What happens when our students enter a world where AI can solve the problems we’ve spent years teaching them to solve? The real challenge won’t be finding the right answer. It will be navigating messy problems filled with competing priorities, incomplete information, unintended consequences, and human needs that don’t fit neatly into a formula.
Students Exploring Messy Problems
I’m a huge fan of student-led sessions in conferences because we often get the nuance, honesty, and contextual elements that are sometimes missing from broader sessions on big topics. So, I’m sitting in a circle of chairs in a session on work-based learning. This is one of the rare conferences where I am not leading multiple sessions. Just a keynote and then a chance to learn from a diverse group of educators (or in this case, students).
At one point, we break into small groups and one of the students leads us in a scenario-based activity. I don’t remember it verbatim but the gist is this: you are working a construction project and the bags are really heavy. A few hours into the first day, you are feeling exhausted and sore. You’re not sure if you should talk to fellow workers or your supervisor. You want to make a good first impression but you also need to advocate for yourself. What should you do?
I love this problem because it doesn’t have a simple solution. Each potential answer actually leads to multiple potential consequences. It’s nuanced. This might be a chance to challenge unethical working conditions but also there’s a need to learn the trade and work within the system where you are essentially powerless.
Each scenario is like this. One deals with co-workers giving you a nickname that you don’t appreciate but it’s also part of the company culture. Another deals with someone who is immature and perhaps even reckless asking for a letter of recommendation.
As each small group discusses their potential solutions, I am struck by the difference between these solutions and many of the problems students solve in school. Here, the answer is not immediate. It requires nuance and complexity. It taps into human interactions and could potentially require contextual understanding and empathy.
What Problems Are We Solving?
I remember sitting in a high school math class trying to solve a problem involving two trains leaving different stations at different speeds. At what point would they meet? I may have been a bit neurospicy, but not the kind who loves trains. I simply didn’t care about the trains. I didn’t know anyone on the trains. I had no idea why they were leaving at different times or why it was my job to figure out when they would meet.
Years later, as a teacher, I remember my students working on a problem where someone was buying an ungodly amount of watermelon based on how many watermelons their neighbor had purchased. Why was this person buying so much watermelon? What kind of fruit-based party would require fourteen watermelons? Why did their neighbor’s watermelon purchases matter? Who eats that much watermelon? And what are the unintended consequences of that much sugar and fiber?
Both problems had a clear process and a single correct answer. And while students certainly need to learn the mathematical concepts involved, I wonder what happens when we reduce problem-solving to a series of disconnected scenarios where the goal is simply to find the right answer.
It had me thinking about word problems in general. We often think students disengage because of language. So we front-load the vocabulary. We teach decoding techniques. Which is fine. They have a place. But . . . what if the bigger issue is with the problem itself.
![]()
We hate stories that go nowhere. It’s a bit like watching Twin Peaks. I know there’s a general mystery but it’s so abstract I can’t follow it. They hate stories that don’t operate the way our world actually operates. This is why I hate Hallmark Movies. They’re predictable and formulaic and they don’t actually tie into the way conflict and complexity play out in our world.
Meanwhile, I actually think students love solving real problems. The type of problems that reflect authentic contexts and invite us to delve into nuance and complexity. And these are the exact kinds of problems our students will need to solve in a world of smart machines.
Callout: wicked problems
Our Students Will Need to Solve “Wicked Problems”
The term wicked problems emerged in the field of urban planning in the 1960s and was later formalized by design theorists Horst Rittel and Melvin Webber in 1973. They used the term to describe complex social challenges that couldn’t be solved through traditional, linear problem-solving approaches. These were the nuanced, complex problems that required deeper empathy and contextual understanding.
Unlike a traditional math problem with a single correct answer, wicked problems are difficult to define and even harder to solve. They often involve competing priorities, incomplete information, interconnected systems, and unintended consequences. A solution that works for one group might create new problems for another. In fact, attempting to solve one part of a wicked problem can actually change the nature of the problem itself. These challenges require more than simply applying a formula or following a set of steps. They require us to navigate uncertainty, consider multiple perspectives, and make decisions when no solution is perfect.
