It depends on the assessment: there is no blanket ban or blanket permission across NCEA. The governing rule is authenticity, so work a student submits must genuinely be their own. Supervised external examinations are closed to AI and devices, while internal assessments vary, so schools set and communicate what each task permits.
Is there a single rule that covers all NCEA assessment?
No, NCEA has no single AI rule, because it covers two very different kinds of assessment. Internally assessed standards are set and marked within your school, often over several lessons, while externally assessed standards are examined or moderated by NZQA. What a student may use, and under what conditions, changes with the format, so the honest answer to ‘is AI allowed’ is always ‘it depends on the specific assessment.’
This matters because students, parents and even colleagues often want a blanket yes or no, and giving one is misleading. A supervised end-of-year exam and a research-based internal task sit at opposite ends of the spectrum: one removes access to any tool, the other may be produced at home over weeks. The framework that ties them together is not a list of banned apps but the requirement that the work is authentically the student’s own.
What is the authenticity rule and why does it matter?
Authenticity is the principle that a student’s assessment work must genuinely be their own, and it is the rule every NCEA judgement rests on. NZQA and schools require that the evidence a grade is based on reflects what the student actually knows and can do. Submitting generative-AI content as your own writing, analysis or ideas breaches that principle in the same way copying a classmate or a website would.
Because authenticity is the anchor, the real question is rarely ‘did a student touch an AI tool’ but ‘can we trust that the assessed work is theirs.’ A breach can be treated as misconduct and can put the grade at risk, so both teachers and students need clarity before the work begins. Framing AI around authenticity, rather than a fixed ban, also keeps your approach workable as tools and official guidance keep changing.
Is AI allowed in NCEA external examinations?
No, in supervised NCEA examinations generative AI is not available or permitted, because students sit them under closed conditions without personal devices or internet access. These exams are designed to capture what a student can do unaided, so the question of AI use largely does not arise: there is simply nothing to type a prompt into. The same applies to any calculator or resource rules the subject already sets.
Not every externally assessed standard is a written exam, though. Some subjects are assessed through portfolios or submissions produced over time and marked externally by NZQA. For those, the supervised-exam logic does not apply, so authenticity requirements and the specific conditions of that standard govern any AI use, including declaring how tools were used. Always check the conditions published for the particular external standard rather than assuming ‘external’ means the same thing everywhere.
Can students use AI in NCEA internal assessments?
Sometimes: internal assessment is where AI use may be possible, but only on the conditions your school and the task set, and only where authenticity can still be assured. Because internal standards are managed within the school, teachers have real discretion over whether a tool may be used, for what, and how its use must be acknowledged. That discretion comes with responsibility, because you must be able to verify that the graded evidence is the student’s own.
In practice this means deciding, task by task, where AI could support learning without undermining what you are actually assessing. A tool that suggests essay structure is very different from one that writes the analysis a standard is meant to measure. Where you do permit AI, students should record and declare how they used it, and you should design the task so their own thinking stays visible and verifiable.
- Whether AI may be used at all, and for which parts of the task.
- What counts as acceptable support versus doing the assessed thinking for the student.
- How students must acknowledge or declare any AI use.
- How you will check authenticity, such as drafts, checkpoints or discussion.
How can teachers verify authenticity and design AI-resilient tasks?
You verify authenticity by building process evidence into the task rather than judging a final product alone. When you can see drafts, planning, in-class writing and a student’s own voice develop over time, a piece that suddenly appears polished and generic stands out. Designing for this up front is far more reliable than trying to detect AI after submission, since detection tools are inconsistent and can wrongly accuse honest students.
Practical task design also reduces the incentive to outsource work. Tasks anchored in class texts, personal reflection, local context or specific evidence are harder to generate convincingly and easier to verify. Short verbal check-ins, where a student explains their reasoning, confirm authenticity quickly without a formal investigation. The goal is not surveillance but assessment you can genuinely stand behind.
- Collect drafts and planning so you can see the work develop.
- Include in-class or supervised checkpoints for the key parts.
- Anchor tasks in specific texts, evidence or personal context.
- Use brief oral conversations to confirm a student's understanding.
How should teachers set clear AI expectations with students?
Set expectations before the work starts, in writing, and specific to each assessment. Students cannot follow rules they were never told, and vague guidance invites both accidental breaches and unfair suspicion. State plainly whether AI may be used, for which parts, and how any use must be declared, then connect it back to the authenticity principle so the reasoning is clear rather than arbitrary.
It also helps to teach responsible use rather than only prohibiting it. Many students will meet these tools in study and work regardless, so showing where AI genuinely supports learning, and where it quietly replaces the thinking a standard measures, builds judgement they can carry into every task. Consistent, well-communicated expectations protect honest students and make any genuine breach easier to address fairly.
How can a tool like JeddAI help teachers manage all this?
A tool like JeddAI helps on the teacher’s side of assessment, not the student’s. It drafts feedback and marking aligned to your own rubric, success criteria and comment banks, then leaves you to review, edit and decide the final judgement, so the professional call and the accountability stay with you. That frees time you can spend designing authentic tasks and checking student work.
Used this way, AI supports consistent, criteria-based marking without touching the authenticity of what students submit. If you want to apply your criteria consistently and reclaim marking time while staying in control, you can Get started with JeddAI.
| Aspect | Internal assessment | External examination |
|---|---|---|
| Who sets conditions | The school and classroom teacher | NZQA, under national examination rules |
| Where it is done | In class or over time, often at school and home | In a supervised examination room |
| Device and AI access | Depends on the task's stated conditions | No personal devices, so no generative AI |
| Authenticity check | Teacher verifies via drafts, checkpoints and discussion | Assured by supervised, closed conditions |
| If AI is permitted | Only where declared and the work stays the student's own | Not applicable in supervised exams |
Frequently asked questions
Will a student be penalised for using AI in NCEA?
If AI use breaches authenticity or the stated conditions of the assessment, it can be treated as misconduct and may put the grade at risk. Consequences follow your school's process and NZQA's rules, so clear expectations up front protect everyone.
Can teachers use AI to help mark NCEA internal assessments?
Teachers can use tools to support marking and feedback, but the professional judgement and the final grade remain the teacher's, and internal moderation still applies. The tool assists the decision; it does not make it.
Are NCEA external exams open-book or online?
Most external written examinations are supervised and closed-book, with no personal devices or internet. Some subjects are instead assessed through externally marked portfolios, which carry their own conditions and authenticity requirements.
Where can teachers find the current AI rules for NCEA?
Check NZQA's guidance and the specific conditions published for each standard, alongside your school's own policy. Because guidance on generative AI keeps evolving, confirm the current rules rather than relying on last year's.
Get started with Jeddle
Jeddle gives teachers and students instant, syllabus-aligned feedback powered by JeddAI.
Looking for study material? Browse Jeddle's Australian-English subject resources, or explore more articles on AI in Education.
