AI in Education in New Zealand: A Living Whitepaper
Updated for publication on 2 April 2026
Executive Snapshot
AI adoption in New Zealand education has continued to mature since the last update on 9 March 2026. The sector is now clearly beyond first-wave debate and into a more operational phase: schools are formalising acceptable-use rules, tertiary providers are normalising controlled and disclosed AI use in assessment, national curriculum work is embedding AI more explicitly, and system agencies such as NZQA are already running AI in production workflows. Taken together, the evidence suggests a sector that is active, uneven, and increasingly structured, rather than experimental in a loose or ad hoc sense. (education.govt.nz)
Three developments stand out right now. First, AI literacy has moved further into live rollout through Day of AI Aotearoa and Kōtui Ako’s AI Digital Literacy offering. Second, the senior secondary curriculum has become more concrete, with published Years 11–13 subject descriptors now showing where AI sits in future subject architecture. Third, evidence-gathering is still catching up to practice: ERO is actively studying how schools and students are using AI, while Years 0–10 draft curriculum consultation remains open until 24 April 2026. (dayofai.org)
What Has Changed Since 9 March 2026
- AI literacy has shifted from “initiative” to “live program delivery.” On 9 March 2026, Day of AI Australia and Day of AI Aotearoa New Zealand announced free, country-specific, curriculum-aligned AI literacy programs; the New Zealand program follows a 2025 pilot in 10 schools and is designed for Years 5–10. Kōtui Ako’s nationwide AI Digital Literacy classes also began on 9 March 2026. (dayofai.org)
- Senior secondary curriculum design is now more visible. On 26 March 2026, Ministry-linked curriculum material pointed schools to published Years 11–13 subject descriptors. Those descriptors embed AI across subjects including Digital Technologies, Computer Science, and Media, Journalism & Communications. (tahurangi.education.govt.nz)
- Monitoring remains live rather than retrospective. ERO’s current work programme explicitly includes the question “How are schools and students using AI?”, and Term 1 in-class trialling of draft Years 0–10 learning areas is underway while consultation stays open to Friday, 24 April 2026. (evidence.ero.govt.nz)
Overall State of Adoption
Schools: widespread use, but formalisation still uneven
The strongest current national benchmark remains TALIS 2024. It shows that nearly two-thirds of New Zealand Year 1–10 teachers had used AI in the prior year, and 69% of Year 7–10 teachers had used AI, compared with 36% across the OECD lower-secondary average. Among New Zealand teachers who used AI, the most common uses were generating lesson plans (78%) and learning about or summarising a topic (73%), while only 12% used AI to assess student work. (educationcounts.govt.nz)
Primary-school evidence points in the same direction, but with important nuance. NZCER’s nationally representative 2024 primary teacher survey found 46% of teachers reported using AI tools in teaching, while a separate NZCER AI-focused study of a cohort disproportionately interested in AI found 69% used generative AI at least weekly, including 18% daily and 26% several times a week. These are not directly comparable datasets, but together they show that AI use is no longer marginal in schools, even if intensity varies sharply by teacher confidence and local context. (nzcer.org.nz)
The pattern of adoption is also consistent: New Zealand schools are using AI mainly for planning, drafting, adaptation, and support, not for handing over final professional judgement. That lines up with Ministry guidance, which keeps teachers central, requires human oversight, and warns against putting personal data into tools. (nzcer.org.nz)
Tertiary: from prohibition and detection to governed use
In tertiary education, the centre of gravity continues to move toward managed use rather than blanket bans. Massey University now uses an Artificial Intelligence Use Framework with assessment-level categories ranging from No AI to AI planning, AI collaboration, and AI exploration, while still prohibiting AI in secured assessments such as exams and practicals. Students are required to keep records of tools, prompts, outputs, and how they modified AI-generated material. (massey.ac.nz)
Lincoln University’s current guidance points in a similar direction. It tells students to follow course-specific rules, assume AI is not allowed where guidance is absent, acknowledge all use, and avoid relying on AI for whole assessments or undisclosed translation. The emphasis is on disclosure, examiner judgement, and assessment-specific rules rather than faith in generic detectors. (ltl.lincoln.ac.nz)
System agencies: the most mature operational AI use
The most mature AI deployment in the sector remains at NZQA. NZQA says it used an AI-powered automated text scoring tool to mark more than 55,000 writing literacy assessments in May 2025, with experienced human markers double-checking over a third of results, especially near achievement boundaries. NZQA presents this as a quality-and-timeliness tool, not a replacement for human judgement. (www2.nzqa.govt.nz)
