If you have been accused of using AI at a University of California campus, the guidance issued by UC teaching centers in 2023 is directly relevant to your case. Those documents did not create new student rights, but they did tell UC faculty something important: an AI detector score is not, on its own, evidence of misconduct. That framing matters when the accusation against you rests on a Turnitin flag.
What UC guidance actually says
In 2023, several UC teaching and learning centers, including the UC Berkeley Center for Teaching and Learning and the UCLA Center for the Advancement of Teaching, published faculty-facing guidance on generative AI in coursework. The consistent message across those documents was that AI detection tools are unreliable enough that a detector score should not be treated as proof of a violation. Faculty were advised to design assignments that made AI use less tempting or more visible, and to talk with students rather than reflexively file academic misconduct reports based on a detector output.
This was not a systemwide ban on detectors. UC did not prohibit faculty from running work through Turnitin or GPTZero. What the guidance did was signal, in writing, that campus teaching authorities considered detectors unreliable as standalone evidence. That is a useful institutional acknowledgment, and it is one you can cite.
Why UC teaching centers took this position
The UC guidance did not emerge in a vacuum. It tracked the peer-reviewed evidence that had accumulated by mid-2023. The Weber-Wulff et al. study in the International Journal of Educational Integrity tested fourteen AI detectors and concluded that none reached a level of reliability appropriate for high-stakes decisions. The Stanford study by Liang and colleagues, published in Patterns, found that GPT detectors flagged non-native English writing at strikingly higher rates than native English writing.
UC teaching centers were paying attention to that literature. Their guidance reflects it: cautious framing, an emphasis on process evidence and pedagogy, and a clear reluctance to endorse detector-driven enforcement.
How UC conduct policy treats evidence
UC campuses generally apply a preponderance of the evidence standard in academic misconduct cases: the panel must find it more likely than not that a violation occurred. That is a lower bar than criminal proof, but it is still a bar. A detector score alone does not clear it when the underlying tool has documented false positive rates, and when the student produces contemporaneous evidence of authorship.
Every UC campus also gives accused students a set of procedural rights: notice of the specific allegation, access to the evidence being used against them, the opportunity to respond in writing and in person, and the right to appeal. Those rights are laid out in campus procedures and echoed in the procedural rights FAQ. Requesting the underlying detector report, the name of the tool, and any human review notes is not adversarial; it is a normal exercise of those rights.
How to cite UC guidance in your response
You are not arguing that the guidance forbids the accusation. You are arguing that your own campus's teaching authority has publicly acknowledged the tool's limits, and that a fair hearing has to weigh that acknowledgment against the detector output. A concise citation is more effective than a long quotation. Your response should:
- Name the specific UC teaching center document you are citing and the year it was published
- Quote one or two sentences that directly address detector reliability
- Pair the citation with the peer-reviewed research (Weber-Wulff 2023, Liang 2023) that supports the same conclusion
- Connect the point to your case: the accusation rests on a detector score, and both your campus guidance and independent research treat that score as insufficient on its own
- Present your process evidence (drafts, version history, notes, research records) as the affirmative account of how the work was produced
Pairing guidance with process evidence
Guidance and research explain why a detector is unreliable. Process evidence explains why your specific paper is yours. The two arguments reinforce each other. Google Docs version history, browser research trail, library checkout records, annotated PDFs, handwritten notes, and dated drafts all belong in the package. Detectors cannot see any of that, which is one of the reasons UC teaching centers were skeptical of them in the first place.
What the 2023 guidance does not do
The UC teaching guidance is not a get-out clause. It does not prohibit faculty from making AI-related allegations, and it does not automatically invalidate a Turnitin flag. A hearing panel is free to weigh a detector score alongside other evidence, and the ultimate finding will turn on the record as a whole.
What the guidance does is shift the framing. When a detector score is the primary or sole basis for an accusation, your campus's own teaching authority has said that is not enough. That is a reasonable thing to point out in a response letter, and it is a reasonable thing to raise on appeal if the hearing outcome ignored it. If you have already been found responsible and are preparing to challenge the finding, the appeal package walks through the grounds most likely to succeed at UC campuses.
Using this in your defense
A UC accusation built on a detector score is not automatically a winning case for either side. It is a case that turns on evidence, procedure, and how well the record is developed. Campus teaching guidance is one input. Peer-reviewed research is another. Your own process evidence is the third, and often the most decisive. If you are drafting a response, NotBot generates a personalized defense package that pulls the relevant UC guidance, the detector research, and your writing process into a coherent letter, evidence guide, and hearing brief, ready in about a minute.
If your case involves potential suspension, expulsion, or immigration consequences, the UC guidance and the research supporting it belong in your response, but consulting an education law attorney before the hearing is worth serious consideration.
Build your UC defense package
A personalized response that cites UC guidance, the research, and your writing process, ready in minutes.
Get your defense package$49 one-time · Generated in 60 seconds