Recent research by NZCER offers a glimpse into the perception and use of generative artificial
intelligence (AI) primary classrooms (Years 5–8). Survey responses were received from 266
primary school teachers, who were interested in the area, and 147 students spread across four
case study schools.

 Earlier this year, NZCER invited Kiwi teachers to participate in a survey to help build a picture of current perceptions and uses of AI in teaching and learning. Research objectives included:

• To what extent are AI tools being used in New Zealand’s primary schools?
• How do teachers and students perceive AI?
• How are these tools shaping teaching and learning experiences?
• What opportunities and challenges are teachers and students encountering?

In addition to examining the current state of AI adoption, the project also considered its implications for effectiveness, equity, inclusivity, and ethical use in the classroom.

And the survey says …

The results have now been published in the report ‘Generative artificial intelligence in Aotearoa New Zealand primary schools – Teacher and student survey findings’ (see link at end to download your copy).

“Taken together,” said the report, “the surveys suggest that generative AI is in use in New Zealand primary schools by both teachers and students, as well as pointing to gaps in policy, resourcing, and professional learning.”

Teacher findings

• Surveyed primary teachers were experimenting widely with generative AI tools, chiefly for lesson planning, assessment design, and personalising learning materials.
• Most teachers relied on free chatbots, especially ChatGPT and Google Gemini, supplemented by AI-enabled content platforms, such as Twinkl.
• Three-quarters of responding teachers have no school-funded premium AI tool access, and a similar proportion do not pay for their own tools, leaving most teachers using older, less capable large language models (LLMs) that are often more prone to error or bias.
• Even among the survey’s group of disproportionately interested teachers, fewer than half felt confident teaching responsible AI use, and most (85 per cent) wanted more training in the use of generative AI tools.
• Teachers seemed unsure about what environmental impact generative AI might have and whether the educational benefits outweigh environmental costs.

Student findings

• Roughly nine in 10 a-konga had heard of AI and more than half reported using generative AI tools, though regular (at least weekly) users were still in the minority.
• Self-reported student use was higher outside school than inside, with activities ranging from drawing images and chatting with AI tools ‘like a friend’ to fixing writing or getting homework ideas.
• Many viewed AI as helpful yet also ‘a bit like cheating’ and most were unsure of the rules at their schools.
• Most students did not recall adult guidance on when or how to use AI.

Report recommendations

“This report suggests a need for centrally supported professional learning, better privacy-protected access to premium LLMs, culturally supportive school and national policies, and iterative integration of ageappropriate AI critical literacy in the curriculum,” said NZCER. “Aotearoa New Zealand could potentially draw on emerging frameworks in countries such as Australia, Canada, and the UK, tailoring these to honour Te Tiriti o Waitangi and other aspirations for the education of tamariki.”

Download the report at nzcer.org.nz/ research/publications/generative-artificial-intelligence-aotearoa-new-zealand-primary-schools

INTERFACE August 2025