Reference Chapter 124 · Teaching authentic work and resolving suspected misconduct without confusing an allegation, an algorithmic flag and a proven case
A teacher receives a polished English essay from a student whose earlier writing was much less fluent. The school suspects generative AI use and threatens an automatic zero. Another pupil copies a paragraph from a website because nobody taught them how to acknowledge sources. A third student’s legitimate group project resembles their partner’s work because the assignment required collaboration but the marking rules never explained individual authorship. A fourth is accused of cheating because a plagiarism detector highlights common scientific terminology.
The gold standard is a national school academic integrity and fair-review system that teaches authentic scholarship, states assistance rules before an assignment, collects reliable evidence, distinguishes mistakes from deliberate misconduct, protects students’ rights and makes proportionate decisions with a real appeal route. It must protect the meaning of academic qualifications without treating unusual writing style, high achievement or a software similarity score as proof of wrongdoing.
This chapter concerns ordinary school tasks, coursework, student research, take-home assessments and school-level alleged cheating. It complements Chapter 05: Assessment and Qualifications Integrity, Chapter 59: National Examination Administration, Chapter 75: School-Based Assessment and Chapter 81: AI Governance. Those cover qualification validity and automated decisions; this chapter focuses on the student-facing education, reporting, evidence, fairness and learning-repair service.
1. Define integrity as an educational capability
Academic integrity involves honest representation of who did the thinking, how sources contributed and which assistance was permitted. It should be taught, not assumed to be an instinctive skill every child possesses.
The International Baccalaureate academic integrity guidance, updated 26 August 2026, treats integrity as a shared responsibility across the school and emphasises authentic student work and credit for the ideas or contributions of others.
A National School Authentic Work Standard should define the learning purpose, relevant subject methods, permitted assistance, acknowledgement, supervision, suspected breach handling, student explanation and review.
The national curriculum should teach responsible use of sources and collaborative knowledge, not only publish lists of penalties. Younger learners need more explicit modelling than experienced research students.
2. Distinguish an ordinary learning error from a disciplinary case
A child who forgets to reference one source may need teaching, while a student who knowingly submits purchased work as their own presents a different concern.
A Academic Integrity Case Classification should distinguish:
- Incomplete or inaccurate citation without sufficient evidence of deliberate deception.
- Plagiarism or misrepresentation of another person’s text, ideas, images, data or work.
- Unauthorised assistance by a parent, tutor, peer or digital system.
- Collusion where independent work was required.
- Unpermitted communication or materials during an assessment.
- Fabrication of data, sources or results.
- Misrepresentation of attendance, identity or authorship.
- Improper use of generative AI under the actual assignment rules.
- Suspected breaches involving external regulated coursework or examinations.
- Staff or institutional maladministration that compromises candidate fairness.
The label should not decide guilt in advance. It identifies the process and competent authority needed.
3. Publish rules before the student starts work
An assignment cannot fairly be judged by assistance restrictions that are disclosed only after submission.
A Task Authorship and Assistance Brief should state what is assessed, whether collaboration is allowed, how to use sources, whether calculators or AI tools are permitted, what counts as the student’s own work and what records or explanation may be needed.
An English essay may allow research but require independent composition. A science project may permit group data collection while assessing individual interpretation. A Mathematics assessment may prohibit computational tools when independent manipulation is the construct.
Avoid vague directions such as “do not cheat” when the permissible boundary depends on the actual task.
4. Teach citation and source use with real examples
Students need opportunities to practise summarising, paraphrasing, quoting, referencing and distinguishing source evidence from their own reasoning.
An Age-Appropriate Attribution Lesson can compare an original text with copied wording, a correctly attributed paraphrase and an unsupported claim. The teacher should show how references help readers evaluate evidence.
No single international referencing style is appropriate for every school subject and age. The IB guidance on acknowledging sources recognises different disciplinary styles while requiring clear acknowledgement.
A student should know the rule before the work is marked, not discover it only during an accusation.
5. Separate legitimate collaboration from collusion
Working together can be a learning goal in science, design and project-based education. Some tasks nevertheless require each student to produce an individual answer.
A Group Work Authorship Record should identify shared inputs, individual outputs, permissible discussion and the way each person’s capability is assessed.
If four pupils jointly collected water-quality data, the school should explain which graphs and interpretations are expected to be individual. Similar factual measurements are not automatically evidence of wrongdoing.
Teachers should avoid designing ambiguous group tasks and then treating students’ resulting similarity as deliberate cheating.
The task must make the legitimate collaboration boundary visible.
