Nice post, Nick. As I prepare to teach again, I'm coming back to this and some of the other substacks in your network that focus directly on instruction. Your point here about process, is no surprise given my commitments to pragmatism, right in line with how I approach teaching. The other aspect, which is implicit in your treatment of individual grades and products, is the social nature of learning. How do we carve out space and time for the social process of writing through peer review and group engagement. How do new AI tools help or hinder social activity in a classroom?
Thanks, Rob!!! I will hold onto your thought on the social nature of learning for a later post!!! Be well!!! I hope your planning is going well. Luckily, you have a long runway and a growing teacher network to lean into. Let us know how we can be of assistance as you dig into the work.
Great post Nick - it's so interesting to hear your learnings from the front line of education - it really adds substance to the commentary!
I fully agree with the feeling of wooziness with switching between multiple tools to get writing. Do you think that "in-built" co-pilots could help with this? Or does it conflict with some of the frameworks you describe?
Thanks, Zan. Definitely feels like the front line these days.
I am excited about the possibility of copilots. For my own workcycle and students. I am very interested to see how students respond to these systems. But to the point of my post, the larger focus needs to be shifting the prevailing focus on product so that these work spaces become less efficiency oriented. In a product-oriented context, copilot spaces are irresistible. In a process-oriented context, they are more instrumental towards particular kinds of workflows.
Nice. Lots of transferrable insights here. Mollick's Co-Intelligence implicitly endorses the continued need to immerse ourselves in process-oriented workflows to see the best outcomes with AI. Efficiency-driven workflows may be advantageous in the short-term, but in the long-term is a very risky endeavor for a whole host of reasons.
I've been working with AI a lot longer than most, as there have been AI built into marketing and copywriting tools, albeit rudimentary in comparison to the latest, for years.
Everyone has their own workflows, but I've landed in the same general zone as you speak to here. If you jump to AI too quickly, your own creativity will naturally atrophy. If you reply on it too much in the end, your abilities to recognize what is good work will, and you'll be less and less useful over time, and might as well just be replaced entirely with AI.
But if you use it where it's strengths shine, especially in brainstorming, grammar checking and what I call 'prototyping', in conjunction with you own work at the start, collaborating with it as of it's a sort of assistant/intern (or 'minion', as I enjoy saying), then combine it coherently yourself and use it to help with spit, polish, grammar and fact checking, that's roughly the way to go. And its key to leave space in there for ideas to germinate.
Yes, my blog is attempting to navigate some real complex terrain. Am I talking about high school students, college students, or professionals? I think all these audiences bump up against product-oriented outcomes and thus lean into AI for efficiency, but the effects are more devastating for persons who have not yet consolidated their core writing competencies. My goal in the next many posts will be creating a culture within the classroom where AI can become a tool again, but doing so means changing many of the foundational truths of the traditional writing classroom: how we teach, assign, and assess writing. As to my own writing practice, I plan on doing a post sometime this summer about my "Fill Up the Bucket" Principle. I find when working with AI that it scales up its editing and polishing functionalities when you offer it increasingly sophisticated inputs. To me, it feels like working through the first couple questions on a computerized standardized test. Once you get the AI humming at a certain level, it can become a substantial writing partner. I know others have pointed this out, but I wanted to document this with my own experiences.
"but the effects are more devastating for persons who have not yet consolidated their core writing competencies."
And never will if AI is used too early in the process, and - maybe as importantly - they won't be able to recognize good or bad writing enough to even use the tools well.
I try not to over-optimize prompts right away, and work through the subject matter like more of a conversation with a research assistant. There's a term for this, but I forget what it is.
You nailed it: A lot of discourse around AI (not only in the classroom) seems to be on the outcome that AI tools help you achieve - faster, more efficiently, etc.
But I very much prefer the process-focused approach. That's what my "Minimum Viable Prompt" post some months ago was all about - encouraging people to embrace the process, enjoy the trial-and-error, and the learning-by-doing that comes with it.
I see how this approach lends itself very well to the educational priorities in modern classrooms.
Peer engagement in a writing community is one of the twin engines that makes a conventional classroom effective. But any sort of “prompt” is counterproductive because it assumes the teacher owns the key that starts the engine. Teaching writing well means providing a full bodied environment that assumes the writer is the owner of the process, working through Murray’s magic for a reason only the writer knows.
I like this new focus a lot, and I'm looking forward to seeing where you take this.
It seems to me like a process-oriented approach is huge! Not only will it give students insight into the process they are using -- which I assume is key in encouraging students to be autonomous -- but I could also imagine it would help students with meta-cognition -- empowering students to think about their own thinking. And that, to me, seems like such a great gift to give. It might be the key to being a self-motivated life-long learner.
I do wonder however, how struggling students will go with a process-oriented approach. I can imagine the average and above average students will benefit from this approach, but I do wonder how below average students will do. I have nothing to support these concerns, I'm just wondering. Is there much research on this?
The short answer is yes. The research on basic writers is deep—what you are calling “below average writers.” Two points. Thinking of writers as “average” or above or below is treacherous. The concept crystallizes and reduces an individual writer as a point on a standardized curve, and it disguises an emphasis on product. Writers who resist and resign need responsive instruction as individuals, not group instruction for summative assessment, even more than fluent, confident, motivated writers.
Nice post, Nick. As I prepare to teach again, I'm coming back to this and some of the other substacks in your network that focus directly on instruction. Your point here about process, is no surprise given my commitments to pragmatism, right in line with how I approach teaching. The other aspect, which is implicit in your treatment of individual grades and products, is the social nature of learning. How do we carve out space and time for the social process of writing through peer review and group engagement. How do new AI tools help or hinder social activity in a classroom?