NZQA is also using AI in service delivery. Its chatbot Awhina, first introduced in 2023, began a generative-AI upgrade in April 2025 so it could search public NZQA web content, rank relevant information, and return clearer answers with links to sources. More broadly, NZQA reports that almost all secondary schools and kura are now offering some digital assessments, which gives the agency a stronger digital base for future AI-enabled assessment operations. (www2.nzqa.govt.nz)
Policy, Governance, and Curriculum
Ministry guidance is clearer, but still principles-based
The Ministry of Education’s current guidance tells schools to:
- avoid entering personal data into AI tools,
- check outputs for inaccuracies,
- expect cultural bias,
- keep teachers central, and
- maintain a school policy covering purpose, scope, privacy, and acceptable use. (education.govt.nz)
On assessment, the Ministry remains cautious. Schools with consent to assess must include acceptable GenAI use in authenticity policy, and the Ministry states that GenAI is not permitted in NCEA external assessment. For internal contexts, AI use must not undermine the principle that assessed evidence is the student’s own work. (education.govt.nz)
Privacy and Māori data sovereignty are becoming defining governance issues
The Office of the Privacy Commissioner expects agencies using generative AI to secure senior approval, test necessity and proportionality, conduct a privacy impact assessment, be transparent, and engage with affected communities including Māori. (privacy.org.nz)
Tāhūrangi’s AI introduction material goes further in the education context, stating that Māori data is a taonga and that teachers need to understand Māori data sovereignty when making AI choices; it also says AI cannot replace engagement with local whānau, hapū, and iwi. This remains one of the clearest ways in which the New Zealand education conversation differs from more generic international AI-in-education models. (newzealandcurriculum.tahurangi.education.govt.nz)
Curriculum direction is now clearer than it was in early March
The Years 0–10 draft curriculum consultation remains open until 24 April 2026, and the draft Technology material for Years 9–10 explicitly includes AI-related concepts. (newzealandcurriculum.tahurangi.education.govt.nz)
The more important update since early March is at Years 11–13. The newly published subject descriptors show AI being embedded rather than isolated. Digital Technologies (Year 11) includes emerging technologies and ethical implications; Computer Science (Years 12–13) explicitly includes artificial intelligence and intelligent systems; and Media, Journalism & Communications addresses AI-generated texts, bias, misinformation, and source verification. Based on the currently published subject list, the system appears to be moving toward AI across subjects, not a standalone generative-AI school subject at this stage. That last point is an inference from the published descriptors rather than an explicit Ministry statement. (tahurangi.education.govt.nz)
Research Overview
Official system evidence
TALIS 2024 shows New Zealand teachers are relatively advanced in AI use and also highly alert to risks. New Zealand Year 7–10 teachers were more likely than OECD peers to see AI as useful for lesson planning and adapting material for students, but they were also more likely to worry that AI enables students to misrepresent others’ work as their own and can make incorrect recommendations. (educationcounts.govt.nz)
NZCER evidence: adoption is real, but capability gaps are still large
NZCER’s AI-focused primary-school study found that surveyed teachers were using AI chiefly for lesson planning (82%), assessment design (66%), personalising learning (65%), research for teaching (59%), and student reports (51%). It also found that 68% had observed inaccuracies in AI-generated material. (nzcer.org.nz)
The same report highlights governance weakness beneath the experimentation. Most respondents with policy visibility were unsure whether Māori data sovereignty had been considered in school AI policy, with 62% answering “I don’t know”, and 85% of the AI-focused teacher cohort wanted more training in using generative AI tools. Student use was also reported as higher outside school than inside, reinforcing that schools are no longer dealing with a purely classroom-contained issue. (nzcer.org.nz)
NZCER’s nationally representative primary teacher survey adds a second, more conservative snapshot: 46% of primary teachers reported using AI tools in teaching, and 53% selected using AI as a professional learning priority. That suggests AI has entered the mainstream, but staff capability is still not keeping pace with adoption. (nzcer.org.nz)
Case Studies
1) Aotea College: policy-led secondary governance
Aotea College created a staff AI group and built a school-wide response that includes updated authenticity statements, staff guidance on referencing AI, verbal questioning, extra assessment checkpoints, and supervised assessment components where needed. (web-assets.education.govt.nz)
