After Redefinition Comes Removal: Why AI Is Forcing Schools to Rethink the Work That No Longer Matters
I’ve spent most of my career talking about how technology can help transform learning and, more importantly, how it can transform learners. I’ve also spent the last two decades watching schools adopt new tools in hopes that something finally shifts. Sometimes it does. Most of the time it doesn’t.
But every once in a while, we hit an inflection point.
I remember one of those moments vividly. As a new teacher, I was part of the Classrooms for the Future Grant. Overnight, my classroom went from overhead projectors, chalk dust, and paperback novels (which we still love btw) to a Smartboard and a rolling laptop cart. It felt like I had walked into the future. Everything I did previously as a teacher suddenly seemed obsolete. I redesigned lessons. I changed my workflow. My students interacted with content differently.
And during that first year, I was introduced to the SAMR Model.
Substitution. Augmentation. Modification. Redefinition. I’ve always loved this graphic from Carl Hooker (and analogy).
I understood it immediately. Most of what I was doing with my shiny new tools fell somewhere between augmentation and modification. But then came those rare moments, the redefinition ones, where the lesson changed because the tool made something possible that simply wasn’t possible before.
When we did the Flat Classroom Project and my students met kids from 7 different countries, Skyped to plan and research, collaborated on Wikis, and co-created videos and blogs together…that was redefinition in action.
Those moments were magical.
Fast-forward twenty years.
We are living through another shift even larger than the “classrooms for the future” era. Artificial Intelligence isn’t just adding features to our work. It is altering the very nature of the work itself.
And yet, something feels incomplete.
Because this time, for the first time, it seems like there’s a level beyond Redefinition.
A level SAMR never anticipated: Elimination.
The Missing Level on SAMR: Elimination
If you have experimented with AI agents even briefly, you know how different this shift feels. This isn’t just about doing old tasks more efficiently. This isn’t even about redesigning tasks in more powerful ways.
It’s about tasks disappearing.
We’ve seen plenty of commentary about AI “personalizing learning,” “saving teachers time,” or “providing tutors for every kid.” That may be true. But the first and loudest impact isn’t coming from personalization or tutoring.
It’s coming from delegation.
Students can delegate unwanted work to an AI agent. Adults can do the same. And not in the hypothetical future, it’s right now.
I keep coming back to something Dan Meyer wrote about Khanmigo. He made the point that we’re evaluating AI as if the primary question is: How well does it teach? That’s not the real disruption.
The real disruption is that our learners, and our teachers, no longer need to “go to” an AI app at all. The AI can be everywhere, in everything, running behind the scenes.
Tools like Lindy AI or AI browsers like Atlas, Comet, or my favorite “Strawberry”, don’t just generate answers. They take action. They open tabs. They move files. They build slides. They make phone calls and send emails. They log into systems. They complete tasks. They automate workflows. They can even do your holiday shopping :)
For the first time, the “work” of school (the part that many learners see as busywork, compliance work, repetitive work) can be completely automated.
An AI agent can write an essay in your voice, respond like you in a video, complete the homework, solve the math steps, draft the lab report, annotate the reading, generate the slides, or create a summary of the historical chapter.
AI can do many of the tasks we assign.
So the question isn’t, “How will AI fit into school?”
The question is, “What happens to school when the assigned work no longer requires a human learner?”
That’s the essence of Elimination: removing tasks that no longer require human effort because the technology can complete them independently, and sometimes even better, faster, and more consistently.
AI Changes the Homework Equation First
Let’s get concrete.
Any work we assign that students do not want to do, do not have time to do, or do not see value in doing can be outsourced to an AI agent. Not “maybe.” Not “sometime soon.”
Now.
This means homework that is graded, compliance-driven, and primarily designed for practice is already compromised. Which I think you and I both already know.
We can’t pretend students won’t use these tools. They already are. Not in malicious ways. Not even in dishonest ways. They are simply doing what adults are doing: outsourcing tasks that don’t feel worth the effort.
This is the shift we are refusing to name.
Students aren’t trying to cheat (ok, maybe there are sometimes, but that didn’t start with AI). They are trying to manage. They are trying to keep up. They are trying to offload friction much in the same way we offload friction whenever we use AI to sort email, draft templates, or summarize meetings.
