Plagiarism and AI at Monash: every text assignment is checked at the moment you submit

The short answer

At Monash, every text-based assignment is run through a similarity-detection system at the moment you submit it. It is not optional, and it is not something a marker chooses to do. That is written into the policy, and only a dean can approve an exception.

The good news sits right beside it. Monash tells markers to use their own academic judgement “rather than relying on the reports’ raw scores”, requires misconduct to be “knowing and/or reckless”, and expressly lets you run your own work through the same system before you submit. Use that entitlement.

The check is compulsory, and automatic

Monash defines the tool class in its own policy dictionary, and names the products:

“Similarity-detection system — A system (e.g. Turnitin or MOSS) that compares text, data, code or other elements in a student assessment against various sources including the internet, published works, commercial databases and assessments previously submitted through the system.”

Then it makes submission through that system mandatory:

“1.23. Students must submit assessments online through the learning management system and assessments will be processed using a similarity-detection system at the point of submission to promote honest academic practice and assist in detecting suspected breaches of academic integrity. Exceptions can be approved by the dean (or delegate) based on legitimate pedagogical reasons where online submission or the use of a similarity-detection system is not appropriate.”

The Assessment Regime Procedure says the same thing to staff:

“2.8.6. Text-based assignments must be submitted electronically and the system must use a similarity-detection system to identify potential academic integrity issues.”

And chief examiners must ensure “all text-based assignments, including eAssessments, that can be submitted through a similarity-detection system are submitted through the system provided (e.g. Turnitin). For non-text-based assignments, if alternative systems have been put in place by the faculty, these must be used (e.g. MOSS)”.

Notice what MOSS is for: code. If you submit programming work at Monash, it is compared against other submissions too.

The similarity report is not the only detection method Monash lists, and the third one is worth knowing about:

“2.3. Faculties must have processes in place to prevent and detect breaches of academic integrity. Detection methods include but are not limited to: · referring to the similarity report from a similarity-detection system; · random sampling of submitted assessment tasks to check for similarities; · comparing the performance of a student across a number of tasks; and · searching the Internet or electronic databases using likely search phrases to identify sources of suspected plagiarism.”

“Comparing the performance of a student across a number of tasks” is a human, comparative method — it is what a teaching team notices when one piece does not sound like the others. It has nothing to do with a classifier, and it is probably the most common way a question gets asked in the first place.

Markers are told not to rely on the score

This is Monash’s operative instruction to the people marking your work, and it is a good sentence to know:

“staff who are marking assessment, including eAssessments, must consider the reports from similarity-detection systems when assessing work and must apply academic judgement to determine whether there is evidence of a breach of academic integrity, rather than relying on the reports’ raw scores”

A similarity percentage is an input, not a verdict. Quotations, technical terms and reference lists all generate matches; that is how the tool works. If you would like the longer version of why a number on its own decides nothing, see what a Turnitin score actually means.

There is a second clause with a sharper edge, because it rules out an entire class of complaint being turned into a breach:

“4.7. Stylistic or formatting requirements (such as word counts or the use of a particular referencing style) are not considered conditions for an assessment and must not be used to determine whether a breach of assessment conditions has occurred. Instead, they should be included in the assessment criteria for the assessment task.”

Style is a marking matter, not an integrity matter. If your prose reads as unusually even or formal — which is exactly what careful writers and second-language writers often produce — that is a question about the assessment criteria, not a route to a misconduct finding under 4.7.

Two more clauses complete the evidential architecture. Clause 3.3: “Decision makers will determine the appropriate weight to give each piece of evidence, considering the relevance and reliability of the evidence, the concurrence of the evidence with the circumstances being investigated, whether the evidence is corroborated…” And clause 3.4: “Decision makers are not bound by the rules of evidence.”

Taken together: a tool output is one uncorroborated input, and the finding is a human judgement made on the balance of probabilities.

No published position on AI detection

We looked, and we will report it precisely rather than helpfully.

Across the full text of the Assessment and Academic Integrity Policy, the Student Academic Integrity Procedure (v4.1), the Assessment Regime Procedure (v12.0), the Student Code of Conduct (v1.2), the Hearing and Appeals Panels Procedure and the University-wide Artificial Intelligence Operations Policy, none mentions an AI writing detector, an AI indicator, or AI detection of any kind. The only tooling named anywhere is the similarity-detection system — Turnitin or MOSS — and a similarity check is a different instrument from an AI-writing classifier. Monash names the first and takes no position on the second.

