Vanderbilt AI Case Dismissed After Drafts and Browser History

A Vanderbilt University student accused of using AI on a written assignment saw the case dismissed after producing saved drafts and browser history that documented the writing process from research to final submission. The pattern is now familiar across campuses: a detector flag triggers an accusation, and process evidence closes it. What makes the Vanderbilt outcome notable is that it happened at an institution that had already disabled Turnitin's AI detection feature in August 2023 over reliability concerns.

What happened

The accusation followed a common script. An instructor ran the student's paper through an AI detection tool, saw a score above the threshold the instructor treated as significant, and referred the matter for review. The student responded by producing three things: saved Google Docs revision history showing the document evolving over multiple sessions, browser history confirming research on the topic in the days before the paper was drafted, and a short written narrative describing how the paper came together.

Reviewers accepted the evidence and closed the case without a finding. No formal hearing was required.

Why saved drafts and browser history were decisive

AI detectors produce a score. Process evidence produces a timeline. Reviewers weighing a probabilistic score against a documented sequence of drafts, edits, and searches tend to follow the timeline, because it answers a question the detector cannot: how did this specific document come to exist?

Two categories of evidence carry the most weight in these cases:

• Version history from a cloud editor. Google Docs, Microsoft Word (with AutoSave), and Notion all store granular revision data. A document that grew over multiple sessions, with visible edits, deletions, and reorganizations, does not match the profile of pasted AI output.
• Browser history from the drafting period. Searches for the topic, database access to library resources, and visits to source materials cited in the paper establish that the research actually happened and that the writer engaged with the sources.

Neither artifact is difficult to produce if you know to preserve it. Both become inaccessible if you delete history, edit the document further after the accusation, or wait past your browser's default retention window.

Why the detector flagged human writing in the first place

AI detectors measure statistical properties of text: how predictable word sequences are (perplexity) and how much sentence length varies (burstiness). Writing that is careful, structured, and free of idiomatic clutter can score low on perplexity because it uses common, high-probability phrasing. That description fits most competent academic writing.

Peer-reviewed research has documented significant false positive rates across major detectors, including studies by Weber-Wulff and colleagues (2023) in the International Journal of Educational Integrity and Liang and colleagues (2023) in Patterns. The 2023 accuracy research on AI detectors found that no tested tool met the reliability threshold researchers considered appropriate for institutional decision-making.

What to preserve, and when

The Vanderbilt outcome depended on evidence that already existed at the moment the accusation arrived. If you have been accused, or think an accusation is likely, act on the following before anything else:

1. Stop editing the flagged document. Every save can overwrite version history and alter metadata.
2. Export version history immediately. In Google Docs, use File → Version history → See version history, then screenshot and download each version. In Word, check File → Info → Version History or your OneDrive backups.
3. Export browser history for the drafting window. Chrome, Safari, and Firefox all allow export. Do it before your default retention window closes.
4. Preserve secondary evidence. Library database access logs, citation manager entries (Zotero, Mendeley), messages with classmates or instructors about the assignment, and photos of handwritten notes with timestamps.
5. Email a copy to yourself. A timestamped email establishes a third-party record of when the evidence existed.

For the full 48-hour preservation sequence, see our guide on gathering evidence after an AI detection accusation.

What Vanderbilt policy requires

Vanderbilt's Honor Code governs academic integrity cases through the Honor Council, which handles allegations under a procedure that requires evidence beyond a bare detector score. The August 2023 decision by the Center for Teaching to disable Turnitin's AI detector was grounded explicitly in the tool's unreliability. That institutional position is directly citable in a response letter: if the university itself concluded the detector was not reliable enough for routine use, an accusation resting on a comparable third-party detector faces the same reliability problem.

A response that pairs institutional context with process evidence gives reviewers a clean path to closing the case without a formal finding.

If this is you at Vanderbilt or elsewhere

The evidence that closed the Vanderbilt case is available to most students who write in a cloud editor and research online. The barrier is usually not access to evidence but knowing to preserve it quickly and present it in a form reviewers can use. Your written response should identify the specific detector, cite the research on false positives, walk the reviewer through your process timeline, and attach the artifacts in an indexed order.

Understanding your procedural rights before a hearing matters as much as the evidence itself, especially for questions like what information you can request, what standard of proof applies, and whether a human reviewer examined the flagged sections before the referral. If you are preparing a written response, NotBot generates a personalized defense package that names the detector, cites the research, and structures your process evidence into a document reviewers will actually read.

If the proposed sanction is suspension, expulsion, or has visa consequences, consult an education law attorney before your hearing.