UNG AI Accusation Dropped After Handwritten Notes and Browser History

A student accused of AI use produces a stack of handwritten notes and a browser history full of research tabs. The accusation gets dropped. That sequence, whether it plays out at the University of North Georgia or anywhere else, is not luck. It is what process evidence does when a detector score is the only thing on the other side of the table.

The pattern: detector flag, process evidence, withdrawal

Across public and reported cases at regional universities like the University of North Georgia, the sequence looks similar. An assignment is submitted through Turnitin or a comparable platform. The instructor sees an AI indicator score above whatever threshold they treat as significant. An email or Student Conduct referral follows. The student responds with material the detector cannot see: a spiral notebook, printed articles with margin notes, a Google Docs version history, a browser history that shows hours of research on the topic. The instructor withdraws the allegation, sometimes before it ever reaches a formal hearing.

This pattern is not unique to UNG. It repeats at institutions across the country and mirrors what happened in the widely reported UC Davis handwritten draft case. The common thread is that a probabilistic score, standing alone, does not survive contact with concrete evidence of a writing process.

Why a detector score is not enough on its own

AI detectors output a probability, not a fact. Peer-reviewed research by Weber-Wulff and colleagues (2023), published in the International Journal of Educational Integrity, tested fourteen detection tools and concluded that none of them met a reliability standard suitable for institutional decisions. A 2023 Stanford paper by Liang et al. in Patterns found that detectors flagged non-native English writing at strikingly higher rates than writing by native speakers.

A UNG instructor operating in good faith may still treat a high detector score as suggestive. What the research shows, and what the retraction pattern reflects in practice, is that suggestive is not the same as sufficient. When a student produces contemporaneous evidence of the drafting process, the detector score loses whatever weight it appeared to carry.

What handwritten notes and browser history actually prove

The strength of the evidence in these cases is not the notebook itself. It is what the notebook demonstrates about time, sequence, and intent. Handwritten notes with dates, page numbers from library sources, and marginal questions show that the student engaged with the material over hours or days. Browser history with search queries, article opens, and citation lookups shows the same thing on the digital side.

This kind of evidence is difficult to fabricate after the fact and easy to verify. It is also directly responsive to the theory of the accusation. If the claim is that a student pasted a prompt into a language model and submitted the output, minutes of research and hand-drafted outlines are inconsistent with that story.

Evidence that tends to close these cases

• Handwritten notes or outlines with dated entries and source references
• Google Docs or Word version history showing incremental drafting over time
• Browser history covering research, article opens, and citation searches during the drafting window
• Library records or database access logs for cited sources
• Emails, LMS logs, or office hours records showing engagement with the assignment

What UNG policy requires before a finding

The University of North Georgia's Student Code of Conduct places academic misconduct within a broader disciplinary framework governed by University System of Georgia policy. Findings require a preponderance of the evidence, meaning it is more likely than not that a violation occurred. That standard is not met by a detector percentage alone when the student produces credible process evidence to the contrary.

The code also gives students the right to see the evidence against them, to respond in writing, and to appeal a finding on procedural or evidentiary grounds. If you have already received a finding rather than an initial accusation, the appeal package covers the grounds that matter after the fact.

What to do if this is you at UNG or a similar institution

1. Preserve everything now. Do not edit or delete drafts, notes, or browser history.
2. Export your Google Docs version history and download your browser history for the drafting window.
3. Photograph any handwritten notes with visible dates and page numbers.
4. Read the Student Code of Conduct and your syllabus. Confirm what the policy actually prohibits and what standard of evidence applies.
5. Request the detector name, score, and any human review notes in writing.
6. Write a response that pairs the process evidence with the research on detector reliability.

Reviewing the procedural rights FAQ before you write anything is worth the ten minutes. If you are preparing your written response, NotBot generates a personalized defense package that documents your writing process, addresses the specific detector that flagged you, and cites the peer-reviewed research on detection reliability.

If your case carries potential suspension, expulsion, or visa consequences, the research and process evidence still form the core of your response, but consulting an education law attorney before your hearing is advisable.