One honest limit on that: Monash’s student-facing web guidance sits on hosts that refused our automated requests, so we could not check whether an operational statement exists there. Unreadable is not the same as absent. If it matters to you, ask your chief examiner directly and get the answer in writing.

What Monash does specify instead is above — human judgement over raw scores, corroboration, and comparison across a student’s own tasks. If you are worried about how a general-purpose detector might read your writing, our notes on how accurate AI detectors really are and on AI detectors and non-native speakers are the general picture, not a statement about Monash’s tooling.

Clause 4.6: what proofreading may and may not do

This is the clause most likely to affect a student using any writing tool — ours included — and it gets its own section because burying it would be dishonest.

Start with collusion, because that is where the boundary is enforced:

“4.4. Collusion is unauthorised collaboration with another person, where students work together on an assessment task which is then presented as a student’s own individual work. … Collusion occurs when a student, without the authorisation of teaching staff: … allows another person or entity to produce or edit their work, except where Disability Support Services has approved the use of a scribe…”

Then the carve-out, and its limits:

“4.6. Unless otherwise stated in the learning management system, students can have another person or entity proofread their work. Acceptable proofreading is the process of identifying errors and suggesting corrections to spelling and grammar. Proofreading by another person or entity does not include: · rewriting passages of text in order to clarify meaning; · changing the words used by the student; · rearranging or reformatting text, code or other material; · adding material or references to the original work; or · checking calculations or formulas.”

Read the phrase “another person or entity”. That is drafted to catch software, not just people. So clause 4.6 is not only about the friend who reads your draft — it is the rule that governs what a tool may do to your text at Monash.

The permitted zone is narrow and precise:

• Permitted: identifying errors and suggesting corrections to spelling and grammar.
• Not permitted: rewriting to clarify meaning, changing your words, rearranging or reformatting, adding material or references, checking calculations.

We sell a checking tool, so here is the straight version. Our check marks filler, hedging and tired transitions in your own sentences and suggests plainer wording — nothing is rewritten for you. “Suggests plainer wording” is very close to the line clause 4.6 draws, and a suggestion that changes the words you used is on the wrong side of it if you accept it. A paraphrasing tool, a rewriter or a “humanizer” is squarely outside the permitted zone at Monash — those change your words by design, which is the second bullet of the exclusion list.

And note the opening words: “Unless otherwise stated in the learning management system”. Your unit can switch even this default off. Check the LMS for your assessment before you rely on the default.

The scribe exception is the one accessibility route named in the clause, and it runs through Disability Support Services approval — get the approval rather than assuming it.

Related: Monash’s plagiarism definition catches paraphrasing that stays too close to the source even when you cite it. Clause 4.2 says plagiarism occurs when a student “paraphrases other people’s work and/or ideas without referencing the source” and also when a student “provides reference to the source but inappropriately paraphrases other people’s work and/or ideas by not expressing them in the student’s own words”. A citation does not rescue a lightly reworded sentence. That is precisely what a spinner produces.

“Knowing and/or reckless” is the threshold

Monash sets a state-of-mind requirement for a misconduct finding, which is a real protection and worth understanding properly.

“2.11. The RO must consider the relevant evidence related to the allegation. If the RO is satisfied that it is more likely than not that the alleged misconduct occurred and the conduct was knowing and/or reckless, the RO will find the allegation of misconduct proven based on the balance of probabilities.”

And the standard of proof is stated plainly, and contrasted:

“3.6. In determining suspected academic integrity breaches or alleged academic misconduct, the decision maker must be satisfied that it is more likely than not that the alleged conduct took place; the burden of proof required is the balance of probabilities, not beyond reasonable doubt.”

Two elements, then. Did it happen, more likely than not? And was the conduct knowing or reckless?

“Knowing” is straightforward — you understood what you were doing. “Reckless” is the one to think about: it describes going ahead without caring whether the rule was being broken. Not checking whether AI was permitted, when the information was there to check, is closer to reckless than to careless.