6. Explain generative AI use according to the awarding authority
Some programmes allow properly acknowledged AI assistance; others restrict or prohibit AI for specified assessment outcomes. There is no universal school-wide rule that correctly answers every task.
The IB’s AI and assessment statement, updated February 2026 explains that the IB does not ban AI tools generally but does not treat AI-produced text or images as a student’s own unaided work; such contributions must be clearly credited in accordance with the programme’s rules.
A Student AI Assistance Boundary should specify allowed brainstorming, language support, editing, translation, generated content, calculations, source evaluation and required disclosure by task.
A school cannot import the IB policy unchanged into a different national examination. It must apply its actual syllabus and awarding body’s authority.
7. Protect independent assessment validity
If the intended outcome is writing a persuasive argument independently, submitting a generated essay may not demonstrate that skill. If the task assesses critical editing of generated material, AI output may be a legitimate starting point.
A Construct-Preserving Assessment Check should identify the academic capability, permitted tools, supervision, sources of evidence and alternative ways to show authentic understanding.
In Mathematics, a learner who can explain why an equation remains balanced after multiplying both sides by the same nonzero number has demonstrated reasoning. A tool that directly solves the entire exercise may remove the evidence the teacher needs.
Authenticity and accessibility can coexist: a justified assistive input tool should not be treated as misconduct merely because it changes the physical way work is produced.
8. Distinguish suspicion from verified evidence
A sudden change in writing style, a high text similarity percentage or an unusual solution may justify careful review. None alone proves deliberate misconduct.
A Academic Integrity Evidence Receipt should identify the assignment, rule that allegedly applied, specific observations, relevant sources, permitted tools, student work and the proper reviewing officer.
The school should preserve original submissions and avoid modifying evidence. It should also consider innocent explanations, including tutoring, language development, legitimate editing and different forms of accessibility support.
Accusation should never be presented as a final finding before the student has the required opportunity to respond.
9. Treat AI-detector scores as uncertain indicators
AI-detection products may produce false positives or misclassify work, particularly where writing is formulaic, heavily edited or uses a learner’s developing language.
A Detection Tool Reliability Gate should identify the software’s intended function, technical limitations, error rates under relevant conditions, privacy, human review and prohibited uses.
No school should automatically impose a zero mark, misconduct record or exclusion solely because a proprietary detector returns a percentage. A system that cannot explain its evidence is inappropriate as final authority over a consequential decision.
A school may instead examine permitted drafts, classroom work, source records and the student’s ability to explain the ideas. These also require fair interpretation and must not become improvised traps.
The wider human-review principles are in AI Governance and Human Review.
10. Give the student a meaningful chance to explain
A fair process should communicate the relevant concern in language and form the student can understand, consistent with national law and age.
A Student Academic Integrity Response Meeting should identify the allegation, available evidence, expected student response, relevant parent or support involvement, impartial reviewer and proper record.
A student might explain that a teacher permitted group brainstorming but required individual writing. Another may acknowledge copying because they had not understood how to paraphrase.
Listen without assuming every denial is true or every confession is reliable. Professional judgement and evidence should guide the decision.
11. Keep teachers and investigators within their authority
A classroom teacher can recognise suspicious work and collect legitimate academic evidence. A serious high-stakes misconduct case may need a designated school leader, awarding body or formal disciplinary process.
A Academic Integrity Decision-Rights Map should identify the teacher, subject lead, school reviewer, student safeguarding officer, awarding body and relevant appeal authority.
The Cambridge International malpractice guidance distinguishes candidate malpractice from centre maladministration and places reporting duties on the head of centre under its procedures.
The school must follow its actual examination and employment rules rather than let any one employee invent a penalty on the spot.
12. Separate classroom teaching errors from external examination offences
A pupil’s draft essay submitted for formative feedback differs from coursework already authenticated for a regulated qualification.
A Formal Coursework Status Check should identify whether the work is practice, internal school assessment, externally moderated coursework or national examination evidence.
Cambridge’s preventing plagiarism guidance explains that suspected plagiarism discovered before assessed coursework submission and cases identified afterward may require different handling. Its exact procedures apply to Cambridge programmes, not every national qualification.
A school should not promise a simple classroom rewrite where the awarding body has already assumed formal control over the assessment.
13. Establish a secure and proportional evidence process
An allegation may involve student drafts, sources, email, collaboration documents or submission logs. Collecting unnecessary private information can create new harms.
A School Integrity Evidence Management Plan should set lawful scope, authorised reviewers, secure retention, relevance, student privacy, opportunity for correction and final disposition.
Do not demand unrestricted access to a child’s personal device or private accounts without legitimate authority. Do not publish accused students’ names in group chats or school-wide notices.