Why it matters: this is a strong example of New Zealand schools moving from reactive concern to institution-level governance and process design. (web-assets.education.govt.nz)
2) Hobsonville Point Secondary School: authenticity through checkpoints and version history
Hobsonville Point Secondary School responded to increased plagiarism linked to GenAI by requiring checkpoints, original working documents with full version history, verbal checks of understanding, and restorative processes involving families. The Ministry case study says these interventions produced a significant drop in inappropriate AI use in assessments. (web-assets.education.govt.nz)
Why it matters: Hobsonville offers a durable model for authenticity control without reverting to blanket pen-and-paper responses. (web-assets.education.govt.nz)
3) New Windsor School and the Ako Hiko cluster: primary-sector experimentation
Education Gazette’s profile of New Windsor School shows AI being used in a primary setting for literacy, creativity, and critical engagement, including writing support and creative tools, alongside cyber-safety and discernment. The school’s work sits within a wider cluster model rather than isolated teacher experimentation. (gazette.education.govt.nz)
4) NZQA: AI in national assessment operations
NZQA’s automated text scoring is still the most consequential production AI use in the sector because it operates inside a national assessment workflow, at scale, with explicit human checking and quality thresholds. (www2.nzqa.govt.nz)
5) Universities: service AI and structured AI use
The University of Auckland’s AI Assistant is now handling more than 9,000 conversations and 60,000 searches each month, while also operating under a documented privacy framework. Separately, the University is leading a $1 million project to build an AI-powered te reo Māori pronunciation coaching tool with real-time personalised feedback. (auckland.ac.nz)
Why it matters: tertiary AI adoption is splitting into at least two tracks—student-service automation and learning innovation, including culturally grounded language revitalisation work. (auckland.ac.nz)
Emerging Trends
1) AI adoption is strongest in augmentation, not substitution
The dominant uses remain planning, summarisation, support, drafting, tutoring, and service navigation. High-stakes delegation of professional judgement is still limited and tightly governed. (educationcounts.govt.nz)
2) Assessment redesign is replacing detector-led enforcement
Across schools and universities, the most credible responses now involve checkpoints, oral explanation, version history, disclosure rules, and assessment-specific permissions. (web-assets.education.govt.nz)
3) AI literacy is moving toward system status
Day of AI Aotearoa, Kōtui Ako, Tāhūrangi AI resources, and draft curriculum material all point to AI literacy becoming a mainstream capability rather than a niche digital-technology topic. (dayofai.org)
4) New Zealand’s distinctive edge is cultural and ethical framing
Te Tiriti, Māori data sovereignty, te reo Māori revitalisation, and culturally grounded design are increasingly central to what “responsible AI in education” means in Aotearoa. But research still shows that policy practice often lags those stated values. (newzealandcurriculum.tahurangi.education.govt.nz)
5) The system is still in design mode
Even with visible progress, national operating models are not fully settled. ERO is still studying usage, curriculum consultation is still open through 24 April 2026, and Phase 5 senior curriculum development will continue through grouped draft releases in May, June, and July 2026. (evidence.ero.govt.nz)
Conclusion
As of 2 April 2026, AI in education in New Zealand is best described as broadly adopted, strategically cautious, and institutionally maturing. Schools are using AI widely enough that the question is no longer whether AI is present, but how it is governed. Tertiary institutions are building more explicit assessment frameworks. NZQA has already demonstrated that tightly bounded, human-supervised AI can operate inside nationally significant education workflows. (educationcounts.govt.nz)
The most important change since the previous update is not a single tool or headline. It is the growing shape of a national model: AI literacy rollout is live, curriculum architecture is getting more explicit, and governance is shifting from improvised reaction to designed practice. New Zealand is not moving toward an “AI replaces educators” model. It is moving toward a model in which AI supports planning, explanation, service delivery, and selected assessment processes, while human judgement, authenticity, privacy, and cultural legitimacy become harder—not softer—requirements. (dayofai.org)
The sector’s defining challenge is now implementation quality. The evidence suggests New Zealand has passed the experimentation threshold, but it has not yet reached a fully coherent national operating model. That makes the current moment important: the foundations for scaled, trustworthy AI in education are being built now, and the institutions moving fastest are those treating AI as a whole-system education design issue, not just a classroom productivity tool. (evidence.ero.govt.nz)