If you’ve been a subscriber to my FutureMakers newsletter or my blog posts for long you’ll know that two of the important issues I’ve been writing about more recently are the purpose of education and the role of AI in education into the future. This isn’t because these are simply that latest ‘buzz’ words or topics, but because I believe they do and will matter significantly in terms of the way we design learning and the systems that support it into the future.
At the end of last week I had the privilege of participating in a webinar hosted by a US-based organisation I’ve engaged with in the past called KnowledgeWorks. In the webinar two researchers – Rebecca Winthrop from the Brookings Institution and Katherine Prince from KnowledgeWorks – brought complementary lenses to what is genuinely one of the most important questions of our time: what does the rise of generative AI mean for how we educate young people, and what should we be doing about it right now?
If you work in education and haven’t yet had a serious, structured conversation about AI with your colleagues, this webinar is a good place to start. I’d encourage you to watch it in full (see link at the end of this post).
In this post I want to share what struck me most from this webinar, and then – because I think we’ve spent enough time just watching and listening – I want to challenge you to actually do something with it.
The Landscape: Exciting, Troubling, and Urgent
Katherine Prince opened by sketching a horizon that is both exhilarating and unsettling. Katherine suggests the next decade will likely bring AI that adapts in real time to students’ moods and learning needs, that mentors teacher candidates, that fluidly matches learners with people and places. It could mean extraordinary personalisation. It could also mean minimally viable public education held together by AGI teachers while human educators disappear from communities that need them most.
She emphasised that the future is not predetermined. But it is being shaped right now, by decisions – or non-decisions – that we are all making.
Rebecca Winthrop grounded us in the present. Her Brookings Global Task Force on AI and Education consulted over 500 students, teachers, caregivers, and technologists across 50 countries, and reviewed more than 400 studies. The findings are nuanced, but her headline is not: right now, the risks overshadow the benefits.
Her five key takeaways are worth considering further:
- It’s genuinely confusing. The line between AI for learning, for communication, and for entertainment has dissolved. Students are using general-purpose chatbots – designed for adults, not for learning – in ways that are largely invisible to schools.
- AI can genuinely help learning, but only in narrow, strategic conditions – mediated by skilled teachers, paired with vetted content, embedded in good pedagogy. That’s not what most student AI use looks like.
- The risks are serious and distinct. This isn’t like the calculator. Cognitive stunting (Winthrop’s preferred term over “offloading”), reduced creativity, social development undermined by sycophantic AI companions, amplified bias, eroding trust between teachers and students – these aren’t hypothetical. There’s already research tracking them.
- Motivation and engagement are at the heart of it. If 50% of middle and high school students are already in “passenger mode” – coasting, doing the bare minimum (a reference to her book, The Disengaged Teen) – AI that does the work for them won’t fix that. It will deepen it.
- We can still bend the arc. It’s early days. The recommendations – she calls them Prosper, Prepare, Protect – are practical and actionable. But only if we act.
The Tensions That Confront Us
Katherine then introduced three tensions that I think deserve space in any educator’s thinking: innovation versus protection; systemic vision versus lagging implementation; and efficiency versus human development.
That third one is the one I keep coming back to, and it connects directly to the conversation that closed the webinar about the purpose of education.
Both researchers landed in the same place, independently: education is fundamentally about human development – developing the self, and learning to live with others. Rebecca framed it this way: school may be the only space in many communities where young people encounter people genuinely different from their family or neighbours. That social fabric, that citizenship function, that opportunity to rub shoulders with difference – AI cannot substitute for it, and we shouldn’t let it.
The Purpose Question: A Particular Challenge for NZ Educators
Here in Aotearoa, this tension has a particular sharpness to it. Our current system is heavily oriented toward preparing students for employment the future of work. The pedagogical framing that dominates is largely transactional – collect the credits, get the qualification, access the next thing. Rebecca named this directly, saying that students have learned to treat school as cashing in points to buy rewards. When AI can do the transaction faster and better than they can, why would they bother? That’s not a student problem. That’s a design problem. And it’s one that the current direction of our education system risks entrenching rather than solving.
