At UC Irvine the clearest written rule for dissertations comes from a department. On November 9, 2026 we found no campus-wide rule on AI in theses or dissertations. The Department of Sociology published guidelines in May 2026: advisor and advisee agree on how AI may be used, some uses such as improving clarity or grammar are generally acceptable, and any use is disclosed in a short statement. Students in other departments follow their own program's rules or ask their advisor.
If your AI statement clearly doesn't cover what you describe elsewhere in your thesis, Esmond points it out.
We looked for a Graduate Division or campus-wide rule on AI in theses and dissertations and found none on November 9, 2026. It does not prove that nothing exists. What we did find is a department text with clear rules for doctoral work, and a central page on data security that applies to every AI tool you use with UCI information.
The rules below belong to the Department of Sociology and apply to its graduate students. Other departments may have different rules, stricter or looser.
A hypothetical example of a commentYour acknowledgments say AI helped with grammar, and your methods chapter says a tool suggested the three clusters you then interpret. Does the disclosure cover both, and did your advisor agree to the second one?
A first comment on one part of your own draft, right away and without an account.
Get a free previewThe department's Guidelines on the Use of Artificial Intelligence (AI) Tools in Qualifying Exams, Second-Year Papers, and Dissertation Research and Writing are dated May 2026. For the papers and the dissertation the starting point is a conversation: “In second-year papers and dissertations, the appropriate use of AI tools (e.g., large language models and other generative text tools [e.g., ChatGPT]) should be negotiated between advisor and advisee.”
The department calls these guidelines and expects them to be adapted by subfield, cluster, advisor and advisee. One requirement holds throughout: “Throughout, the only imperative is to be explicit: to discuss and disclose clearly and directly.”
The guidelines list uses that are typically acceptable in limited, clearly defined ways, and uses that conflict with the UCI Academic Senate Policy on Academic Integrity. In our words:
| Use | Sociology guidelines | What to settle with your advisor |
|---|---|---|
| Brainstorming research questions | Generally acceptable | Whether you mention it in the disclosure |
| Clarity, grammar or organization of your text | Generally acceptable | Which tools, and for which chapters |
| Coding, data organization and similar research tasks | Generally acceptable, unless listed under the guardrails | How you document the steps |
| Text, arguments, interpretations or findings presented as your own | Not to be generated by AI | Nothing: this is excluded |
| Fake data, invented citations or findings | Not to be generated by AI | Nothing: this is excluded |
| IRB-protected, confidential data | Not to be uploaded | How your data stay out of any tool |
Any use of AI tools is disclosed. The guidelines say where and how: “Include a concise statement in the methodology or acknowledgments section describing how AI tools were used”. The statement names the tools and the purposes, and confirms that you reviewed, verified and edited all AI-assisted content.
The guidelines also keep responsibility with the author: for the accuracy, originality and integrity of the paper or dissertation, for every claim, analysis and conclusion, and for proper citation of sources. They advise checking journal and database AI policies too, since some data providers set their own terms.
For qualifying exams the department moves the decision to the cluster coordinator and recommends a written policy. Its sample agreement allows tools such as word processing, spelling and grammar review and reference managers, and says large language models are generally not allowed unless you make a case for a specific use. Keep the two apart: what your advisor agreed for the dissertation does not carry over to an exam.
UCI Information Security has a page on using AI tools securely. It explains that a provider keeps a copy of what you submit and may train on it, and it lists which data protection levels each service may receive. Services under a UCI contract, such as ZotGPT, are cleared for more levels than outside services, which are listed for the lowest level only. The page states: “Using UCI contracted and supported services are always preferred.”
The same page says any service where a third-party supplier has access to University information needs review as an approved vendor and use case. Protected research data goes into no outside tool, and the Sociology guardrails say the same about IRB-protected data. Whether your own draft counts as University information, and at which protection level, is not settled by these texts. Ask your advisor before you upload it to Esmond or any service UCI has not contracted.
UC Irvine departments set their own expectations for theses and dissertations. To find yours:
This example is made up. A Sociology doctoral student uses an AI tool to brainstorm questions about neighborhood change, which the guidelines treat as generally acceptable. Months later he asks the same tool to propose a typology for his interview data and builds chapter four around it. The first use fits the acceptable list. The second comes close to an interpretation presented as his own, which the guardrails exclude, and he never discussed it with his advisor. His acknowledgments mention brainstorming only.
Esmond evaluates every assignment before you submit. It never writes it.
Esmond is an AI tool that doesn't write for you. If your course doesn't allow any AI tools, don't use Esmond for that work. If you're unsure, ask your instructor or advisor.
Esmond evaluates the content of a draft from within the field and points out factual errors, inconsistencies and blind spots; read how Esmond works and look at the sample report. In Sociology, agree its use with your advisor like any other tool, then name it in the statement in your methodology or acknowledgments. Esmond is an outside service, so leave IRB-protected, confidential data out of what you upload, as the department and UCI Information Security ask, and ask your advisor first whether your draft counts as University information under the security office's levels. In other departments, follow your program's text or ask your advisor first, and see whether you can use Esmond. UC Irvine has not reviewed or approved Esmond.
We found no campus-wide rule on November 9, 2026. The Department of Sociology has published guidelines for second-year papers, dissertations and qualifying exams; other departments may have their own.
The guidelines list improving clarity, grammar or organization as generally acceptable, after you agree it with your advisor. You still disclose it.
In a concise statement in the methodology or acknowledgments section, naming the tools and purposes.
Sociology excludes uploading IRB-protected, confidential data, and UCI Information Security limits outside AI services to the lowest data level.
Read on November 9, 2026.
Get your paper evaluated before you submit