The 1 Reason Every School Should Teach Systems Thinking
AI will not replace thinkers, but it will expose who never learned to think.
Back in the early 1970s, a team at MIT built a computer model to capture how the world behaves over time.
They called that model World3. It pulled together five big things we normally learn in separate subjects- population, food production, industrial output, pollution, and non‑renewable resources- into one connected system.
Donella Meadows, then a young systems scientist and the lead author, explored how those forces pushed and pulled on each other over roughly two centuries of simulated history. The book that emerged from this work, The Limits to Growth, sold millions of copies and sparked a debate that still hasn’t really stopped.
A lot of that debate has revolved around one big question: whether its long‑term projections actually got the future right.
By contrast, far fewer people have asked what might be the more useful question: how that team learned to see the world as one connected system, and why kids still rarely get taught to think that way at school.
That’s the heart of what I’m arguing in this week’s edition. My point is simple: schools should teach systems thinking and give it the same weight as any core skill. Systems thinking prepares young people to make sense of a deeply connected world, where the big problems play out through relationships, feedback loops, and knock‑on effects that no single fact can ever explain on its own. Take AI: it’s made systems thinking much more urgent because a child who learns to see systems now is far better equipped for a future where much routine ‘thinking’ gets handed off to machines.
Facts in isolation are not understanding.
A lot of schools are very good at teaching the pieces: a fact here, a formula there, and things to memorise for a test. That approach works fine when a problem stays put and fits neatly inside one subject. But it starts to fall apart as soon as the problem won’t sit still and spills over into other areas.
Back in 1973, Horst Rittel and Melvin Webber gave these kinds of problems a name. They called them “wicked” because they’re fuzzy at the edges, don’t have a clear finish line, and can’t be solved with a simple true‑or‑false answer. Derek and Laura Cabrera push this further, arguing that wicked problems often show up when there’s a gap between how a real system works and how we think it works. We end up grabbing the wrong fix simply because we’ve misread the system we’re trying to change. A finance chief solves it as a money problem. A technology chief solves it as a tech problem. But the real issue often lies in the space between those perspectives, and no one feels responsible for it.
Take a story many economists and policy writers love. In a popular anecdote from British‑ruled India, the government wanted fewer cobras in Delhi and offered a reward for every dead snake. In the tale, people start breeding cobras just to cash in on the reward. When the officials shut down the scheme, the story says the breeders dumped their now‑worthless snakes, and Delhi ended up with more cobras than it had started with. It’s a neat little warning about how a quick fix can end up feeding the very problem it was supposed to solve.
The twist is the part that matters. The breeding part very likely never happened. German economist Horst Siebert popularised the term “cobra effect” in 2001, and the story itself seems to trace back to a hedged 19th‑century newspaper piece that naturalists at the time thought was unlikely. We took a messy, uncertain bit of history and squeezed it into one clean cause, because a simple story is easier to remember than a real system. That storytelling habit is exactly the kind of thing systems thinking teaches you to notice and challenge.
Climate change is both a chemistry problem and an economics problem, with each side feeding into the other over decades. Inequality works the same way, and so does a pandemic or a supply chain that breaks on the far side of the planet. These kinds of problems live in the relationships between the pieces, not in any one piece on its own. A child who’s only taught to break things into parts will often miss what’s really going on in the way those parts connect.
Systems thinking is the habit of looking at those connections first. It gets you asking what influences what, where the delays sit, and what’s likely to happen after that obvious first move. That uses a very different mental muscle from memorising facts, and many schools don’t give that muscle nearly as much of a workout as they could.
This can be taught, and taught early.
The pushback I hear most often is this: systems thinking sounds like something for professors and consultants, not something you’d try with a bunch of nine‑year‑olds.
It’s a fair concern, but there are decades of classroom practice that show kids can handle it. The Waters Foundation has been helping schools teach systems thinking to children since the late ’80s. In the Catalina Foothills district in Arizona, US, systems thinking programs have been run in classrooms since around 1989. In Waters Foundation examples from Tucson, kindergarteners chart how story characters such as the gingerbread man change over time, learning to see change unfold – that’s one way into systems thinking. In the US, the National Research Council’s Framework for K–12 Science Education, under the National Academies, now treats “systems and system models” as core ideas in school science.
This is not only a US story. In 2022, the European Commission published GreenComp, its official sustainability competence framework, and named systems thinking as one of 12 core competences every learner should build. A 2024 European Commission study mapped 39 countries and found that elements of systems thinking are already present in most national curricula. UNESCO goes further, listing systems thinking among the 8 core competences it says every learner should develop. The policy backing is there. The gap is in how evenly schools actually teach it.
None of this needs the jargon. If you give a young child simple language for a loop, you’re handing them a tool they can use for the rest of their life. They start to notice deeper causes instead of just pointing to the closest symptom. They start asking what a decision will do a few steps down the line, not just right now.
We already teach reading and basic arithmetic early on, because people need them almost everywhere in life. Understanding how things connect belongs in that same ‘must‑have’ group, but only some school systems genuinely treat it that way.
AI has raised the stakes faster than most school systems can realistically keep up.
For most of the history of schooling, systems thinking has been treated as a nice‑to‑have, hard to squeeze into the timetable, something for the “next” reform round. AI has pretty much killed off that luxury of taking our time.
An AI model is a system embedded within larger systems: the classroom, the platform, the economy, and the rules that society wraps around it. A young person who just leans on AI risks turning into a passive consumer of whatever it spits out, even the slick, confident answer that happens to be wrong. A young person who can see the system around the tool can do far more with it. They start asking where the answer came from, what’s missing, and who actually gets affected if we act on it. They can check the reasoning instead of just swallowing a plausible‑sounding output whole.
That gap, between passive use and active questioning, is what will end up mattering most. AI is very good at fast, pattern-heavy tasks. The sort of routine cognitive work many jobs used to depend on. What AI still can’t do is decide what really matters, weigh the ugly trade‑offs, or take responsibility for the fallout. Those are still very human skills, and they’re exactly the kind of muscles that systems thinking helps you build. As AI takes over more of the routine work, being able to reason about the whole system stops being a nice extra and becomes what separates the people who steer from the people who get steered.
None of this has to wait for a national curriculum overhaul, and chances are, you don’t run one anyway. What you probably run is something closer to the ground: a team, a department, a hiring plan. The same logic applies just as much in those settings. Instead of only hiring the deepest technical specialist, start valuing the person who can see how all the parts connect. Give more credit to the answer that follows a problem several steps back, rather than the tidy single cause that fits nicely on one slide or the one with the most detail on one piece. You might not be able to rewrite the syllabus, but you can decide what your organisation treats as real intelligence.
So we come back to Donella Meadows and the model she made famous.
For over 50 years, we have argued about whether those curves were accurate. We’ve spent far less of that time teaching the mindset that produced them; the ability to hold a whole system in view and think through how it moves. The next generation is going to inherit more complexity, coming at them faster, with more powerful tools in their hands than any generation before. We can keep throwing facts at them and hope they somehow assemble the bigger picture on their own. Or we can choose to teach them, early and deliberately, how to see the whole system they’re operating in.
One path prepares them for the world they’re really going to live in; the other mostly prepares them for another test.
I write The Foresight-Driven Enterprise for leaders who are building what comes next. If that sounds like you, subscribe, and you’ll get a free copy of Working with Complex Systems, my ebook on systems thinking models worth knowing, including Donella Meadows’ own