Evidence from an external tool must be evaluated for accuracy and provenance rather than treated as automatically reliable.
14. Choose educational corrections where appropriate
Some breaches reveal a genuine learning gap: the student does not know how to cite, distinguish sources or produce independent analysis.
A Proportionate Academic Learning Repair may include instruction, revision of a draft, a new independent task or an appropriate academic consequence within the applicable rules.
The response should be differentiated by age, evidence, consequence, prior instruction and the nature of the breach. Do not treat an inadvertent primary-school referencing error as identical to intentional substitution of another person’s work for high-stakes certification.
Fair teaching-oriented correction protects integrity because students can learn the underlying standard.
15. Apply formal consequences through lawful rules
Where a breach is proven, penalties may affect marks, task resubmission, eligibility or school conduct depending on the programme and authority.
A Misconduct Decision Record should identify the rule, evidence, student response, finding, reason, authorised sanction, permitted mitigation, communication and review rights.
A school’s desire to make an example of one pupil is not a substitute for proportionate, legally defensible process.
Do not retroactively change penalty rules to match public anger about AI use. The student should know the applicable rule and have the review opportunities the system requires.
16. Protect dignity and safety throughout the case
Academic allegations can affect friendships, confidence, family trust and a student’s willingness to seek help. Some students may have legitimate disability or language support requirements.
A Student Well-being and Fairness Check should prevent humiliating public accusations, unnecessary disclosure, discriminatory assumptions and the use of coercive questioning.
Support needs should be handled through appropriate professionals without conflating legitimate accommodations with dishonest assistance.
A student can be held accountable while still being treated with respect. A school must protect both the integrity of evidence and the learner’s dignity.
17. Prevent conflicts of interest in decisions
A teacher who has a significant personal dispute with a family may not be the best sole judge of an alleged serious breach. A senior official may be closely involved in an error being investigated.
A Reviewer Independence and Conflict Record should identify personal or material conflicts, appropriate referral and separation of investigation and appeal when required.
A fair review does not necessarily require an external tribunal for every classroom error. It does require procedures proportionate to the consequence and actual legal duties.
The school should protect students and staff against improper retaliation for raising a legitimate question about assessment quality.
18. Give clear and accessible appeals
A family may dispute the evidence, the finding, the penalty or a procedural error. The correct review route depends on whether the case is internal or belongs to an awarding body.
A Student Misconduct Appeal Notice should state the decision, reasoning, competent reviewer, relevant deadline, allowable grounds and interim effect under actual rules.
The ministry should not invent an appeal right where law does not provide it, nor suggest that a school meeting replaces a formal qualification appeal.
An appropriate review might uphold the finding, change a penalty, require a new hearing or correct a record.
See Education Complaints and Independent Remedies.
19. Correct student records without permanent stigma
A misconduct finding may be overturned, a reported error corrected or a penalty expire under relevant policy. Records need accurate status and lawful retention.
A Academic Integrity Records Correction Route should identify the original case, final outcome, authorised change, notices to relevant recipients and any protected audit history required.
A school should not leave an unsupported accusation in an ordinary progress report after the case has been closed without a finding. Nor should it destroy legally required assessment evidence simply to avoid discomfort.
Academic transcripts and formal qualifications may have separate record and correction requirements.
The long-term records framework is Academic Records and Credential Preservation.
20. Audit school assessment design for incentives to cheat
A take-home task asking every student for an identical generic essay can be easily outsourced. A better assessment may require a specific subject demonstration, staged drafting or discussion of actual evidence.
An Integrity-by-Design Assessment Review should examine construct, permitted assistance, difficulty, realistic time, accessible resources, authentic checkpoints and the value of independent explanation.
No assessment can prevent every misuse. But clear teaching, purposeful task design and appropriate feedback reduce ambiguity and improve the school’s ability to verify understanding.
The school must avoid making surveillance or intrusive technology the default response to every integrity concern.
21. Include parents and tutors in honest assistance rules
A caregiver who edits a student’s spelling, a tutor who explains algebra and a parent who writes an entire essay for the pupil play different roles.
A School–Home Assistance Guide should state which kinds of help are appropriate for each task and when the student’s independent skill must remain visible.
Families should not be required to supply specialist teaching to complete ordinary school assignments. Teachers should ensure learners have been taught necessary knowledge and can seek legitimate clarification.
An external tutor may support understanding but must not falsely authenticate work that the student did not produce.
22. Develop a national integrity training and reporting system
A National Academic Integrity Capability Plan should train students, teachers, school leaders, assessment officers and families in legitimate rules, evidence handling, AI limits, procedural fairness and repair.
