← Back to blog

Plagiarism and AI at UCLA: undisclosed is plagiarism, prohibited is cheating

|9 min read

Check your writing for AI text — free

The first 1,500 words are free, with no sign-up. Every verdict shows how often it is wrong about verified human writing — a figure no other detector publishes.

We are building a writing workspace: your Word or LaTeX document, your PDFs beside it, and an assistant that can only cite what is actually in them — see it and get notified.

The short answer

UCLA puts AI into its conduct code twice. Using it without saying so is plagiarism; using it where your instructor forbade it is cheating. Those are two separate offences and you can commit either without committing the other.

And the definition removes a defence most students reach for first: plagiarism at UCLA applies “with or without the intent to deceive”.

Two offences, not one

Most universities in this series either fold AI into plagiarism or give it a category of its own. UCLA does both, in two clauses, and the split is clean.

Undisclosed use is plagiarism — § 102.01c:

“Plagiarism includes, but is not limited to, the use of another person’s work (including words, ideas, designs, or data) without giving appropriate attribution or citation. This includes […] use of artificial intelligence to compose work (including words, ideas, designs, or data) without giving appropriate attribution or citation; or representing an altered but identifiable work of another person or the Student’s own previous work as if it were the Student’s original or new work.”

Prohibited use is cheating — § 102.01a, as an example of failing to follow instructions:

“…or the failure to observe the expressed procedures or instructions of an academic exercise (e.g., unauthorized or prohibited use of artificial intelligence to compose or complete assignments, examination instructions regarding alternate seating, or conversation during an examination).”

The practical consequence is worth spelling out. Disclosing your AI use does not cure a prohibition: if the assignment said no AI, saying you used it makes you honest, not compliant. And permission does not cure non-disclosure: if AI was allowed but you did not attribute it, that is the plagiarism clause.

You need both — permission and attribution. Check the assignment for the first, and cite for the second.

The same clause catches reusing your own earlier work — “the Student’s own previous work as if it were the Student’s original or new work” — and sets a default worth knowing: “Unless otherwise specified by the faculty member, all submissions, whether in draft or final form […] must either be the Student’s own work or must clearly acknowledge the source.” Note that it reaches drafts.

UCLA publishes no prescribed citation form for AI, so use a clear, specific acknowledgement — our AI disclosure statement guide covers what one should contain.

Intent is expressly irrelevant

Four words inside the plagiarism definition do a lot of work: “with or without the intent to deceive”.

That closes off the most common instinct — “I wasn’t trying to cheat”. At UCLA that sentence is not a defence to the finding. It may be highly relevant to the sanction, which is a separate stage, and that is where it should be argued.

Practically, it means carelessness carries real risk. A citation you meant to add and forgot, a paragraph pasted from a chatbot as a placeholder and never rewritten, a reused paragraph from last term’s essay — all sit inside the definition regardless of what you intended.

The best protection is the ordinary one: keep drafts and version history, so you can show how the work came together. See how to prove you wrote it yourself and, if a case has started, what to do when you are accused.

If a detector score is what started this, our guide to an AI detector flagging your thesis sets out what that number does and does not establish on its own.

Degree revocation, and its safeguard

UCLA is the only university in our US set with an explicit, procedurally specified power to take a degree back:

“If, after a degree has been awarded, a degree recipient is found responsible for a violation of the UCLA Student Conduct Code involving academic dishonesty or fraud committed while a Student, then the Dean of Students may impose, as a sanction, a revocation of the degree, subject to the following procedures: The Dean of Students will submit a recommendation of revocation of the degree to the Academic Senate and obtain the concurrence of the Chair of the Academic Senate.”

The safeguard matters as much as the power. This is not something a conduct officer can do alone — the Chair of the Academic Senate must agree, and for a graduate or professional degree the Dean of the Graduate Division is brought in too.

A more immediate rule affects anyone close to finishing: “no degree may be conferred on a Student until all allegations against a Student and any assigned sanctions and conditions have been fully resolved.” An open case stops your graduation, whatever it eventually decides.

The rest of the ladder runs: Warning; Disciplinary Probation; Deferred Suspension; Deferred Dismissal; Suspension — during which you may not “transfer or register for courses at another campus or location of the University of California”; and Dismissal, for “an indefinite period”, with readmission anywhere in the UC system requiring “the specific approval of the Chancellor”.

One clarification on records: UCLA works through holds — which “may prevent the Student from registering; and/or from obtaining transcripts, University services, or a diploma” — rather than a permanent transcript notation. No standing disciplinary notation is documented.

The university must prove it

The code defines the standard rather than merely naming it, which is unusual and useful:

“The Standard of proof that will be used in a Student Conduct Review and in Student Conduct Hearings is ‘preponderance of evidence,’ i.e., the University must prove that it is more likely than not that the student committed the misconduct of which they are accused.”

The burden is UCLA’s. And there is a named outcome for weak cases — “Insufficient Information means that the available information is insufficient to support the potential finding of a violation […] based on the preponderance of evidence standard.”

