M.027 The reflection machine
Friction with a purpose in higher education
You are reading Molekyl, finally unfinished thinking on strategy, creativity and technology. Subscribe here to get new posts in your inbox.
This week, NHH had the pleasure of welcoming the Minister of Research and Higher Education, Sigrun Aasland. As part of her visit, the two of us were asked to tell her how we work with AI at NHH.
On the surface, it was a simple ask with a clear delivery. We could have just walked her through what we have done with AI lately: AI in NHH’s strategy process, AI to tutor large workshops with hundreds of BSc students, AI to sharpen the feedback on written assignments, AI as innovation coaches and co-founders in executive workshops, AI as an organization that must be lead, and more.
Below the surface, the ask set something else in motion. It made us want to not only show what we do and how we do it, but also to explain “the why” behind these efforts.
So we took the simple ask and complicated it. From giving a 30-minute talk, to designing an AI-fuelled learning experience for the visit, building on the very same principles we use in our own teaching.
We kicked it off a couple of days before the visit, when we invited the minister to a 30-minute pre-meet about AI in education. The twist was that she didn’t meet with us, or anyone else from NHH. She met a bespoke AI-agent we built for the occasion. An AI instructed to nudge her to reflect on the questions we intended to discuss in our talk, and to help us better understand her thinking as we designed it.
The aim was to turn a showcase of the what and the how into a demonstration of the deeper why. Why we use AI this way. Why we think it works. And why we think AI can be the light at the end of the tunnel for higher education, and not just an approaching train.
There is only so much one can fit into a 30-minute talk, so we wrote this post to flesh out our thinking more fully. Which makes the post itself part of that same learning experience.
So let’s begin where the talk began, with the same question we put to the minister: what do we actually want our young adults to walk away with after their years at a place like NHH?
What are we educating for?
One answer to this question is that NHH and other higher education institutions want our students to learn something that will help them in their later life and careers. In doing so, questions play a key role.
Education is all about questions. Some questions are related to facts, and “the what”. What is an interest rate, or what is a large language model? Other questions are related to skills and processes, or “the how”. How do central banks set the interest rate, or how to train a large language model? Yet other questions are related to the deeper understanding behind the “what” and the “how”, namely the “why”-questions. Why do interest rates fluctuate, or why is a large language model able to solve problems its builders never designed for it to solve?
The higher up the ladder we climb in the education system, the more important the final type is. The “why”-question. As opposed to knowing a “what” or a “how”, “the why” represents the deeper wisdom that lets us carry a lesson from one area into a completely new one and still make good judgment. It’s where deep understanding is made actionable. Where our graduates leave university with a toolbox that stretches beyond the actual questions, contexts and cases they actually studied in school.
Well, in theory, that is. In practice, it’s far from straightforward. Information we can get from a book. Know-how we can get from imitation and deliberate practice. But wisdom? That is trickier.
Education for wisdom
There are many competing definitions of wisdom floating around, but we like to think of it as the following simple equation:
wisdom = experience × reflection on that experience
Crude, yes, but more than sufficient to highlight some of the distinct features of wisdom. One is that experience is an essential part of wisdom accumulation, meaning that wisdom isn’t just something you get from knowing a lot. Another is that reflecting on that experience is what ultimately determines how much wisdom you get out of it. A third is that wisdom is the product of the two terms, not the sum. Implying that if either term is zero, so is the acquired wisdom.
While experiences are abundant in life, reflection is in relatively shorter supply. It’s the cognitively demanding process of thinking of a set of “why”-questions related to an experience. Reflection can happen naturally, and it can be deliberately designed and curated. Both processes will lead to higher wisdom.
When reflection happens naturally, it’s often because an experience has some sort of friction baked into it. When something is hard, unpleasant, or surprising, our System 2 tends to kick in and our brain naturally starts to ask why. If a book is hard to read, we wonder why we’re bothering. If something is scary, we ask why it’s worth the risk. If a situation is uncomfortable, we ask why we think so.
In comparison, we are less likely to naturally reflect on frictionless experiences. We can leave the house in the morning, take the everyday route, and arrive at the office half an hour later without being able to tell what happened in between. Hence, we tend to acquire more wisdom from the tough lessons of life, than the easy ones.
Deliberate reflection, on the other hand, is often associated with learning situations. Socrates and his reflective questions is perhaps the most famous example, but humans have lived this out in practice for much, much longer. In the old days when the elder, the parent or the master sat with their kids or apprentices and asked them the tough questions about an experience, or in recent times when Navy SEALs have targeted debriefs after operations or exercises. Why didn’t it work? Why did we do that and not that? Why did it end up this way?