Conduct that is neither is not nothing, but it goes down a different road. Monash calls the lower rung “inappropriate academic practice”:

“1.2. Breaches of academic integrity occur on a spectrum of severity, with more minor breaches handled as inappropriate academic practice and more serious breaches handled as academic misconduct.” · “1.3. Inappropriate academic practice is conduct that can result in minor, trivial or frivolous academic integrity breaches…” · “1.5. An allegation of academic misconduct is made when a suspected breach of academic integrity is determined to be serious…”

If you are in this position and your case is genuine inexperience, this distinction — and the evidence that supports it — is what you are arguing about. See what to do when you are accused of using AI.

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.

The AI rules, and the published-reasons duty

Monash names generative AI throughout its instruments. The policy clause:

“1.17. When generative artificial intelligence is used by students as part of assessment, it is expected to be used with honesty and in a manner that is responsible and ethical. · 1.17.1. If students use generative artificial intelligence to create work submitted for assessment, it must always be acknowledged. · 1.17.2. Where use of generative artificial intelligence tools is not allowed in an assessment task, the chief examiner will specify the types of generative artificial intelligence tools (e.g. text-generating, image-generating, translators) that cannot be used. Such restrictions are recorded in the assessment documentation in the learning management system. · 1.17.3. The University will provide staff and students with resources to guide their understanding and use of generative artificial intelligence tools.”

Note that translators are named as a type of generative AI tool a chief examiner may prohibit.

AI reaches the misconduct side through the breach-of-conditions limb:

“4.7. Students must not breach the conditions for assessments set by chief examiners… Conditions may include but are not limited to: … not using generative artificial intelligence tools; or acknowledging the use of generative artificial intelligence tools in the creation of work.”

Now the part that works in your favour. Monash puts a real constraint on banning AI — the ban must be reasoned, written down, and published in advance:

“1.7. The chief examiner must determine where and how generative artificial intelligence tools can be used in a unit, an assessment task, or part of an assessment task. · 1.7.1. All specifications of use must be clearly stated in the required section in the learning management system and published before the start of the teaching period or no later than four weeks before the assessment task details are made available to students. · 1.7.2. Where the chief examiner specifies that generative artificial intelligence is not allowed, the decision should be made based on educational reasoning, the assessment task and its function in generating particular evidence of student learning. The reason must be clearly stated in the required section in the learning management system and published before the start of the teaching period…”

So the rule for your assessment should be findable in the LMS, in a specific place, ahead of time — and if AI is prohibited, the reason should be there too. If you cannot find it, that is a fair question to ask your chief examiner, and the answer is worth keeping.

Monash also builds AI into assessment design rather than only into prohibition: “2.8.1. Assessment tasks must be designed to minimise the potential for breaches of academic integrity and assess human ability in the production of the assessment task, where collaboration with artificial intelligence occurs.”

Above all of it sits a University-wide Artificial Intelligence Operations Policy (approved by the Vice-Chancellor 2 September 2024, effective 16 September 2024), which routes the obligation back to the academic instruments: “1.8. Staff and students are responsible for making clear attributions on the use or involvement of AI as specified in the Responsible Conduct of Research Policy, Assessment and Academic Integrity Policy, Student Academic Integrity Procedure…”

The Student Code of Conduct restates the duties in student-facing terms: “only use authorised resources, software and tools for assessment tasks” and “use generative artificial intelligence with honesty and in a manner that is responsible and ethical”.

On contract cheating, Monash is the institution that cites the Commonwealth legislation by its full name — the Tertiary Education Quality and Standards Agency Amendment (Prohibiting Academic Cheating Services) Act 2020 (Com) — in the related-legislation block of the Student Academic Integrity Procedure. Its definition is money-neutral and two-way: “4.9. Contract cheating is the act of engaging, or acting as, a third party to complete some or all of an assessment task on a student’s behalf, irrespective of the third party’s relationship with the student, and whether they are paid or unpaid. Contract cheating includes instances when a student either receives or offers a contract cheating service…” Helping a friend is inside the definition.

