Yes, you raise good questions. In many ways, we have few examples of the kind of work I am trying to characterize. My instances are placeholders that can and should be picked apart by other skillful instructors like yourself. I also am very partial to the intellectual labor of an essay. Multimodality is perhaps a way of preserving or repositioning that labor in a meaningful way in these AI altered circumstances. So much of this deeper work involves building trust and a shared vision with our students. Following their interests to spaces of genuine skill and knowledge development. The path is definitely not clear.
The fact that LLMs can produce essays on demand, answer questions based on a reading, summarize, generate authoritative-sounding research summaries, etc.doesn’t mean that these tasks are not useful for learning. They are in fact very likely to lead to learning when completed by a human. The “deep learning” field trips sound wonderful and appealing, but there are many ways in which they are unlikely to lead to actual learning. Also, spell check and calculators can and do indeed lead to poor spelling and lack of automaticity with math facts, both of which impede the higher order thinking and learning that we all want.
This is true, Miriam. Similarly, cars have resulted in humans walking less and therefore getting less exercise, but I don't see many bike-to-work campaigns gaining swells of adopters. Similarly, Socrates is credited with decrying the technology of writing because it made his students lazy since they no longer had to remember the things they wrote down. Yet we now consider unlettered persons "ignorant" no matter how glorious their memories might be. Technology giveth and technology taketh away.
How we hold onto the essay? Why are we holding onto it so tightly? What is the cognitive work that students engage in when writing essays? That is more important than the form—the long-standing form—the edifice around which we have built our practice. The sooner we recognize that an essay isn’t an end, but a means, we will start to make a productive pivot. If the essay remains, our foundational assumption must be that students will interact with AI while completing such work. How can the form respond to this condition on the ground? Perhaps you have some ideas.
It sounds like you agree that the process of writing an essay (the means to the end) is useful, but you are unsure of what the cognitive work is that makes it useful? I would say that writing helps us transfer information from working memory to long term memory, elaborate on concepts, and connect them to prior knowledge. This helps develop and reinforce understanding. It’s a powerful tool for generating, clarifying and solidifying our thoughts. Importantly, the process can also reveal gaps in our knowledge or misconceptions, and this can promote learning. I don’t know what can replace it, and I don’t have any brilliant ideas for how to hold on to it beyond trying to incorporate more in-class, non AI assisted writing and discussion. This will obviously look different in different grades and contexts.
Miriam, you describe writing an essay more as a tool of reading comprehension than a writing experience. What about writing for an audience? Writing with an intention to do something in the world? Writing to evoke an aesthetic experience? I agree with you. Writing an essay can do all those things you mention to build knowledge as we do when we take in information and organize it in long term memory. But that isn't the work of the writer. Is that really writing instruction?
Great read here. I’m looking for ways to implement what you’re suggesting in my college (online!) writing classes (1101 Composition and 1102 Comp 2 - writing about lit) and in large in-person lit survey classes.
The pattern in software is identical, the breakthrough isn't restricting the tool, it's designing work around what humans uniquely bring.
Your Graduate Profile captures what I want for my own three kids: judgment, context, real thinking. Not just task completion.
The Industrial Revolution example is perfect. When my engineers use AI for syntax while focusing on architecture and real constraints, they become the engineers they always wanted to be. Same principle here.
This matters so much because our kids deserve to think deeply, not just efficiently. Education is where we either get this right or lose a generation to optimization without purpose.
While there's a lot here I'm on board with, the examples highlight why redesign is much harder than folks like to present. While of course we want students to think about, build, and experience the future, we need to share examples of redesign that don't perform a bait-and-switch with targets. Consider the two tasks:
1. Write a short essay answering: "Was the Industrial Revolution more helpful or harmful?"
2. Build a multimodal exhibit with artifacts, sound, and image; Host a "Factory Futures" gallery walk with invited guests; Reflect on how AI could both revive and threaten blue-collar industries.
These tasks don't develop or assess the same learning targets. They might not even belong in the same course. Do we want to understand students' current knowledge and understanding of historical events or do we want to see their thoughts about the future? Both are worthwhile, and the latter isn't inherently more valuable than the former. We still have writing standards, essays continue to be a viable medium (e.g. this Substack), and it remains worthwhile for students to learn and think critically about the past.