Wisdom is the product of experience and reflection on that experience, and friction is the concept that binds them. It can sit in the experience itself, or it can sit in the reflection as a tough why-question.
If we look to higher education in light of this, we can reframe its main role to deliberately add friction to help students acquire wisdom.
For centuries the model for how we did so was Socratic, and focused on adding friction to the reflection term. One master with one or a few apprentices, practicing active, individual tutoring of the why-questions behind whatever you studied.
Then, as demand for highly educated people grew, we traded the one-to-a-few model of individual, reflective tutoring, for a one-to-many model where education was standardized for scale. With this change, the deliberately added friction to the reflection term fell. But the new model also introduced new friction, this time in the education experience itself.
In any mass education system, students have to figure many things out on their own. What to read, how to write, what to prioritize, how to work out what that strange professor actually meant. Some of that friction of the modern higher education experience is designed, like assignments, exams or mandatory classes. But much of it also isn’t. By a lucky stroke, friction in the experience ended up carrying much of the wisdom-building load that designed reflection in one-to-few tutoring used to carry.
And it worked. For a long time.
Until AI came along.
The answering machine
As noted in this earlier post, AI is a friction-reduction machine, driving the cost of answers towards zero.
In many areas of life and society, this is a godsend. We can find “the what” faster than ever, and get easier access to “the how”. Democratising access to information and knowledge. Increasing efficiency of knowledge work. Letting AI do the mundane tasks, while the humans can focus on the higher level tasks at the very top of the knowledge hierarchy.
In education, the same technology is a potential disaster, as it’s not the answers that matter. It’s the processes of arriving at them. In higher education, friction-ridden questions, assignments and work make up the very ladder that allows students to climb all the way to wisdom accumulation. The frictionless elevator might seem nice, but it doesn’t reach the floor we actually want students to visit.
Take the friction out of the education experience, and we are in danger of destroying the oh-so-important reflection that naturally grows out of a friction-ridden experience. The result is experience times zero. Little or no wisdom, no matter how much experience we pile into the study years.
What’s left then is shadow thinking and shadow learning. Our students work with AI, feel productive, get more done than ever before, while quietly losing the ability to dance at the top.
This is the story where AI is the final nail in the coffin of higher education as a place that builds wisdom.
But there is also another story. One where the light in the tunnel isn’t an approaching train, but a new and promising path. And, surprisingly, that path is carved by the very same AI.
The questioning machine
Most discussions about saving higher education from AI are about adding friction back into the experience. Make classes analog and mandatory, kill the home exams and add more oral examinations and seminars. And so forth.
We have done several such efforts ourselves, and they will probably work. At least to some extent.
But adding friction back to the education experience is only one half of what we can do. The other half, which we discuss comparably less, is to explore what we can do at the reflection side of the equation.
If we do, we see that the thing that threatens the whole field, might be the same thing that can save it. AI.
Flip the switch on the answering machine, the one driving friction towards zero everywhere, and we can turn it into a deliberately designed questioning machine. One that hands back the most powerful tool in the education toolbox: individual reflection. Cheaper, and at greater scale, than ever before.
Flipping the logic is, however, the easy part. We can do that with a prompt. Making it actually work is harder. That takes deliberate design, by people who understand how learning happens in a given domain.
But this being hard is not an argument against trying. It’s the argument for giving it a shot.
AI, out of the box, wants easy. Wisdom and education doesn’t. They were never about easy. The friction in getting this right is exactly the friction that builds the wisdom we’ll need to design the higher education of tomorrow.
And it’s worth our best attempt. Because with AI we can bring back the very thing we gave up when we scaled higher education. The one-to-one, reflective tutoring. For everyone. With as much friction as a given learning purpose needs.
AI is a cheating machine. But it can also be a reflection machine. The difference between the two is deliberate design. Which is where some hope sneaks back in for us educators. Done right, AI does not remove us. It moves us. From delivering friction by accident, through the clunky readings and the exam dread, to designing it on purpose.
It’s also what every one of our AI-experiments has in common, including the pre-meet AI we built for the minister. Different on the surface, but built on the same intellectual engine of deliberate, carefully designed Socratic friction, aimed at someone’s reflection.
Experience times reflection
Which brings us back to the minister.
We took our own medicine, and tried to make her visit an education experience too. Designing it on the same principles we use for learning experiences at NHH, and for leadership workshops elsewhere.
We deliberately added friction to both terms of the wisdom equation, for her to leave NHH a little wiser than when she arrived.
Did it go according to plan?
That is a question we will leave to the minister herself to reflect on.
 | A guest post by
|