UC Davis and AI Detection False Accusations: The Pattern

Reports of students being accused of AI use based on detector output, sometimes facing severe sanctions before successfully appealing, have surfaced at universities across the United States. The pattern matters more than any single case: a detector flag, treated as proof, can drive a sanction that does not survive a closer look at the evidence.

The documented pattern across universities

Since the public release of ChatGPT in late 2022, reporters at outlets including the Washington Post, USA Today, Bloomberg, and the Chronicle of Higher Education have documented a recurring sequence at universities including the University of California system, the University of Minnesota, the University of North Georgia, and others. The sequence usually looks like this:

1. A faculty member submits a paper to an AI detector, most often Turnitin or GPTZero.
2. The tool returns a percentage score the instructor interprets as evidence of AI generation.
3. The student receives an accusation, sometimes with a proposed sanction ranging from a zero on the assignment to suspension or expulsion.
4. The student produces process evidence: drafts, version history, notes, library records.
5. On appeal or further review, the sanction is reduced or vacated.

The University of California academic integrity framework gives each campus, including UC Davis, its own student conduct office that handles these cases under the systemwide Policies Applying to Campus Activities, Organizations, and Students (the so-called "PACAOS"). Procedural rights, including the right to review the evidence against you and to appeal, are spelled out there.

Why history and humanities essays get flagged

Humanities essays are over-represented in reported false positive cases for reasons rooted in how detectors work. AI detection tools rely primarily on perplexity (how predictable the word sequences are) and burstiness (variation in sentence structure). Several features common to history essays push text toward the AI-flagged end of those metrics:

• Formal academic register, with consistent tone and limited slang
• Heavy use of dates, names, and standard historical terminology that is highly predictable in context
• Topic sentences and structured paragraphs taught in first-year writing courses
• Quotation and paraphrase of primary sources, which the student did not generate and which can read as stylistically uniform

The 2023 study by Weber-Wulff and colleagues, published in the International Journal of Educational Integrity, tested fourteen detection tools and concluded that none performed reliably enough to support standalone disciplinary use. A separate Stanford study (Liang et al., 2023, Patterns) found false positive rates above 50 percent for some categories of human-written essays. Our review of the research on detector accuracy covers these findings in detail.

The evidence that tends to reverse these cases

In the cases reported by mainstream outlets where accusations have been withdrawn or sanctions overturned on appeal, students typically produced concrete process evidence rather than rebutting the detector mathematically. The most useful artifacts:

If this is you at UC Davis or another UC campus

UC Davis handles academic integrity allegations through its Office of Student Support and Judicial Affairs (OSSJA). The process typically begins with a notification, an opportunity to respond, and (if the case is not resolved) a hearing. Students have the right to review the evidence, to bring an advisor, and to appeal an adverse outcome under the campus implementation of the UC student conduct policy.

Practical steps to take as soon as you receive an allegation:

1. Request, in writing, the specific detector used, the score reported, and any human review notes.
2. Preserve all drafts, version history, notes, and research records immediately. Do not delete or edit anything.
3. Read the UC Davis Code of Academic Conduct and your syllabus carefully for what is actually prohibited. AI tool use is treated differently across courses.
4. Ask whether the instructor consulted with OSSJA before issuing the allegation, and whether the standard of evidence required by policy was met.
5. If the proposed sanction is suspension or expulsion, or if you are an international student whose visa status depends on enrollment, consult an education law attorney before your hearing.

Our procedural rights FAQ covers what to request and when. If you are drafting a written response now, NotBot generates a personalized defense package that addresses your specific detector, your writing process, and the procedural standard your institution applies, with a response letter in three tones, an evidence checklist, and a hearing brief. If you have already received an adverse decision, the appeal package is structured around the grounds UC campuses recognize on review.