Perhaps we shouldn't stop with writing an essay, and students should also build something multimodal, host a gallery walk, and connect their learning to the future. In that case, the hard thing about the redesign isn't "the designing": it's identifying what from our current courses we are de-implementing. Which pieces of the past do we not share with students? Which concepts in science do we strategically under-develop? I'm all for narrowing the curriculum, but de-implementation is a big task. We want to make sure that what is cut isn't just what had been the last unit of a scope and sequence. That process is arduous, challenging, political, and necessary.
Fantastic Nick. This is an invitation. Don't bury your head in the sand or ban. Evolve! I'm interested in the next post too. If ChatGPT can do it. Try it first. You might be surprised at the output and how it su[per charges your creativity as an educator.
Love this perspective. AI isn’t the problem, it’s the low-level tasks that don’t require real thinking. When we design learning that’s messy, engaging, and meaningful, AI can actually make it better instead of replacing it. Thanks for putting this into words so clearly, schools need to hear this right now!!!
Redisgn is the fundamental question and a big ask for educators, many of whom have no interest in engaging with AI in their classrooms in the first place. Are you familiar with the RAIL framework? We did their certification program this spring and the video which really stuck with me is the necessity for "re-imagination" as opposed to "integration" when it comes to assessments. Everyone is talking about integrating AI into the current methods which teachers have used for decades and I agree that I don't think it's going to be sustainable. How do you respond to folks who aren't convinced (or adamantly opposed) to bringing AI directly into their classrooms? Even your example adds a layer of setting up a field trip (which I understand could be designed around) but adds a further burden on teachers that anyone who has organized one can attest to. The bottom line is traditional assignments and modes of assessment are simply easier. Maybe some teachers are going to have a recognition that kids will use AI and they can try to get them not to, but don't want to actually spend class time having kids use the models. I think the ongoing conflict, and it will be critical to gather data, will be the difference between teachers selectively using AI for certain kinds of tasks and skill-building successfully, and those who refuse. It will be a bifurcated landscape compounded by the fact that it will also be uneven across public and private schools, demographics, resources, etc... What may work in one environment may simply not be feasible in another. Complex to be sure.
Wise response as usual. I know how you feel about the integration word. Yes, this is hard work, and teacher bandwidth for this design is limited. Teachers have too many other balls up in the era. The fuel for this redesign needs to emerge through system wide intention. That process as you say will be experimental. Districts with resources will push ahead in this area, as with past moments of redesign and transformation. I will look into the RAIL framework. Sounds intriguing.
Your words, Stephen: The bottom line is traditional assignments and modes of assessment are simply easier. Are you saying teachers want to take the easy way out?
I was struck by the literal meaning of your words. It’s possible that you meant easier in that it doesn’t take so much time and time is such a critical resource for teachers. Or it could mean that teachers generally don’t want to work hard. Easy/hard is a subjective gradient. I wasn’t quite sure what you are saying
Two things. First, not all teachers are convinced that redesign is even necessary, not by a long shot. For the audience on Substack maybe, but hardly a majority. Second, the difference between integrate and redesign is huge. Integrate is possible but if we just do the same thing over and over but with AI - well you know that old definition of insanity. Actual, real redesign - rethinking the entire purpose and goals of learning - is something very, very difficult for most teachers. I was in charge of curriculum development for several years and just getting teachers to think about UbD or TfU or other frameworks for backward design was almost impossible. I know it’s 1 school in a very specific kind of model, but I’ve been around enough teachers and schools to know how hard it is to change your practice. It’s easier because it’s familiar. That’s my point in defending / empathizing with teachers even though I’ve tried to move away from those models for years (with some success). We’ll see if AI moves the needle when so much else has not. To be done at scale requires institutional support and while it’s starting, I’m not convinced it’s going to end well. I wish I was more optimistic.
