Should This Task Be AI-Resistant or AI-Compatible?
A very unscientific but wildly useful guide for teachers, coaches, and school leaders.
The Real Question No One Is Asking
The AI debate in schools usually sounds like this:
“Should we ban AI?”
“Should we teach kids how to use AI?”
“Is AI cheating?”
All fine questions.
But they miss the most important one:
Should this task be AI-resistant… or AI-compatible?
Because the truth is complicated. Some learning gets better with AI. Some learning gets worse with AI. And some learning should politely tell AI to wait in the hallway :)
Let’s make this easier (and more fun). First, a brief overview on the reasons for AI-Resistant and AI-Compatible.
Step 1: Meet the Two Characters in Your Classrooms
AI-Resistant Tasks
These are the tasks where the struggle is the learning.
If AI does the work, the learning collapses like a Jenga tower.
AI-resistant tasks usually involve:
Sense-making
Personal voice
Judgment and discernment
Physical or social interaction
Productive struggle
Examples:
Writing a personal narrative about a moment that changed you
Debating a controversial issue live with peers
Designing and testing a prototype
Solving a novel problem with no obvious path
Explaining why an answer works, not just what the answer is
If a student says:
“Wait… can I use MagicSchool/SchoolAI or Gemini/Copilot for this?”
And your instinctive reaction is:
“Absolutely not and here’s why”
Congrats. You’ve found an AI-resistant task. But, that is not every task is it…
AI-Compatible Tasks
These are tasks where AI acts like a power tool (think using a screwdriver putting together a desk or bed, vs using a drill).
The learning actually accelerates.
AI-compatible tasks usually involve:
Iteration
Feedback
Research synthesis
Planning and outlining
Generating multiple options
Examples:
Brainstorming essay structures
Getting feedback on clarity or organization
Generating practice problems
Summarizing background research
Drafting multiple versions of an email, proposal, or lesson plan
If a student says, “Can I use AI to help me think this through?”
And your reaction is, “Yes (safely) but show me how you used it”.
That’s AI-compatible learning done right.
Step 2: The One Question That Decides Everything
Before assigning anything, we should ask, “If AI does the thinking here, what human thinking disappears?”
If the answer is essential thinking → go AI-resistant.
If the answer is busywork, formatting, or non-essential thinking → go AI-compatible.
This question works for classrooms, leadership work, professional learning, maybe even coaching and parenting.
Step 3: This Isn’t About Control
Here’s the uncomfortable truth: AI didn’t break school assignments. It exposed the ones that were already fragile.
This doesn’t mean they aren’t necessary or have value. It means we cannot live in a fairy-land where we act like AI isn’t available. So, since it is widely available, we need to be INTENTIONAL about task design that should be AI-resistant, just as we are when we want to use AI with a PURPOSE.
AI doesn’t force us to choose ban or embrace. It forces us to choose better design and true learning purpose.
Step 4: A Simple Decision Matrix (Steal This)
Make tasks AI-resistant when the goal is one of the following:
Building identity, voice, or values
Learning how to think, not what to think
Developing judgment and discernment
Strengthening collaboration or communication
Make tasks AI-compatible when the goal is one of the following:
Improving quality through feedback
Increasing speed so thinking can go deeper
Supporting access for diverse learners
Practicing real-world workflows
The key knowing when to switch. And, how to switch effectively.
AI-Powered or Human-Directed? Maybe both…
The best schools won’t be the ones that ban AI entirely or let it do everything.
They’ll be the ones that can confidently say:
“Here’s when we use AI.
Here’s when we don’t.
And here’s why.”
That’s not necessarily an AI skill. But it is the most important skill of our times right now.
That’s discernment.