The declaration, and the module

There is a compulsory declaration at submission, and it has a hard consequence attached:

“2.19. Students must complete the assessment declaration provided when submitting assessment tasks, except for assessments completed and submitted in class. Assessment tasks that are not accompanied by a declaration will not be marked. · The declaration may cover a single assessment task, a group of assessment tasks, or an entire unit, as determined by the chief examiner. It must be submitted online, unless required by the chief examiner to be submitted in hard copy. · For group tasks, each member of the group must complete an assessment declaration.”

“Will not be marked” is the sentence to act on. If you are in a group, every member files their own.

The wording of the declaration itself is not reproduced in the procedure — it is a system artefact in the learning management system — so we cannot quote it here. Read it in the LMS before you tick it, because it is a statement you are making about your own work. Our AI disclosure statement guide covers how to word an acknowledgement in the body of a thesis, which is a separate obligation under clause 1.17.1.

The acknowledgement duty is repeated as a student responsibility: “2.15. Students are responsible for demonstrating appropriate academic practice. Specifically, students must: … appropriately acknowledge the use of generative artificial intelligence tools…” And chief examiners must teach it: students are to be “educated about the responsible and ethical use of generative artificial intelligence tools and how such tools should be acknowledged if they are used”.

There is also a module, enforced by lockout:

“2.14. Students must complete the compulsory introductory module in academic integrity at the start of their studies. Students who do not complete the module in the required timeframe will be blocked from accessing the learning management system. · 2.14.1. Coursework students complete the compulsory Introduction to Academic Integrity module. · 2.14.2. Graduate research students complete the compulsory Research Integrity module.”

You may check your own work first

Monash puts the student self-check in the procedure itself, which is unusual and genuinely useful:

“2.18. Students may use the similarity-detection system prior to the final submission of their assessment task to check and improve their work. The system may take significant time for processing and students are responsible for planning ahead to ensure they submit the task by the deadline.”

Note the framing: “to check and improve”. Monash positions the report as a learning tool, not only as an enforcement artefact. Note also the practical warning in the second sentence — processing takes time, so this is a two-days-early habit, not a five-minutes-early one.

And note the limit, honestly: the entitlement covers similarity only. Nothing in the documents we read offers students a self-service AI-writing check, and Monash takes no published position on such a tool at all.

The published penalty scale

Monash publishes its outcomes in tables, which makes the ladder unusually legible.

At chief examiner level there are three landing points. “No breach of academic integrity — Dismiss the matter and arrange for the work to be marked as normal… The student is not recorded on the academic integrity register.” A minor breach with no prior warning attracts an educative response: “The student receives a warning. The student is recorded on the academic integrity register. The mark awarded reflects the inappropriate academic practice.” And a serious breach — or a minor one where a warning already exists — is referred onward to the Responsible Officer.

A dismissed matter does not follow you: “5.8. A dismissed suspected academic integrity breach or dismissed allegation of academic misconduct must not be considered as a prior warning or finding of misconduct.”

The educative response is expressly not a punishment:

“6.1.1. When marking a task as part of an educative response, the chief examiner or nominee must not impose a penalty. A mark must be awarded that disallows the sections of the work that contain the breach…” · “6.1.2. If disallowing the sections … would result in failure of the task, the chief examiner or nominee may, at their discretion, allow the student to resubmit the work, in whole or in part, for a maximum mark of 50 per cent of the marks available for the task.”

A warning is recorded but does not appear on your transcript: “6.2. The Academic Integrity Warning Outcome is an official record that the student has been counselled… The warning is recorded on the academic integrity register but not on the student’s academic transcript. If the student is suspected on a later occasion… the later occasion must be referred directly to the RO as an allegation of academic misconduct.” One warning changes what happens next time. That is the reason to take a first, small matter seriously.

At Responsible Officer level the published options run: no penalty (“when … the decision maker decides not to impose a penalty because they are satisfied that the misconduct was minor and the student is unlikely to engage in further misconduct”); a record of misconduct; a reprimand; a consent penalty agreed with the student and “intended to educate and rehabilitate” — for example enrolling in a particular unit, developing a study plan, writing a reflection, or seeking academic counselling; disallowance of work or mark, where “the maximum mark a student can receive for a resubmitted task is 50 per cent” and a refused resubmission means zero; a mark of zero for the unit (“Failure of the unit with a mark of zero”); and a fine “not exceeding 5 penalty units”, which “is a debt to the University and failure to pay the fine can result in invalidation of enrolment”.