I hear you. Here’s the problem. Historically, teachers settled into a career are incredibly resistant to change. This is why when a principal gets to open a new school, they are very very picky about who they hire. When I worked in a curriculum office during the Whole Language period in a district with 50,000 students (massive changes on a quick timeline), the curriculum office wrote off older teachers, keeping their eyes on long calendar (five year plan) and betting on pressure from younger teachers as they gained experience. Ultimately, the resistors gained enough political clout that Whole Language went away. The difference is that those who are convinced that redesign isn’t necessary face a whole new reality. AI is like 5000 X’s bigger than Whole Language and it isn’t going away. The longer they hold out, the worse its gets for everybody. This really isn’t a matter of preference except in elite schools like yours. It could explode public schools with more blast than vouchers. Poor and minority children are going to be the victims of these resistors who are not convinced in public schools. Education is a local issue in the main, and most teachers aren’t thinking beyond their school and their job. I hear lots of what you are describing here from individual teachers—how unfair it is to teachers, how they are doing fine as they are, how distasteful they find AI as a cheating machine. The only aspect of this that is even more morally problematic is the response of administrators, who have largely turned their backs on young people as a strategy to save their own worlds. Your concerns are legitimate. Under normal circumstances teachers would be able to keep on keepin’ on as they have for the past 75 years. These aren’t normal circumstances. This is an existential threat to public schools exacerbated by the conservative ideology in the White House pushing vouchers and destroying the Department of Education. Truth be told, I am sick at heart that I’m going to go to my grave watching the system that lifted me up out of an Appalachian-style childhood and gave me the chance to get a PhD. I’m totally helpless to contribute beyond flapping my lips on Substack. My voice on this Substack is not one I enjoy using.
This is so so important for us as educators to learn and pick from. In redesigning instruction, to spark more critical thinking and creativity among students and if that could be achieved by students alongside AI that's okay.
Nice overview, Nick. I'm emerging from my first draft of a plan to teach this fall, including teaching first-semester undergraduates for the first time in over a decade. It is forcing me to think about the transition from HS to college in a way I have not had to. This gives me some good vocabulary and concepts to talk with them about their experience before college and expectations about their education.
I also made the transition from teaching graduate courses to teaching first-semester freshmen just before I retired. I loved it, Rob. It's soooo different!!! So much hope and innocence. It takes a while for them to get jaded.
Yes, you raise good questions. In many ways, we have few examples of the kind of work I am trying to characterize. My instances are placeholders that can and should be picked apart by other skillful instructors like yourself. I also am very partial to the intellectual labor of an essay. Multimodality is perhaps a way of preserving or repositioning that labor in a meaningful way in these AI altered circumstances. So much of this deeper work involves building trust and a shared vision with our students. Following their interests to spaces of genuine skill and knowledge development. The path is definitely not clear.
The fact that LLMs can produce essays on demand, answer questions based on a reading, summarize, generate authoritative-sounding research summaries, etc.doesn’t mean that these tasks are not useful for learning. They are in fact very likely to lead to learning when completed by a human. The “deep learning” field trips sound wonderful and appealing, but there are many ways in which they are unlikely to lead to actual learning. Also, spell check and calculators can and do indeed lead to poor spelling and lack of automaticity with math facts, both of which impede the higher order thinking and learning that we all want.
This is true, Miriam. Similarly, cars have resulted in humans walking less and therefore getting less exercise, but I don't see many bike-to-work campaigns gaining swells of adopters. Similarly, Socrates is credited with decrying the technology of writing because it made his students lazy since they no longer had to remember the things they wrote down. Yet we now consider unlettered persons "ignorant" no matter how glorious their memories might be. Technology giveth and technology taketh away.
How we hold onto the essay? Why are we holding onto it so tightly? What is the cognitive work that students engage in when writing essays? That is more important than the form—the long-standing form—the edifice around which we have built our practice. The sooner we recognize that an essay isn’t an end, but a means, we will start to make a productive pivot. If the essay remains, our foundational assumption must be that students will interact with AI while completing such work. How can the form respond to this condition on the ground? Perhaps you have some ideas.
It sounds like you agree that the process of writing an essay (the means to the end) is useful, but you are unsure of what the cognitive work is that makes it useful? I would say that writing helps us transfer information from working memory to long term memory, elaborate on concepts, and connect them to prior knowledge. This helps develop and reinforce understanding. It’s a powerful tool for generating, clarifying and solidifying our thoughts. Importantly, the process can also reveal gaps in our knowledge or misconceptions, and this can promote learning. I don’t know what can replace it, and I don’t have any brilliant ideas for how to hold on to it beyond trying to incorporate more in-class, non AI assisted writing and discussion. This will obviously look different in different grades and contexts.