Currently, I am watching my son solve these types of wicked problems in his engineering program. With so many students worried about passing vector calculus or making it through organic chemistry, they are suddenly solving these wicked problems and it’s turning out to be just as challenging as the gate-keeping classes they passed in the previous semesters.
In a world of generative AI, the ability to navigate wicked problems becomes even more important. AI can generate answers, analyze information, identify patterns, and suggest potential solutions in seconds. But generating a solution isn’t the same as understanding a problem. AI might offer a technically accurate answer without fully accounting for the human context of the situation. AI can’t read the room and thus can’t understand the competing values that collide in a given scenario.
But this is all the more reason our students will need to solve these complex, wicked problems.
Ten Characteristics of Messy Problems and the Capacities They Develop
When I talk to teachers, school leaders, students, industry leaders, and parental figures, I often hear a theme emerge. Students need to be problem solvers. They need to know how to find the problems, pursue the solutions, and ultimately handle the challenge they face when a solution has unintended consequences.
This is why I love messy problems. These are the learning experiences that lead to the core competencies that we want students to develop. Messy problems look different depending on the context, but they often share certain characteristics. And it’s through these problems that students develop the habits and mindsets they need as problem solvers. So, let’s look at the 10 characteristics of messy problems.
1. Complexity
Messy problems are built on complexity. They require students to look at multiple factors or variables that can lead to multiple solutions. Here, the problems push students to look at multiple perspectives that make them difficult to understand or solve through simple explanations.
In science, students might explore how to restore a local river ecosystem. They need to consider water quality, invasive species, agricultural runoff, wildlife habitats, and the needs of nearby communities. Improving one part of the ecosystem might create challenges elsewhere. Note that it doesn’t have to be a huge project-based learning activity. It could be a shorter 3-5 day deep dive.
In social studies, students might examine how their local city should respond to a housing shortage. They consider population growth, zoning laws, and housing affordability. They could interview competing stakeholders around environmental impacts and urban growth. Each potential solution involves competing interests and trade-offs.
In math, students develop a plan to redesign a school bus system to reduce transportation costs while maintaining reasonable travel times. They analyze distances, student populations, bus capacities, fuel costs, and scheduling constraints. Changing one route can affect the efficiency of the entire system.
2. Multiple Solutions with Competing Priorities
This sense of complexity ultimately leads to solutions that are equally complex and, at times, even fraught. Messy problems rarely have one perfect solution. Instead, they require weighing trade-offs and making decisions when different goals or priorities are in tension.
In the river restoration project, students might discover that reducing agricultural runoff could improve water quality but also place financial burdens on local farmers. It might also mean they have to reduce fishing in the area or limit tourism.
In the housing project, increasing affordable housing might require higher-density development that creates concerns about traffic, infrastructure, and neighborhood character. Affordable housing is critical but people also care about noise pollution and traffic.
In the school bus project, reducing transportation costs might mean longer commutes for some students or earlier pickup times for families. In each case, students have to move beyond simply finding a solution and begin asking which solution makes the most sense, for whom, and at what cost.
3. Unintended Consequences.
So, if students start with complex problems that lead to complex solutions, the inevitable result is a set of unintended consequences. Solving one part of a problem can create new challenges elsewhere. Even well-intentioned solutions can lead to unexpected outcomes.
We see this play out in our world all the time. In England, efforts to reduce ambulance response times led some services to manipulate how calls were classified or response times were recorded. In Mexico City, an effort to reduce air pollution by limiting when people could drive actually led some residents to purchase additional cars, including older vehicles that created even more pollution. In Sri Lanka, a ban on chemical fertilizers designed to promote sustainable agriculture led to declining crop yields and increased food insecurity. In the United States, an additive designed to make gasoline burn more cleanly ended up contaminating groundwater. And in Yellowstone, efforts to eliminate wolves to protect livestock contributed to ecological changes that affected everything from elk populations to vegetation and stream habitats.