Suspension and exclusion are not in that table. They sit with the Student Misconduct Panel. Suspension means the student “is precluded from enrolling in any course or unit for the duration”; exclusion means “the individual ceases to be a student at Monash”. Both are held until the appeal window closes or an appeal is determined — “In exceptional circumstances, the panel may order that the suspension takes effect immediately.”

Monash flags the visa consequence explicitly, which many institutions do not: “4.5. Monash recognises that some of the penalties imposed may impact international student visas. It is the student’s responsibility to meet their visa obligations and to seek advice as early as possible.” If you are on a student visa, get advice the day a notice arrives, not after the outcome.

Penalties can be suspended: “9.5. When imposing a penalty, the RO may suspend the operation of that penalty for a period of up to five years on condition that, during the term of the suspension, the student does not commit another act of misconduct…”

And the same safeguard Melbourne uses appears in the penalty factors: prior record is sealed until the finding is made — “any previous finding of academic misconduct against the student, which the RO will not ascertain until after finding the misconduct proven and before determining the penalty.”

Degree revocation is not listed as a penalty in the documents we read, so we do not describe one.

Process, appeal, and the Ombudsman

The chain runs: chief examiner (or nominee) → Responsible Officer, supported by Student Conduct and Complaints → Student Misconduct Panel → Student Appeals Panel.

Natural justice is spelled out as your rights:

“3.1. All decision makers are bound by the rules of natural justice. This means that … students have the right to: · be informed of the process being followed, referred to support during the process and notified if unexpected delays arise; · be informed of the allegations and evidence against them and provided with a reasonable opportunity to respond; · have their response considered and the allegations determined by an unbiased decision maker who will base decisions only on the relevant evidence.”

Reasons are mandatory: “3.7. Decision makers must document: … both the decision and the reasons for reaching the decision, citing the relevant facts found to be proven; and … the evidence on which the findings were based.”

A notice of allegation must give you the specifics — the relevant sections of the Council Regulations, “the specific allegation made against the student, including the type of misconduct, when and where it is said to have taken place, the role the student is alleged to have played”, and “a copy of the evidence for the misconduct”. If a notice arrives without the evidence attached, ask for it.

If your unit results are finalised before the investigation is, the chief examiner is instructed to assign a grade of Withheld (WH) until the outcome is known.

The outcome notice must include the decision, the reasons “including relevant facts found to be proven”, the evidence, the penalty and its reasons, and “the process and timeframe for lodging an appeal” — normally within seven working days.

The appeal reaches the merits, and the grounds are published:

“5.2. A student can make an appeal on one or more of the following grounds: · there was a reasonable apprehension of bias on the part of the decision maker. However, if the student did not object to a panel member prior to the hearing, an appeal on this ground may only be based on the conduct of the member during the hearing; · there was a breach of procedural fairness that has the potential to affect the decision; · the penalty imposed was excessive; · there is new evidence that was not reasonably available to the student during the investigation or hearing that has the potential to affect the decision; and/or · the decision was not reasonable based on the evidence that was available and relevant at the time of the original decision.”

The deadline is firm: an appeal “must … be submitted to the University Registrar within 20 working days from the date of the notice, unless the student can demonstrate extraordinary circumstances”.

One trade-off to understand before you admit anything: “11.1. … If the student admitted the misconduct during the misconduct investigation, they cannot appeal against the finding but may appeal against the penalty only.” Admitting can be the right call, and it forecloses the appeal on the finding. Get advice first.

Monash grants an express right to legal representation at the appeal hearing — the only such right we found in this Australian set: “6.12. The student has a right to be legally represented at a SAP hearing and must provide details of their legal representative at least four working days before the hearing. If the student is legally represented, Monash may also be legally represented at the SAP hearing.”

The panel can substitute its own decision: “7.2. After hearing the appeal … the SAP may: affirm the original finding … amend the original finding … or set aside the original finding of misconduct and/or penalty and substitute it with its own decision or penalty.”

Beyond that sits the University Student Ombudsman, appointed by Monash University Council “to provide an avenue for an independent and impartial review”. The complaints ladder runs through five stages, ending: “Student complaint stage 4: Internal review by the USO · Student complaint stage 5: External review”. The USO’s grounds mirror the appeal grounds (bias or procedural unfairness, new evidence, unreasonableness), “1.11. The decision of the USO is the final decision of Monash University”, and where a stage 4 complaint does not meet the grounds the USO “will provide reasons along with information on how a student may seek a stage 5 review”.

