When Struggle Gets Outsourced
How AI is challenging math instruction, and why that might be a good thing.
The problem is not that students are using AI. The problem is we’re giving them work that AI is better designed to complete than they are.
Earlier this summer, I had the chance to keynote a math educator conference in New Hampshire. During the Q&A, a teacher asked a question that has stuck with me ever since:
That hit me. Because it’s true. But if the work coming in is AI-generated, copy-pasted, or filtered through a black box, then what exactly are we evaluating and discussing?
🎯 If grades are getting in the way, rethink them
Once students leave the classroom, we lose control over what tools they use. What we can control is the purpose behind the work we assign, and how we assess it.
If grades are nudging students toward speed and correctness rather than curiosity and reflection, then maybe grades are the problem and not AI.
Let’s assess process, not just product.
Let’s reward questions, not just answers.
Let’s make it clear that what matters most isn’t just getting it right, but thinking problems through and sharing personal journeys to solutions.
⚖️ The promise and peril of “help”
Students using AI, finding answers and applying solvers isn’t new or revolutionary. When I was in high school, fellow students took my homework out my locker and copied it. Photomath, Symbolab, and now ChatGPT have simply scaled that behavior, and added an “explanatory” interface to it.
These tools may look like help, but they risk flattening the learning experience. When students bypass the messy process, comparing strategies, defending choices, making mistakes, and revising their thinking, we lose the opportunity to see what students truly understand.
When the struggle is skipped, often the learning is too.
🛠️ Rethinking the role of struggle
The better question isn’t:
“How do we stop students from using AI?”
It’s:
“How do we make the struggle feel worth it?”
Productive struggle isn’t about making math harder. It’s about making the challenge meaningful and helping students build confidence in navigating the unknown, an essential part of mathematical thinking.
As George Pólya outlined, problem solving isn’t a single-step task, it’s a full cycle of sense-making:
Understand the problem → Devise a plan → Carry out the plan → Reflect and revise
This process isn’t just about getting the answer, it’s about building understanding.
It breaks down when the assignment is “finish this worksheet.” Struggle happens when we say:
Be curious.
Be messy.
Make your thinking visible, even if it’s incomplete (in many instances incomplete may be better).
🔁 What If we shifted the assignment?
NCTM’s 2024 position statement on AI challenges us to redesign mathematics, to create learning environments that are humanizing, inclusive, and intellectually rich.
That starts with rethinking what we ask students to do.
Instead of eight problems to get through, what if the homework looked like this?
Try at least two problems (of your choosing) and write down every strategy you considered, even the ones you didn’t use.
Pick one problem and represent it in two different ways.
Solve one problem wrong on purpose and explain the error.
Get stuck on a problem, and write down three “I wonder…” questions.
This kind of work might look different, but it feels more human and more personal.
And that’s exactly what AI can’t replicate.
🤖 It’s about the prompt.
The issue is that students are being given work that encourages them to use AI instead of thinking for themselves.
Our goal can’t just be “complete this task.”
It must be:
“Reveal your thinking.”
“Struggle productively.”
“Make connections.”
“Ask questions.”
The best tech platforms, like Desmos Classroom and Polypad, don’t try to automate learning.
They let students explore, represent, and revise.
They give teachers access to thinking in progress.
They invite discussion, not just evaluation.
They build connections to concepts and procedures.
🧭 Keep It Personal
Learning is personal. It’s rooted in curiosity, struggle, identity, and reflection. It happens not just in the answers students give, but in the questions they ask, the choices they make, and the connections they build along the way.
AI is here to stay. Some students will use it wisely; others will misuse it like any other technology, but the disruption it creates also brings an invitation:
To make our teaching more personal.
To shift from answer-getting to idea-building.
To help students not just solve the problem, but solve their problem, their way, and reflect on what they learned in the process.
AI can simulate solutions. It can generate steps, summarize strategies, and even mimic explanations. But it can’t feel confusion. It can’t wrestle with an idea. It can’t light up with understanding.
Because learning isn’t just about getting it right. It’s about making it yours.
✉️ Want to keep thinking together? Drop a comment, share this post, or connect with me on LinkedIn. Let's build math spaces where struggle isn't outsourced—it’s celebrated.