Making human capability visible
Stewart Sowman-Lund’s recent article on how New Zealand universities are responding to generative AI prompted a useful conversation. Different institutions have adopted different policies, reflecting different assumptions about assessment, responsibility and academic integrity.
Those differences matter.
But I don’t think they are the deepest question.
The conversation often begins here:
Should students be allowed to use AI?
It is an understandable place to start. Institutions need policies. Students need clarity. Teachers need practical guidance.
Yet I suspect we are asking the second question before we’ve answered the first.
A more useful starting point might be this:
What human capabilities should a qualification still develop and be able to demonstrate in an AI-rich world?
That question changes everything.
Once you begin there, AI policy becomes one design decision among many rather than the centre of the discussion.
Critical thinking still matters.
Judgement still matters.
Verification still matters.
Professional responsibility still matters.
The ability to organise complex ideas, defend decisions, recognise error and adapt when circumstances change may become even more valuable precisely because AI can now generate fluent and convincing outputs with extraordinary ease.
The challenge, then, is not simply deciding when AI should or should not be used.
The challenge is ensuring that our approaches to learning and assessment still provide credible evidence of those human capabilities.
For many years, the completed assignment has acted as a proxy for learning.
That assumption has become much less secure.
A polished essay no longer tells us, by itself, what a learner actually understands, how they reached their conclusions, what they can explain without assistance, or how they would respond when the situation changes.
This does not make essays worthless.
It simply means they may no longer carry the evidential weight they once did.
That distinction matters.
The goal is not to catch students using AI.
Nor is it to celebrate AI for its own sake.
The goal is to make human capability visible.
Sometimes that might involve an oral defence.
Sometimes a professional conversation.
Sometimes a workplace demonstration.
Sometimes reflective evidence showing how decisions were made, challenged and refined.
Different disciplines will arrive at different answers.
That is entirely appropriate.
The same question is beginning to emerge beyond universities.
Employers face it.
Professional bodies face it.
Public institutions face it.
Increasingly, the issue is not whether someone can produce a polished output.
It is whether they possess the judgement to recognise when that output is wrong, incomplete, ethically problematic or no longer fit for purpose.
Perhaps, then, the long-term challenge is not writing better AI policies.
It is designing better systems for recognising, developing and assuring human capability under AI-rich conditions.
AI policy will continue to evolve. Human capability is the question that remains.
Kia ora! Hey, I'd love to know what you think.