AI in tertiary education is already here. What’s missing is orientation.
There’s no urgency in this.
AI has already entered tertiary education. Not as a future signal, but as a present condition. Most educators have seen it. Many have tested it. Few feel settled with it.
The pattern is consistent. Curiosity alongside uncertainty. The tools are accessible, but the entry point isn’t obvious. People aren’t asking whether AI will matter. They’re asking where to begin — and what holds.
That question hasn’t been well served.
What’s available tends to drift in two directions. Either technical explanation without context, or application without grounding. In both cases, something is missing. Not capability, but orientation.
At the same time, the systems shaping AI sit elsewhere. The dominant lenses are commercial, global, and culturally narrow. Education in Aotearoa is responding to those systems more than shaping them.
There’s a gap in that.
Not just in skills, but in judgement. Not just in use, but in values.
Over the past few months, a small group of us have been working in that space.
Through the AI for Good platform, we’ve developed a set of learning collections for tertiary education in Aotearoa.
They sit here: in the Ako AI Tertiary learning pathways as part of the AI for Good platform
It’s not a comprehensive curriculum. It’s not a finished system. It’s a way in.
The collections move across six areas:
- teaching and learning
- classroom use
- assessment
- culture and inclusion
- ethics and responsibility
- policy and governance
Each one could stand alone. Together, they form a surface you can move across.
What matters here isn’t the topics. It’s how they’re held.
The assumption underneath the work is simple. Values aren’t something you layer on later. They’re structural.
A kaupapa Māori lens isn’t presented as a module. It sits inside the design. Concepts like manaakitanga, kaitiakitanga, and rangatiratanga don’t appear as references. They shape what is emphasised, what is left out, and where the boundaries are.
That includes a question that doesn’t get asked often enough.
Not what AI can be used for — but what it shouldn’t be used for.
There are domains where these tools don’t belong. There are forms of knowledge that shouldn’t be externalised into systems trained offshore. There are relationships and responsibilities that sit outside the logic of optimisation.
Those aren’t edge cases. They’re central.
At the same time, refusal on its own doesn’t hold.
The capability is here. It will keep expanding. The question isn’t whether AI enters education, but how.
That creates a tension that doesn’t resolve cleanly.
On one side, acceleration — largely commercial.
On the other, the need to hold people, knowledge, and environment with care.
Education sits between those forces. Interpreting. Translating. Responding.
This work doesn’t resolve that tension.
It sits inside it.
There are larger questions already forming.
Data sovereignty. Infrastructure. Whether Aotearoa shapes its own approach, or continues to inherit models developed elsewhere.
Those questions won’t be answered quickly.
For now, this establishes a piece of ground.
A way in that doesn’t require certainty — only attention.
What follows will not come from a single move.
It will be built, iteratively, in the open.
This is one of the first pieces.