This puts educators in an uncomfortable position. Government mandates set the frame – curriculum priorities, assessment structures, accountability measures. And those mandates, right now, skew heavily toward economic utility. Education as workforce preparation, learning as a means to employment. Many educators feel the pull to simply comply, to deliver what is required and leave the bigger questions to someone else.
But this is precisely the moment when professional knowledge matters most. Every educator in this country carries (or should carry) a deep understanding of what learning actually is, what child and adolescent development requires, and what schools are uniquely positioned to provide that no algorithm can replicate. That knowledge isn’t just personal conviction. It’s the thing that makes teaching a profession rather than a delivery service.
Rebecca Winthrop highlighted a insight that I think could really change how we (as educators) approach this. She is emphatic that parents and whānu are the missing piece of this conversation. They are desperate for guidance. They are worried about their children. And critically, as she points out, they are the voters who ultimately determine which politicians hold power and which policy directions get traction. If educators can bring parents and whānau – and the community – into a genuine, honest conversation about the purpose of education, about what their children actually need in an AI world, about the difference between a transactional model and one centred on human development, then something shifts. Parents and whānau who understand the stakes don’t just become allies in the classroom. They become a force in the wider conversation about what we want our system to be.
This is not about educators becoming political activists. It’s about recognising that the professional responsibility to advocate for children doesn’t stop at the classroom door. Sharing what you know – with parents, with your community, with anyone who will listen – is part of the job. And, as the panelists pointed out, the pre-mortem only works if the right people are in the room.
A Pre-Mortem, Not a Post-Mortem
One of the most powerful ideas in the webinar was the framing of Brookings’ research approach. Rather than wait five years for randomised control trials and then pick over what went wrong, they conducted a pre-mortem. This involves asks “What are the failure modes? What do we know now that we can act on before the damage is done?” We didn’t do that with social media. We can do it with AI.
Here are some questions worth pursuing further in your own context:
- If you looked at the assignments you set this term, how many of them could be completed – and would be completed – by a student using AI with minimal cognitive engagement? What would you do differently?
- Does your practice lean toward the transactional model that Rebecca described? If so, what would it look like to shift even one unit toward process, curiosity, and genuine engagement?
- Are your students in explore mode – motivated by the learning journey itself – or predominantly in passenger or achiever mode? What’s your role in that?
- What would an AI-aware, AI-assisted, and AI-resistant approach to your curriculum actually look like in practice?
- Who is missing from your school’s conversation about AI? Students? Families? What would it mean to genuinely bring them in?
These aren’t comfortable questions. They’re also not rhetorical. I’d genuinely like to know what you’re thinking and what you’re doing. Leave a comment, or get in touch directly.
What I’m Taking Away
The recording of this webinar is worth your time. Share it with a colleague. Use it as the basis for a staff conversation. The researchers are generous with their resources – KnowledgeWorks have published a large number of AI-related resources on their site, plus there’s the Brookings tip sheets for families and Rebecca Winthrop’s LinkedIn newsletter are all freely available.
But more than any resource, what I’m taking from this is a sense of urgency about not waiting. The pre-mortem is only useful if we actually act on it. The future Katherine Prince described – vibrant, personalised, genuinely human-centred learning – doesn’t arrive automatically. It has to be built, deliberately, by people like us.
What are you going to do this term?
Watch the webinar recording
References
- Charting a New Course for Education from KnowledgeWorks
- A New Direction for Students in an AI World: Prosper, Prepare, Protect from the Brookings Institution
- The Disengaged Teen, Winthrop and Anderson
If you enjoyed this conversation, you may be interested in the next Webinar event being hosted by KnowledgeWorks. On April 7, they’re partnering with Education Reimagined and a group of educators to explore educator roles within learner-centred education. Register here for “Educator Roles for Learner-Centered Transformation.”
