UCSB Professor Retracts AI Cheating Accusation After Student Response

A UC Santa Barbara student accused of using AI on a course assignment had the allegation withdrawn after producing a complete record of the writing process. The case follows a pattern that has now appeared at multiple UC campuses: a detector flag, a meeting with the instructor, and a retraction once the student presents drafts, version history, and notes that the detector never saw.

What happened

The accusation began the way most do at UC campuses: an assignment was submitted through Turnitin, an AI similarity indicator returned a high percentage, and the instructor opened an academic integrity conversation. The student denied AI use and asked what evidence the instructor was relying on. The answer, in substance, was the detector score and the instructor's reading of the prose.

The student responded by producing the Google Docs version history for the assignment, time-stamped research notes, and a screen recording made during one drafting session for a separate class that demonstrated their normal writing pattern. After reviewing that material, the instructor withdrew the allegation before it was referred to the campus conduct office. No formal finding was entered. The outcome mirrors what has been reported in other UCSB Turnitin false positive cases: the detector flag does not survive contact with process evidence.

Why the retraction happened

Detector scores describe statistical properties of finished text. They cannot describe how the text was produced. Once a student introduces direct evidence of the production process, the detector's signal is no longer the only thing in the record, and an instructor weighing both has to weigh them against each other.

In this case the version history showed incremental edits over several days, with sentences appearing, being revised, being deleted, and being rewritten. AI-pasted text typically arrives in large blocks at single timestamps. The research notes referenced specific page numbers from library sources. The screen recording, made for unrelated reasons, showed the same habits of drafting and self-editing that produced the flagged assignment.

The instructor was not required to retract. They chose to because the evidence shifted the probability assessment. That is the practical mechanism behind almost every documented retraction at UC campuses.

What UCSB policy actually requires

UC Santa Barbara's Student Conduct Code requires a preponderance of the evidence standard for a finding of academic misconduct. A detector score, by itself, is not a finding. It is a signal that triggers further inquiry, and UC guidance has been consistent that detector output should not be treated as proof.

Instructors at UCSB retain discretion over whether to refer a case to the Office of Student Conduct or resolve it informally. The earlier in the process a student introduces credible counter-evidence, the more often the case ends at the instructor level. Procedural rights at the formal stage are covered in our procedural rights FAQ.

Evidence that led to the withdrawal

The materials the student produced fit a pattern that recurs across UC retractions:

• Version history from Google Docs or Microsoft Word showing incremental construction of the text over time
• Research notes with citations that match the sources used in the final assignment
• Independent recordings or artifacts from unrelated work that show the same writing style and habits
• A direct, factual written statement denying AI use and explaining how the assignment was produced

None of these individually disproves AI use. Together they describe a writing process that an AI workflow does not produce. That is what changes the instructor's calculation.

If this is you at UCSB

The window between an instructor accusation and a formal referral is the most important phase of the case. What you produce in that window often determines whether a referral happens at all.

1. Ask the instructor in writing for the specific detector used, the score reported, and the passages they consider problematic.
2. Preserve the full version history of the assignment and all related research materials before opening or editing anything.
3. Write a calm, factual response that addresses the detector's known limitations and presents your process evidence.
4. Request a meeting only after your written response is on the record.
5. If the case moves to the Office of Student Conduct, review the formal procedure carefully and prepare for a hearing.

If you are preparing a written response, NotBot generates a personalized defense package that addresses the specific detector used at UCSB, your writing process, and the procedural standard the Student Conduct Code requires. If the case has already progressed past the instructor stage and a finding has been entered, the appeal package covers what to argue at the appellate stage.

Retractions like this one are not the norm, but they are not rare either. They depend almost entirely on whether the student can produce a process record before the case escalates.