Cases are handled by a Reviewer in the Office of Student Conduct, “generally within 90 business days of issuing a Notice of Allegations”. Contested serious cases go to a Hearing Body whose panel members are appointed by the Vice Chancellor for Student Affairs.

Accepting the outcome closes the door

This is the procedural point most worth understanding before you respond to anything.

If you accept the administrative resolution, “the outcome is final, and no Appeal […] or Hearing […] will follow. The matter is closed.” There is no later route back.

If you do not accept it, the path depends on severity:

  • No suspension or dismissal: you may appeal. Within five days you must say in writing whether you are appealing the sanction only or both finding and sanction; you then have a further five days to submit the written appeal to the Dean of Students.
  • Suspension or dismissal:you may “contest the Reviewer’s determination of responsibility by proceeding to a Hearing”.

And a sanction-only appeal is confined to a single ground: “the proposed sanction is disproportionate given the findings of fact.” If you dispute the facts, you have to say so at the first stage — appealing the sanction alone concedes the finding.

Five days is short. Decide deliberately rather than signing to make it go away.

What UCLA does not publish

Several things we looked for and could not find, stated as absences rather than filled in:

  • Any position on AI detection.The Interim 2026 code returns no matches for “detect” or “Turnitin”, and the Teaching & Learning Center’s AI hub, its AI resources index and the CEILS teaching guide say nothing about detection tools in either direction. That is different from a refusal: Harvard, Stanford and Columbia each published a reliability warning, and UCLA simply has not published on the question.
  • Any similarity vendor. None is named in the code or on the conduct or teaching pages.
  • Any student self-check route, and any repository or retention statement.
  • Any annual case counts.

The nearest thing UCLA publishes to a standard for adopting such tools is one of the six principles of its Advisory Committee on AI in Teaching and Learning, established in spring 2025 by the Vice Provost for Teaching and Learning: “Adoption Decisions Driven by Research”. It names no tool.

Whatever any institution decides, the limit on these tools is the same, and it applies to ours: our published error rates are measured on English and German corpora, and no detector produces proof. See how accurate AI detectors really are and what a Turnitin score actually means.

Which code applies to you

UCLA currently publishes two conduct codes side by side: the 2021 UCLA Student Conduct Code, and the Interim 2026 code, which is the source of everything quoted here.

The Interim 2026 code carries no printed effective date, no adoption date and no approving body. For a document that can revoke a degree, that is a genuine gap — and it makes it hard to be certain from outside which code governs a given incident. The 2021 page notes it supersedes an earlier code “for cases prior to January 15, 2021”, which tells you UCLA does apply codes by incident date.

If a case concerns you, ask the Office of Student Conduct in writing which code applies to your incident date. That is a fair question and the answer should be on the record.

One more jurisdictional wrinkle: complaints go to the Office of Student Conduct “except those in which the Student is subject to an approved disciplinary procedure of a School, College, or other University unit”. If you are in a professional school, its own procedure may displace the central one.

Before you submit

  • You need permission and attribution. They are separate clauses and each fails on its own.
  • Do not rely on not having meant it. The definition says “with or without the intent to deceive”.
  • Attribute in drafts too. The default rule reaches “draft or final form”.
  • Do not reuse your own previous work as new work.
  • If an outcome arrives, do not accept it reflexively. Acceptance is final, and you have five days to say otherwise.

If you want a read on your own text before you hand it in, our AI check is free up to 1,500 words, with no account and no name.

Sources

This page is orientation, not legal advice. What binds you is your assignment instructions and the UCLA Student Conduct Code that applies to your incident date.

Frequently Asked Questions

Is using AI plagiarism at UCLA?

Using it without attribution is. The Interim 2026 Student Conduct Code writes it into the plagiarism definition: “use of artificial intelligence to compose work (including words, ideas, designs, or data) without giving appropriate attribution or citation.” Separately, using AI where your instructor has prohibited it is cheating under a different clause.

Does intent matter for plagiarism at UCLA?

No, and the code says so in the definition itself: work is plagiarism whether presented “with or without the intent to deceive”. Intent may affect the sanction, but it does not affect whether the conduct falls within the definition.

Can UCLA take back a degree?

Yes, and it is the only institution in our US set with an explicit procedure for it. If a graduate is found responsible for academic dishonesty committed while a student, “the Dean of Students may impose, as a sanction, a revocation of the degree” — but only after submitting a recommendation to the Academic Senate and obtaining “the concurrence of the Chair of the Academic Senate”.

Does UCLA use an AI detector?

It does not say. A full-text search of the Interim 2026 code for “detect” and “Turnitin” returns no matches, and UCLA's Teaching & Learning Center AI pages contain no statement about detection tools in either direction. That is a documented absence rather than a stated refusal — unlike Harvard, Stanford and Columbia, each of which has published a reliability warning.

Check your writing for AI text — free

The first 1,500 words are free, with no sign-up. Every verdict shows how often it is wrong about verified human writing — a figure no other detector publishes.

We are building a writing workspace: your Word or LaTeX document, your PDFs beside it, and an assistant that can only cite what is actually in them — see it and get notified.