Miriam, you describe writing an essay more as a tool of reading comprehension than a writing experience. What about writing for an audience? Writing with an intention to do something in the world? Writing to evoke an aesthetic experience? I agree with you. Writing an essay can do all those things you mention to build knowledge as we do when we take in information and organize it in long term memory. But that isn't the work of the writer. Is that really writing instruction?
Great read here. I’m looking for ways to implement what you’re suggesting in my college (online!) writing classes (1101 Composition and 1102 Comp 2 - writing about lit) and in large in-person lit survey classes.
Yes for this!
The pattern in software is identical, the breakthrough isn't restricting the tool, it's designing work around what humans uniquely bring.
Your Graduate Profile captures what I want for my own three kids: judgment, context, real thinking. Not just task completion.
The Industrial Revolution example is perfect. When my engineers use AI for syntax while focusing on architecture and real constraints, they become the engineers they always wanted to be. Same principle here.
This matters so much because our kids deserve to think deeply, not just efficiently. Education is where we either get this right or lose a generation to optimization without purpose.
Thank you for fighting this fight.
What worries me that if we (faculty) don't take control of the redesign, someone else will. That could end very badly.
While there's a lot here I'm on board with, the examples highlight why redesign is much harder than folks like to present. While of course we want students to think about, build, and experience the future, we need to share examples of redesign that don't perform a bait-and-switch with targets. Consider the two tasks:
1. Write a short essay answering: "Was the Industrial Revolution more helpful or harmful?"
2. Build a multimodal exhibit with artifacts, sound, and image; Host a "Factory Futures" gallery walk with invited guests; Reflect on how AI could both revive and threaten blue-collar industries.
These tasks don't develop or assess the same learning targets. They might not even belong in the same course. Do we want to understand students' current knowledge and understanding of historical events or do we want to see their thoughts about the future? Both are worthwhile, and the latter isn't inherently more valuable than the former. We still have writing standards, essays continue to be a viable medium (e.g. this Substack), and it remains worthwhile for students to learn and think critically about the past.
Perhaps we shouldn't stop with writing an essay, and students should also build something multimodal, host a gallery walk, and connect their learning to the future. In that case, the hard thing about the redesign isn't "the designing": it's identifying what from our current courses we are de-implementing. Which pieces of the past do we not share with students? Which concepts in science do we strategically under-develop? I'm all for narrowing the curriculum, but de-implementation is a big task. We want to make sure that what is cut isn't just what had been the last unit of a scope and sequence. That process is arduous, challenging, political, and necessary.
Fantastic Nick. This is an invitation. Don't bury your head in the sand or ban. Evolve! I'm interested in the next post too. If ChatGPT can do it. Try it first. You might be surprised at the output and how it su[per charges your creativity as an educator.
Yes! If only one improvement comes out of AI, I hope it's driving necessary improvements to education to make it valuable instead of busy work.
Love this perspective. AI isn’t the problem, it’s the low-level tasks that don’t require real thinking. When we design learning that’s messy, engaging, and meaningful, AI can actually make it better instead of replacing it. Thanks for putting this into words so clearly, schools need to hear this right now!!!
Redisgn is the fundamental question and a big ask for educators, many of whom have no interest in engaging with AI in their classrooms in the first place. Are you familiar with the RAIL framework? We did their certification program this spring and the video which really stuck with me is the necessity for "re-imagination" as opposed to "integration" when it comes to assessments. Everyone is talking about integrating AI into the current methods which teachers have used for decades and I agree that I don't think it's going to be sustainable. How do you respond to folks who aren't convinced (or adamantly opposed) to bringing AI directly into their classrooms? Even your example adds a layer of setting up a field trip (which I understand could be designed around) but adds a further burden on teachers that anyone who has organized one can attest to. The bottom line is traditional assignments and modes of assessment are simply easier. Maybe some teachers are going to have a recognition that kids will use AI and they can try to get them not to, but don't want to actually spend class time having kids use the models. I think the ongoing conflict, and it will be critical to gather data, will be the difference between teachers selectively using AI for certain kinds of tasks and skill-building successfully, and those who refuse. It will be a bifurcated landscape compounded by the fact that it will also be uneven across public and private schools, demographics, resources, etc... What may work in one environment may simply not be feasible in another. Complex to be sure.