Thanks, Rob!!! I will hold onto your thought on the social nature of learning for a later post!!! Be well!!! I hope your planning is going well. Luckily, you have a long runway and a growing teacher network to lean into. Let us know how we can be of assistance as you dig into the work.
Great post Nick - it's so interesting to hear your learnings from the front line of education - it really adds substance to the commentary!
I fully agree with the feeling of wooziness with switching between multiple tools to get writing. Do you think that "in-built" co-pilots could help with this? Or does it conflict with some of the frameworks you describe?
Thanks, Zan. Definitely feels like the front line these days.
I am excited about the possibility of copilots. For my own workcycle and students. I am very interested to see how students respond to these systems. But to the point of my post, the larger focus needs to be shifting the prevailing focus on product so that these work spaces become less efficiency oriented. In a product-oriented context, copilot spaces are irresistible. In a process-oriented context, they are more instrumental towards particular kinds of workflows.
I see - that's a helpful clarification - thank you!
Fantastic. This breed of in-depth exploration is precisely what is needed, not just in education, but in virtually every industry.
Nice. Lots of transferrable insights here. Mollick's Co-Intelligence implicitly endorses the continued need to immerse ourselves in process-oriented workflows to see the best outcomes with AI. Efficiency-driven workflows may be advantageous in the short-term, but in the long-term is a very risky endeavor for a whole host of reasons.
I've been working with AI a lot longer than most, as there have been AI built into marketing and copywriting tools, albeit rudimentary in comparison to the latest, for years.
Everyone has their own workflows, but I've landed in the same general zone as you speak to here. If you jump to AI too quickly, your own creativity will naturally atrophy. If you reply on it too much in the end, your abilities to recognize what is good work will, and you'll be less and less useful over time, and might as well just be replaced entirely with AI.
But if you use it where it's strengths shine, especially in brainstorming, grammar checking and what I call 'prototyping', in conjunction with you own work at the start, collaborating with it as of it's a sort of assistant/intern (or 'minion', as I enjoy saying), then combine it coherently yourself and use it to help with spit, polish, grammar and fact checking, that's roughly the way to go. And its key to leave space in there for ideas to germinate.
Yes, my blog is attempting to navigate some real complex terrain. Am I talking about high school students, college students, or professionals? I think all these audiences bump up against product-oriented outcomes and thus lean into AI for efficiency, but the effects are more devastating for persons who have not yet consolidated their core writing competencies. My goal in the next many posts will be creating a culture within the classroom where AI can become a tool again, but doing so means changing many of the foundational truths of the traditional writing classroom: how we teach, assign, and assess writing. As to my own writing practice, I plan on doing a post sometime this summer about my "Fill Up the Bucket" Principle. I find when working with AI that it scales up its editing and polishing functionalities when you offer it increasingly sophisticated inputs. To me, it feels like working through the first couple questions on a computerized standardized test. Once you get the AI humming at a certain level, it can become a substantial writing partner. I know others have pointed this out, but I wanted to document this with my own experiences.
Preach.
"but the effects are more devastating for persons who have not yet consolidated their core writing competencies."
And never will if AI is used too early in the process, and - maybe as importantly - they won't be able to recognize good or bad writing enough to even use the tools well.
I try not to over-optimize prompts right away, and work through the subject matter like more of a conversation with a research assistant. There's a term for this, but I forget what it is.
Yes!
You nailed it: A lot of discourse around AI (not only in the classroom) seems to be on the outcome that AI tools help you achieve - faster, more efficiently, etc.
But I very much prefer the process-focused approach. That's what my "Minimum Viable Prompt" post some months ago was all about - encouraging people to embrace the process, enjoy the trial-and-error, and the learning-by-doing that comes with it.
I see how this approach lends itself very well to the educational priorities in modern classrooms.
Thanks, Daniel!! I appreciate your support!!!
Peer engagement in a writing community is one of the twin engines that makes a conventional classroom effective. But any sort of “prompt” is counterproductive because it assumes the teacher owns the key that starts the engine. Teaching writing well means providing a full bodied environment that assumes the writer is the owner of the process, working through Murray’s magic for a reason only the writer knows.
Found this super insightful
Thanks, Finn!!! It is good to hear from you!!!
I like this new focus a lot, and I'm looking forward to seeing where you take this.
It seems to me like a process-oriented approach is huge! Not only will it give students insight into the process they are using -- which I assume is key in encouraging students to be autonomous -- but I could also imagine it would help students with meta-cognition -- empowering students to think about their own thinking. And that, to me, seems like such a great gift to give. It might be the key to being a self-motivated life-long learner.
I do wonder however, how struggling students will go with a process-oriented approach. I can imagine the average and above average students will benefit from this approach, but I do wonder how below average students will do. I have nothing to support these concerns, I'm just wondering. Is there much research on this?
The short answer is yes. The research on basic writers is deep—what you are calling “below average writers.” Two points. Thinking of writers as “average” or above or below is treacherous. The concept crystallizes and reduces an individual writer as a point on a standardized curve, and it disguises an emphasis on product. Writers who resist and resign need responsive instruction as individuals, not group instruction for summative assessment, even more than fluent, confident, motivated writers.
Ah, yes, good points. Old ways of thinking are hard to break
It's interesting how powerful AI can be when it's limitations are best understood.