Students at the Malaysia campus have a separate route: “11.2. At the Malaysia campus, a student dissatisfied with any disciplinary proceeding may submit an appeal in writing to the Malaysian [authority]…”

Two housekeeping facts we would rather state than skip. The Student Academic Integrity Procedure carries a review date of 30 June 2025 that has passed, though the document remains current with minor amendments effective 14 March 2026; the Assessment Regime Procedure’s review date of 1 January 2024 has also passed, with a major amendment effective 27 July 2026. The documents are current. Their scheduled reviews are overdue.

Before you submit

• Assume the similarity check runs. It happens at the point of submission for every text-based assignment.
• Use clause 2.18 and self-check early. You may run your own work through the system first; processing “may take significant time”, so build in a day or two.
• Read clause 4.6 before you use any writing tool. Permitted proofreading is spelling and grammar only — by a person or entity. Rewriting, rewording, restructuring and adding material are excluded.
• Find the AI specification in the LMS. It must be there before the teaching period or four weeks before the task details — with the reason, if AI is prohibited.
• File the assessment declaration. Work submitted without one “will not be marked”, and in a group everyone files their own.
• Acknowledge any generative AI use. Clause 1.17.1 says “it must always be acknowledged”.
• If a notice arrives, diarise 20 working days for an appeal to the University Registrar — and get advice before admitting anything.

If you want a read on your own text before you hand it in, our AI check is free up to 1,500 words and asks for no name or account. Two honest caveats. Our published error rates are measured on English and German corpora, and no detector — ours included — produces proof. And clause 4.6 is the rule that matters at Monash: use a tool to see where your writing stands and to fix spelling and grammar, not to have your words changed for you. If you also want a check on your sources, see verifying AI citations, and on keeping the evidence of your own process, how to prove you wrote it yourself.

Sources

• Monash University: Assessment and Academic Integrity Policy — version 2.1 (administrative amendments effective 20 September 2024); approved by Academic Board, Meeting 1/2020, 19 February 2020; effective 1 January 2021; owner: Deputy Vice-Chancellor (Education). Source of clauses 1.16, 1.17, 1.21, 1.23 and the similarity-detection system definition.
• Monash University: Student Academic Integrity Procedure — version 4.1 (minor amendments effective 14 March 2026); approved by the University Education Committee, Meeting 5/2022, 20 June 2022; effective 30 June 2022; review date 30 June 2025, which has passed. Source of clauses 2.3, 2.7, 2.11, 2.14, 2.15, 2.18, 2.19, 3.1–3.7, 4.2, 4.4, 4.6, 4.7, 4.9, 5.7, 5.8, 6.1, 6.2, 9.3, 9.5, 11.1 and 12.4.
• Monash University: Assessment Regime Procedure — version 12.0 (major amendment effective 27 July 2026); Learning and Teaching Committee, Meeting 9/2019, 2 December 2019; effective 1 January 2021; review date 1 January 2024, which has passed. Source of clauses 1.7 and 2.8.
• Monash University: Student Academic and General Misconduct: Hearing and Appeals Panels Procedure — the Student Misconduct Panel penalties, the appeal grounds, and the right to legal representation.
• Monash University: Student Code of Conduct — version 1.2 (minor amendment effective 6 December 2025); approved by the Vice-Chancellor 21 February 2025, effective 24 February 2025.
• Monash University: University Student Ombudsman Policy — the USO is appointed by Monash University Council; the five-stage complaints ladder ending in external review.
• Monash University: Artificial Intelligence Operations Policy — version 1.2 (administrative amendment effective 27 February 2025); approved by the Vice-Chancellor 2 September 2024, effective 16 September 2024, review 16 September 2027; owner: Chief Analytics Officer.

This page is orientation, not legal advice. What binds you is your unit’s assessment documentation in the learning management system, the Monash Assessment and Academic Integrity Policy, the Student Academic Integrity Procedure, the Assessment Regime Procedure, the Student Code of Conduct and the Monash University (Council) Regulations.