Wise response as usual. I know how you feel about the integration word. Yes, this is hard work, and teacher bandwidth for this design is limited. Teachers have too many other balls up in the era. The fuel for this redesign needs to emerge through system wide intention. That process as you say will be experimental. Districts with resources will push ahead in this area, as with past moments of redesign and transformation. I will look into the RAIL framework. Sounds intriguing.
Thanks, Nick. I just learned we have a mutual friend - I sent the info in a DM.
Your words, Stephen: The bottom line is traditional assignments and modes of assessment are simply easier. Are you saying teachers want to take the easy way out?
I’m assuming you’re being deliberately rhetorical here, right?
I was struck by the literal meaning of your words. It’s possible that you meant easier in that it doesn’t take so much time and time is such a critical resource for teachers. Or it could mean that teachers generally don’t want to work hard. Easy/hard is a subjective gradient. I wasn’t quite sure what you are saying
Two things. First, not all teachers are convinced that redesign is even necessary, not by a long shot. For the audience on Substack maybe, but hardly a majority. Second, the difference between integrate and redesign is huge. Integrate is possible but if we just do the same thing over and over but with AI - well you know that old definition of insanity. Actual, real redesign - rethinking the entire purpose and goals of learning - is something very, very difficult for most teachers. I was in charge of curriculum development for several years and just getting teachers to think about UbD or TfU or other frameworks for backward design was almost impossible. I know it’s 1 school in a very specific kind of model, but I’ve been around enough teachers and schools to know how hard it is to change your practice. It’s easier because it’s familiar. That’s my point in defending / empathizing with teachers even though I’ve tried to move away from those models for years (with some success). We’ll see if AI moves the needle when so much else has not. To be done at scale requires institutional support and while it’s starting, I’m not convinced it’s going to end well. I wish I was more optimistic.
I hear you. Here’s the problem. Historically, teachers settled into a career are incredibly resistant to change. This is why when a principal gets to open a new school, they are very very picky about who they hire. When I worked in a curriculum office during the Whole Language period in a district with 50,000 students (massive changes on a quick timeline), the curriculum office wrote off older teachers, keeping their eyes on long calendar (five year plan) and betting on pressure from younger teachers as they gained experience. Ultimately, the resistors gained enough political clout that Whole Language went away. The difference is that those who are convinced that redesign isn’t necessary face a whole new reality. AI is like 5000 X’s bigger than Whole Language and it isn’t going away. The longer they hold out, the worse its gets for everybody. This really isn’t a matter of preference except in elite schools like yours. It could explode public schools with more blast than vouchers. Poor and minority children are going to be the victims of these resistors who are not convinced in public schools. Education is a local issue in the main, and most teachers aren’t thinking beyond their school and their job. I hear lots of what you are describing here from individual teachers—how unfair it is to teachers, how they are doing fine as they are, how distasteful they find AI as a cheating machine. The only aspect of this that is even more morally problematic is the response of administrators, who have largely turned their backs on young people as a strategy to save their own worlds. Your concerns are legitimate. Under normal circumstances teachers would be able to keep on keepin’ on as they have for the past 75 years. These aren’t normal circumstances. This is an existential threat to public schools exacerbated by the conservative ideology in the White House pushing vouchers and destroying the Department of Education. Truth be told, I am sick at heart that I’m going to go to my grave watching the system that lifted me up out of an Appalachian-style childhood and gave me the chance to get a PhD. I’m totally helpless to contribute beyond flapping my lips on Substack. My voice on this Substack is not one I enjoy using.
This is so so important for us as educators to learn and pick from. In redesigning instruction, to spark more critical thinking and creativity among students and if that could be achieved by students alongside AI that's okay.
Nice overview, Nick. I'm emerging from my first draft of a plan to teach this fall, including teaching first-semester undergraduates for the first time in over a decade. It is forcing me to think about the transition from HS to college in a way I have not had to. This gives me some good vocabulary and concepts to talk with them about their experience before college and expectations about their education.
I also made the transition from teaching graduate courses to teaching first-semester freshmen just before I retired. I loved it, Rob. It's soooo different!!! So much hope and innocence. It takes a while for them to get jaded.
It will be good if in a post a task is thought the way it is explained. It will be helpful.
“to design learning that asks students to think more deeply and act more purposefully”
This was an excellent article!