What Happened to My Agency? Also, Why Do I Have 400 Photos of Mountains?
Some conferences leave you with a notebook full of ideas. Others leave you with new friendships, new questions and an alarming number of photos of mountains that all seemed completely necessary at the time. ASEF ClassNet19 in Innsbruck somehow managed all three, and looking back, I think that combination captures what made the experience so valuable.
This year’s theme, Future-Ready Teaching: Human Agency in the Age of AI, could hardly have been more closely aligned with the questions I have been wrestling with in my own teaching in Aotearoa New Zealand. As AI becomes increasingly capable and increasingly present in classrooms, I keep returning to one deceptively simple question: how do we use these tools to expand what students can do without quietly removing the thinking, judgement and decision-making that make the learning genuinely theirs?
ClassNet19 helped me explore that question alongside teachers from across Asia and Europe, and perhaps more importantly, it reminded me that this is a question we should be solving alone. Something is reassuring about sitting around a table with teachers from very different systems and cultures and discovering that, despite all those differences, many of us are lying awake thinking about remarkably similar things. Apparently worrying about whether AI is helping students think or simply helping them finish faster is now a genuinely international pastime.
This year was also slightly different for me because it was my first experience attending ClassNet as a mentor. I will admit that the word “mentor” made me feel as though I should arrive carrying wisdom and certainty. The reality is I learned at least as much from the teachers I worked with as I could ever hope to contribute myself, and that quickly became one of the most meaningful parts of the experience.
I had the privilege of supporting two teams, Neural Nomads and HumanAIze Blue, as they developed their Innovative Teaching Practices. They approached the challenge differently, but both explored something increasingly important: what should remain distinctly human when AI can generate, translate, suggest, summarise and create almost instantly? That question sounds philosophical, but in the classroom it becomes very practical very quickly.
One idea that particularly stayed with me came through the HumanAIze work. Rather than simply asking students to use AI, we encouraged them to make their judgement visible by deciding whether to Accept, Adapt or Reject what the AI suggested and, crucially, explain why. It sounds like a relatively small change, but it shifted the focus of the learning quite significantly because the important output was no longer simply what the AI produced; it was the human decision that followed.
Students were therefore asked to consider why they trusted a response, what evidence supported it, what the AI might have overlooked and whether they could improve on what it had offered. Sometimes the best decision was to use the suggestion. Sometimes it was to change it. Sometimes it was to confidently decide the AI was wrong and head in another direction. That final option is especially important, because I do not want students leaving school believing that confidence should always belong to the machine.
Comparing, questioning, arguing, reconsidering and deciding are where the learning sits. If we remove all of that in the pursuit of efficiency, we may end up with students producing better-looking work while having made fewer meaningful decisions themselves.
That was probably my strongest takeaway from ClassNet19. Our challenge is not simply to teach students how to use AI effectively. It is to help them develop the judgement to know when to accept what it offers, when to adapt it and when to say, with good reason, “No, I think there is a better way.” Prompting may be useful, but judgement is much harder to outsource, and I increasingly think that is where some of the most important learning now needs to happen.
One of the strengths of ASEF ClassNet is that these conversations do not happen within a single national or cultural perspective. Sitting with teachers from countries with different education systems, languages, resources and expectations quickly challenges the idea that there is one universal answer to how AI should be used in education. A strategy that works beautifully in one classroom may make very little sense somewhere else.
Those conversations matter because they move us away from talking about “AI in education” as though education were one uniform thing. Instead, we begin to see how teachers and students in different contexts make thoughtful decisions about technology based on culture, community, resources and student need. That seems far more useful than simply chasing the newest tool and assuming innovation will somehow emerge automatically.
This is also where ASEF ClassNet becomes about much more than technology. The technology may help connect us, but the programme is really about people: teachers sharing practice, students sharing perspectives, cultures meeting rather than simply observing each other from a distance, and teachers discovering that the questions they are asking locally are often being asked by colleagues thousands of kilometres away.
There was something quietly reassuring in recognising that concerns about human agency are not unique to Aotearoa, nor to one curriculum or one group of teachers. Across the programme, I kept hearing variations of the same fundamental question: how do we prepare young people to live and learn alongside increasingly capable AI while ensuring they continue to believe that their own ideas, choices and voices matter?
I do not think ClassNet19 gave us a neat answer to that question, and I am increasingly suspicious of educational events that claim to provide one anyway. What it did provide was something more useful: a community of teachers willing to keep exploring the question together and testing ideas in the classroom.
Perhaps the future of education with AI, then, is not about making learning completely frictionless. In fact, I suspect we sometimes need to deliberately preserve the right kind of friction: the pause before accepting an answer, the conversation with someone who sees the world differently, the uncertainty that requires a student to make a choice and the moment when a student decides, “Actually, I disagree.”
Those moments may not look particularly efficient, and they are unlikely to feature in a glossy technology demonstration, but they are deeply human. If anything, the rapid development of AI has made me more convinced that our role as teachers is not simply to help students work with increasingly intelligent machines, but to create learning environments where their own judgement remains necessary.
I am enormously grateful to the ASEF ClassNet team, the teachers involved, my fellow mentors, and particularly the Neural Nomads and HumanAIze Blue teams for allowing me to share a small part of their journey. The generosity with which people shared their ideas, challenges, cultures and classrooms made the experience meaningful.
If we want education to remain genuinely human in the age of AI, perhaps one of the most important things we can preserve is our willingness to keep questioning not only what the technology can do, but what happens to our own agency while it is doing it.
You can check out my mentees' projects below:
Comments