In each case, the original solution made sense within a particular set of priorities. But the people implementing these solutions didn’t fully anticipate how their decisions might affect other parts of the system. This is why students need to develop foresight. They need to move beyond asking, “Will this solution work?” and begin asking, “What else might happen if it does?”
As educators, we can tie these examples to our content area and have students discuss them in open-ended Socratic Seminars
We can design problem-based simulations where students make these decisions and then deal with the consequences as a result.
4. Incomplete or Conflicting Information
When I think of well-crafted math problems, we often strive to make sure students can access the prior knowledge. We define terms and front-load vocabulary. We try to avoid adding extraneous information because it can feel unfair. Then we tie the math problem precisely back to the current unit or lesson plan.
These are all excellent strategies. And yet, problems rarely work that way in our world. We are often stuck with incomplete information. At times it might even be contradictory. We have to toggle between multiple strategies that connect to our own prior learning.
Here’s where it helps to craft problems that encourage students to examine whether they have all the facts and ask questions that might lead to new facts. It helps to tie in learning from multiple units where they have to combine strategies or processes that were previously siloed.
Imagine a middle school math class where students are tasked with determining whether their school should install solar panels. They receive information about the cost of installation, average monthly electricity bills, and estimates of potential energy savings. But some of the information is missing. They don’t know how much electricity the school uses during different seasons or how long the panels will last. To make matters more complicated, two different companies provide conflicting estimates about the potential savings.
Students have to figure out what information they still need, what questions to ask, and which sources they can trust. Along the way, they draw on percentages, ratios, equations, and data analysis from previous units. Instead of simply plugging numbers into a formula, they have to determine which mathematical strategies make sense, recognize the limitations of their data, and ultimately make a recommendation based on the best information available.
5. Interconnected Systems
As we think about the unintended consequences, a core reason for this is that the complexity of the problem and the competing interests involved, are all part of a larger reality of interconnected systems. Messy problems rarely exist in isolation. They are part of larger systems where changes in one area can affect other people, processes, or outcomes.
It’s like changing one bus route in a city. You might shorten the commute for one neighborhood while creating overcrowding at another stop, disrupting transfers, and making it harder for people across town to get to work on time. Or perhaps it’s more like trying to add an image to a Word Document in 2001. Sure, you solve the issue. You now have a picture in the middle of an essay. But along the way, you’ve completely screwed up the formatting. Somehow the text is flying off the pages, way outside the margin. The font is now 24 px. And the tabs are three inches into the page.
As teachers, we can design systems-rich problems that force students to engage in systems thinking. Imagine a social studies simulation where students explore what would happen if their city transitioned entirely to self-driving electric cars. At first glance, it sounds like a great idea. Fewer accidents, less pollution, and more efficient transportation. But what happens to human autonomy when people no longer control their own vehicles? What happens to the thousands of people who make a living driving? How do cities maintain their roads when gas tax revenue disappears? What happens to public transportation, car ownership, or the way we design our neighborhoods? Suddenly, what seemed like a straightforward technological solution becomes a much larger conversation about economics, government, infrastructure, and the ways we live together.
6. Context-Dependent Challenges
The Home Depot built a massive business in the United States around the idea that people want to tackle home improvement projects themselves. So, when the company expanded into China, it brought along the same do-it-yourself model that had worked so well in America. But here’s the thing. China isn’t the US and many Chinese consumers preferred hiring professionals to complete renovations (partly because labor was more affordable). By 2012, Home Depot had closed its remaining big-box stores in China. This is a small reminder that solutions that work in one context often fail miserably in another.
Messy problems tie directly to a specific context. They require students to empathize with a community and understand the unique challenges of a specific location in a specific time. After all, The Home Depot and Lowe’s have both tapped into an American desire for personal DIY projects (something deeply cultural) mixed with a reality of larger homes with expensive labor costs (or economic context). Those simply weren’t the case in China.
This type of context-dependent problem is the type that AI struggles to tackle. Generative AI can’t read the room. It offers predictive solutions based on general knowledge. But increasingly, our students will need to solve messy problems that are distinctly context-dependent. This is part of why I love design thinking projects. They push students to develop empathy with a specific group as they engage in contextual thinking.