If outsourcing is easy, human nature will take over.
So the real issue isn’t integrity. The real issue is relevance.
If an assignment is easy to outsource, then we must ask why the learner needs to do it in the first place.
That’s the Elimination level of SAMR. Tasks that can be completed fully, invisibly, and instantly by an AI agent are tasks we must reconsider, or rethink how we do it in AI-resistant ways.
Teachers Are Living the Same Shift
Let’s be honest: educators are doing the same thing. And maybe they should in some circumstances.
AI can now do all of the following.
Generate unit plans
Draft rubrics
Provide feedback
Analyze student work
Enter grades
Create presentations
Prepare slides
Design assessments
Email parents
Build scaffolds
Rewrite curriculum
Summarize meetings
Draft reports
Personalize accommodations
Generate exemplars
And it’s getting better at these tasks daily.
If you had access to an AI agent that could give feedback, grade student work, and enter results directly into your gradebook, how many teachers would use that?
The answer isn’t hypothetical. We know.
Many would. Not all of course. But many already are.
And not because they are lazy. Because they are exhausted. Because they are overworked. Because the system has piled on tasks that are so far removed to the true work of teaching: mentoring, designing experiences, and helping learners grow.
Elimination is coming for teacher tasks as well. Not the human ones, but the compliance ones.
This is not a threat. It’s an opportunity.
But only if we design for it.
What Gets Eliminated in an AI-Saturated School?
If we follow the logic out, here’s what AI is likely to eliminate:
1. Low-engagement, high-output assignments
Worksheets. Formulaic essays. Chapter summaries. Packet-based units. These can already be completed by an AI agent without human involvement.
2. One-size-fits-all homework
Again, the outsourcing cost is too low. If it doesn’t matter to the learner, an AI agent can complete it faster than they can.
3. Administrative overload
Much of the teacher workload that contributes to burnout. The emails, incessant planning, feedback cycles, and documentation are already automatable.
4. Traditional long-form papers without process
Essays still matter. Writing still matters. Thinking still matters. But if we only assess the final product, the process can be done entirely by AI. That’s not what we want, and it defeats the purpose of assigning a long-form writing piece in the first place.
5. Compliance-driven learning
The kind where students complete tasks to get points, not purpose. AI makes compliance trivial. Which means compliance learning will collapse first.
6. Tasks that don’t build skills transferable to the real world
If the only purpose of the task is “school for school’s sake,” AI is going to expose it.
This can feel uncomfortable, but it’s actually clarifying. It forces a conversation we’ve avoided for years.
If a task can be completed entirely by AI, did it ever require a learner? Should it now require a learner. The answers will be debated, but that’s the point. Healthy conversation around what matters.
Redefinition Must Come Before Elimination
Before we talk about eliminating tasks, we need to talk about redefining them.
Redefinition was always the aspirational level of SAMR. It is the place where technology didn’t just make learning more efficient, but fundamentally different. The Google Doc allowed collaborative annotations and real-time writing. The laptop cart allowed multimodal projects. The internet allowed real-world connections.
AI, however, redefines differently.
It redefines by shifting the cognitive load.
An example might look like this. Let’s focus on feedback and assessment.
SUBSTITUTION
The scantron machine took the education world by storm, because it substituted the task of manually grading, and saved teachers time and mental energy. EZ-grader and Zipgrade did much of the same.
AUGMENTATION
In rolls, automated quizzes on Google Forms, Canvas, and other LMS tools. AI products like Frizzle allow teachers to take a picture of their student work and have it immediately grade it.
MODIFICATION
Now a student comes into math class, and hops on Snorkl for 5 minutes to answer a problem or two. The student uses the digital whiteboard to answer and show work. They received immediate feedback inside of Snorkl from the AI that sees their work in real-time. The teacher views a dashboard to see how all of the students are doing on these two questions. Within five minutes the student and teacher both have a really strong idea of where they are at, and what is the next best course of action.
REDEFINITION
Teacher goes to Goblins App to create their own AI clone. It’s kind of fun to create a Pixar looking and talking head, but if you aren’t into that thing, you can use one of their avatars and do the exact same thing.