Relevant indicators include students’ understanding of permitted assistance, valid task design, serious cases reviewed on time, corrected false accusations, authorisation of penalties, appeals and educational outcomes.
Avoid using raw numbers of cases to rank schools. A school that teaches reporting responsibly may record more initial concerns but fewer substantiated problems.
The aim is trustworthy learning and fair decisions, not an artificially low misconduct count.
23. Worked case: the AI detector accuses the wrong English student
Imagine a Secondary 2 student submitting a well-organised English argument after practising several drafts with teacher feedback. An automated detector marks the work as likely generated and the school immediately awards zero.
A gold-standard review recognises that the detector result is uncertain. The subject lead checks the assignment rules, dated drafts, authorised feedback and the student’s explanation of the argument. The student demonstrates independent control of the sources and paragraphs.
The school corrects the grade through its formal record route, acknowledges the procedural error appropriately and changes policy so that no future serious finding rests on a detector percentage alone.
The academic standard remains strong, while legitimate student work is protected.
24. Common failure modes and repairs
Unclear rules produce retrospective accusations
Failure: staff judge a task using undisclosed collaboration or AI restrictions. Repair: publish an assignment-specific assistance brief.
Similar text is automatically declared plagiarism
Failure: quotations, common terminology or legitimate shared data are ignored. Repair: examine sources and individual authorship carefully.
AI flags become automatic penalties
Failure: uncertain software output replaces human fact-finding. Repair: require reliable evidence and authorised review.
Classroom practice is confused with regulated coursework
Failure: the school applies the wrong process and jeopardises qualification integrity. Repair: check assessment status and awarding-body authority.
Children are publicly shamed before a finding
Failure: allegations become disciplinary reputation. Repair: protect confidentiality and fair response.
Penalties teach nothing
Failure: repeat errors persist because attribution and authentic work were never taught. Repair: include age-appropriate academic integrity instruction.
25. The first 180 days
Days 1–30: inventory existing integrity policies, assessment providers, AI rules, misconduct procedures and urgent cases with possible unfair penalties.
Days 31–60: inspect representative tasks for clear assistance requirements and review selected cases for reliable evidence, student response and proper authority.
Days 61–90: publish a National Authentic Work Standard, academic misconduct decision-rights map and proportional evidence and appeal procedures.
Days 91–120: pilot student citation instruction, teacher training and fair human review of AI-related concerns and coursework rules.
Days 121–180: verify corrections, appeals, better task design and independent learning evidence; revise rules where student understanding or procedural safeguards remain weak.
26. Gold-standard acceptance tests
- Are students explicitly taught academic integrity and attribution?
- Does every consequential task disclose permitted assistance and authorship requirements?
- Are collaboration, plagiarism, AI misuse and staff maladministration distinguished?
- Does a reported concern identify the actual rule and relevant evidence?
- Are students heard appropriately before a serious finding?
- Are algorithmic or similarity scores prevented from becoming automatic verdicts?
- Do regulated assessments follow the competent awarding body’s procedures?
- Are findings and sanctions proportionate, authorised and properly recorded?
- Can families use legitimate review and correction routes?
- Does the system preserve genuine student knowledge while reducing repeated dishonest practice?
27. Parent and student FAQ
Does using ChatGPT automatically count as cheating?
Not in every task or programme. The decisive question is what the school and awarding body permit and whether the student honestly discloses AI-generated content and demonstrates the assessed capability.
Is a high similarity or AI-detection score proof?
No. Such tools can provide leads for review but are not, on their own, reliable final proof of misconduct.
Can a teacher give zero immediately?
The answer depends on the actual task and authorised rules. A serious misconduct allegation should follow the required evidence, decision and review procedure.
Can a student receive help from a tutor?
Often yes for legitimate teaching and practice. But submitting work substantially completed by another person as independently authored may breach the relevant rules.
28. Reference library and related chapters
Primary sources are International Baccalaureate, Academic Integrity (updated August 2026), IB, AI in Assessment and Education (updated February 2026), Cambridge International, Preventing Plagiarism Guidance and Cambridge International, Malpractice and Maladministration Procedures. These are awarding-body examples and should not be represented as binding rules for all national examinations.
This is Chapter 124 of The Gold Standard of a Ministry of Education Reference Manual. Also read Homework Policy and Family Burden, Examination Item Banks and Score Comparability and Student Discipline and Procedural Fairness.
Final principle
The gold standard of academic integrity is not the harshest penalty or the most sensitive plagiarism detector. It is an education system in which students understand what authentic work means, teachers design tasks that reveal actual capability and every allegation is handled through trustworthy evidence and fair human judgement.