7. Competing Human Needs and Perspectives
Where I live in Oregon, many schools are facing huge challenges of declining enrollment mixed with budget shortfalls. Their solutions range from closing schools to laying off teachers to increasing class sizes. We all agree on the main problem but the challenge is in how a solution impacts various stakeholders.
Messy problems invite us to make sense out of the competing human needs and perspectives. Different people experience the same problem differently. They have different needs and values. They have different lived experiences (often tied to their age) which shape their priorities. These create conflict, making it difficult to find solutions that work for everyone.
Our students can problem-solve in a way that helps them see how their solutions will impact different types of stakeholders in different way. It might be a shared social studies and math problem about urban growth boundaries and rising housing prices. It could be a science question about conservation and farming. Whether it’s a major project or a short problem-based sprint, these moments helps students see how various stakeholder needs often conflict.
8. Creative Constraints or Limitations
One of the most counterintuitive ideas I explored in a recent blog post is how constraints often lead to better problem-solving. This is a key component of divergent thinking. It’s the notion of thinking outside the box by repurposing the box.
Messy problems deliberately incorporate these creative constraints. They often involve limited time, resources, money, or available options. During our short Maker Monday activities, I used to pull Constraint Cards (you must take out one item, you need to add one item, you can only collaborate in silence, etc.) in a way that pushed students to think creatively.
9. Problems That Cross Disciplines and Domains
Messy problems don’t fit neatly into a single subject or field. They require drawing on knowledge, skills, and perspectives from multiple disciplines. This cross-pollination of content areas and disciplines ends up sparking better divergent thinking and ultimately more innovative solutions.
As educators, we can craft interdisciplinary PBL units that ask students to think like historians while also engaging in statistical thinking. We can tie together the content from science with history.
10. Problems with Ambiguity
Messy problems are not always clearly defined. Sometimes the greatest challenge is identifying the actual problem, determining what matters most, and deciding where to begin. What initially appears to be the problem might actually be a symptom of something much larger. Students need opportunities to ask questions, challenge assumptions, and explore different ways of framing the problem before jumping into solutions. This requires a willingness to sit with uncertainty and recognize that defining the problem is often just as important as solving it.
Who Do Our Students Become as Problem-Solvers?
Ultimately, messy problems help our students develop key capacities that they will need in a complicated and constantly changing world. In other words, we design messy problems with key characteristics. But those characteristics become the conditions that help students develop the competencies students need.
start here
| Characteristics of Messy Problems | Capacities Students Develop |
|---|---|
| Complexity | Nuanced thinking: recognizing ambiguity and resisting oversimplification |
| Multiple solutions with competing priorities | Wisdom: weighing competing priorities and making sound judgments when no solution is perfect |
| Unintended consequences | Foresight: anticipating potential outcomes and ripple effects |
| Incomplete or conflicting information | Intellectual humility: recognizing the limits of one’s knowledge and remaining open to revision |
| Interconnected systems | Systems thinking: understanding relationships, dependencies, and feedback loops |
| Context-dependent challenges | Contextual understanding: recognizing how circumstances shape problems and solutions |
| Competing human needs and perspectives | Empathy: understanding the experiences and needs of those affected |
| Creative constraints or limitations | Divergent thinking: generating multiple possibilities |
| Problems that cross disciplines and domains | Connective thinking: combining ideas and knowledge from different fields |
| Problems with ambiguity | Strategic framing: the ability to define the core problem and determine the appropriate approach |
stop here
How Do We Get There?
All of this sounds great in theory, but what does it actually look like in the classroom? How do we design learning experiences where students wrestle with messy problems while still meeting our standards and working within the constraints of our curriculum? The good news is that we don’t have to completely overhaul our teaching to make this happen. Next week, I’ll share specific strategies for incorporating messy problems into the classroom, including problem-based learning, scenario explorations, Socratic Seminars, design thinking projects, and work-based learning.