the draft
• Messages or emails to classmates, TAs, or the instructor discussing the assignment while it was in progress

Each item on its own can be explained away. Together, they describe a real writing process that no chatbot participated in. For a full step-by-step on preserving these artifacts before they expire from retention windows, see our guide on how to gather evidence after an AI accusation.

If this is you at Ohio State

Do not respond emotionally to the first email from your instructor. Ask, in writing, three specific things: the detector that produced the score, the exact percentage or classification returned, and whether any human review of the writing occurred before the allegation was made. Then start assembling process evidence. Do not open the flagged document to look at it; every save can overwrite version history that supports you.

If a formal COAM process is opened rather than an informal resolution, the standard for a finding is higher than the instructor's initial suspicion. Bring the peer-reviewed research on detector accuracy, your process evidence, and a clear written account of how you wrote the essay. If a finding is issued, the appeal package covers the grounds that matter at that stage.

If you are preparing a written response, NotBot generates a personalized defense package that names the detector used, cites the research on its limitations, and documents your writing process in the format Ohio State's process expects, ready in about a minute.

If your case involves potential suspension, expulsion, or visa consequences, consult an education law attorney before your hearing. Research and process evidence support your argument, but they do not replace legal counsel when the stakes are that high.