Students log onto Goblins and take a diagnostic. The app now knows their foundational gaps. It works alongside the content and curriculum of the teacher, assigning work, giving feedback, and helping the teacher understand the best use for in-class time and explicit instruction.
This can all happen right now. So, how are you redefining the learning experience?
If Smartboards helped us present differently, and devices helped students create differently, AI helps students think differently…or more specifically, think at different parts of the problem-solving sequence.
AI can take over the rote parts: outlining, formatting, transcribing, drafting, sorting, summarizing, and generating. Which means learners can spend more time in meaning-making, analysis, evaluation, creativity, and reflection (if we design for that).
It also doesn’t mean we get rid of the in-person outlining, formatting, drafting, and summarizing. This should still be a focus, and build from the basic principles of learning. But, it should not be the end goal of what the learner does in a school.
This is where the opportunity lies.
If AI eliminates tasks that didn’t matter much, it can also elevate the ones that do.
But only if we redesign.
What Should Replace the Eliminated Work?
Three core shifts must happen.
1. From Products to Processes
If AI can replicate the product, the value must exist in the process. Teachers need visibility into the thinking, iteration, and decision-making. This requires checkpoints, conferences, drafts, revisions, and reflection. Not just uploading a final file.
This can happen in-person or with tools like Draftback and Snorkl.
2. From Compliance to Engagement
The assignments that survive AI are the ones learners actually want to do. Tasks with personal relevance, meaningful challenges, authentic audiences, and ownership opportunities cannot be outsourced.
Students don’t outsource work they love.
Of course they’ll still have to do some work they don’t love, but it should be as a stepping stone to do that meaningful and relevant task they are excited about.
3. From Isolated Tasks to Human-Centered Experiences
Collaboration, conversation, problem-solving, role-play, simulations, debates, fieldwork, mentorship, community partnerships — AI can support these, but it cannot replace the human element.
Meaningful learning is social. And social learning is very hard to automate.
The Real Purpose of School Must Shift
We often talk about preparing students for the future of work. But the future of work is shaped by the same forces shaping the future of school.
In both worlds, AI will do the following:
automate routine tasks
accelerate work cycles
eliminate redundant processes
elevate cognitive-rich tasks
reward adaptability, creativity, critical thinking, and problem-solving
value human skills over mechanical ones
So instead of preparing learners to complete tasks, we should prepare them to create, critique, collaborate, defend, and innovate.
Instead of preparing them for a world where they “do” work, we should prepare them for a world where they design work, direct work, and decide what’s worth doing.
Discernment has never been more important than it is right now.
This is the shift from compliance to agency.
From tasks to thinking.
From assignments to experiences.
From product to process.
From old SAMR to a new framework where Redefinition is no longer the ceiling, because Elimination is the floor beneath us.
The Moment We’re In
We are not in an AI future. We are in an AI present.
Students already have access to tools that can complete traditional school tasks with ease. Teachers already have access to tools that can automate large portions of their workload. Administrators already have access to tools that can manage systems and data.
We can either pretend this isn’t happening, or acknowledge the reality and redesign around what matters most: meaningful, relevant, human-centered learning.
The schools that thrive in the next decade will not be the ones that completely ban AI or cling to traditional assignments. They will be the ones that do the hard work of reimagining what learning is for.
This moment is not about control.
It is about clarity.
When the tasks that don’t matter fall away, the tasks that do matter become unavoidable.
The question for every school is simple:
What remains when the busywork disappears?
Call to Action for K–12 Educators and School Leaders
Start an Elimination Audit. This week.
Identify one assignment, workflow, or task that could be fully completed by an AI agent (student or teacher).
Ask why that task exists. If the only answer is “because we’ve always done it,” it’s a candidate for elimination.
Redesign what remains around process, purpose, and engagement.
Share your redesigned version with your team or department.
Make this a monthly or quarterly habit. One eliminated task per month = a fundamentally different school year.
AI will eliminate parts of school (and here is the important part) with or without us.
Our job is to make sure it eliminates the right parts.
If we do this well, what remains will finally be the learning that matters. If we don’t do it well, then we’ll be fighting for a learning experience that gets less and less human every year.
