This makes a ton of sense. I’m in the middle of creating an online course for screenwriters using AI and have come to the realization that — if I want them to preserve their creative voice — they need to constantly improve their process and THEN use AI (product). You’ve stated this very very well.
Brilliant push back on the 'process over product' silver bullet. We can't let 'process' become just another compliance checkbox. Your point about emergence in creativity is essential. Thanks for the nuance!
That '6 assignments under a trench coat' line is so spot on. With 20+ years in the ELA classroom, I’ve been deep in thought lately about how we need to reshape the writing process we’re teaching to better fit this new reality.
I just finished Eric Hudson’s series on AI skills (I'm late finding the series. He posted it almost a year ago), and his take on 'playfulness' and the 'extended mind' feel like such a helpful bridge here. They offer a way to move toward a more creative, expanded partnership with these tools rather than just turning thinking into another compliance checkbox. Thank you for this piece, which has further pushed my thinking.
I don’t really have an answer for agents yet, and I’m not sure anyone does.
For me, the hope is to design better learning experiences.
And I will say that shifting to process hasn’t ballooned my work at all, especially with the shift to student agency and alternative assessment a bit more. I just spread out the work differently.
Nice piece, Jason. This feels important because the tendency you identify of treating process as a step-by-step formula and opposing it to product is, as Owen on LinkedIn says, merely recreating the problem. That's especially true if you're using a product to structure and track the completion of the process.
The point is to collapse the distinction and embrace the ungraded messiness that occurs when you turn a group of students on to a problem with no clear outcome or assurance of success.
That "third step" you outline is the key. It cuts against the systems we have set up that expect replication and control. The problem is not so much theoretical as our habits of mind about what we expect students to do. The older students get, the more their habits of mind are a factor. My own hot take is that students are the biggest barrier to adopting active learning approaches because it feels so alien and strange, and they have become addicted to being ranked and measured.
I can see the process model being appropriate in some contexts, but at some point, we assume the student knows the process. If we're still wanting to see how the student is making sausage in a 400-level course, we clearly are into surveillance. For a 100-level class, it makes sense. "This is how you do things in this field...."
As a doctoral student, my first class was a process class. For a 15-20 page research paper we had to submit an initial 20-source bibliography, then an annotated bibliography, then an outline, then the final product. IN other words, 4 total products. Fine. Still felt a bit elementary for a doctoral program, to be honest. But now, while those scaffolded steps are still available for most major papers, (if the student desires feedback) the only graded piece is the final product.
Interesting. I abandoned product long ago in favour of process. In recent years the first question I ask students is to outline their writing process: the answers are not surprisingly fairly basic but it lays down a benchmark before we work through the steps. I have a blog from 2014 describing my thoughts at that time, which added to my colleagues conviction that I was a crazy loon! https://mikecosgrave.com/blog2006/random/why-i-never-set-essays-anymore/
I think you and I are exploring two sides of the same coin: you’re looking at process-based assessment, and I’m looking at AI-enhanced processes to support cognition. It would be a fun conversation. 😊
One nuance worth naming is that the learning processes we ask of adults (like university students) are often quite different from what’s developmentally appropriate in K-12 contexts. For adult learners, it makes a lot of sense to deliberately require the use of AI and ask students to intentionally design their own learning processes (like your SEWPS).
In K-12 environments, however, I think more scaffolded processes are still essential, especially when the goal is to ensure AI-use supports effort rather than replaces it. As students mature and demonstrate the ability to manage their own learning with this emergent technology, it then makes sense to gradually shift toward having them design their own processes. That very well may include high school students, but I would say that we generally aren't there yet.
From my experience, most young people are not necessarily fluent in using AI to support learning (as opposed to just output). Neither are most teachers yet. I wonder if we will eventually get to a point where young people are generally more savvy with their use of AI and can use it effectively without as much scaffolding.
I also want to mention that the idea that process design should be explicitly about supporting different forms of thinking. Everything leading up to the creation of a product, whether that’s an artifact of thinking or a polished final outcome, is part of the "process." As you point out, not everything needs to be documented, but when documentation supports thinking for the learner and/or provides visibility for the teacher, it is worth doing. There are many things that could be an artifact of thinking too: screenshots, chats, sketches, etc. You get the point. I think I go back to Project Zero and how thinking routines can make thinking visible. I would say processes documentation can and should be in that line. I think one of the most powerful things process documentation can lead to, as your Transparency Statement document implies, is the metacognition that can come from it. I love hearing kids have a cognitive shift when they realize that they learned quite a bit and see their own growth. I imagine you and I are generally in agreement-- process should be about thinking not about forcing students to prove the didn't use AI.
I’d love to talk shop with you on your/my podcast sometime. I'm also writing a book called AI-Enhanced Processes (currently working on second edition). If you're keen, I'd love to be a critical friend or collaborator with you since we're both working on similar projects with what seem to be similar guiding questions.
Anyway, I'm going on and on. Thanks again for the great article and hope to speak with you soon!
This is a very reflective post that has me also thinking about a lot of assumptions I’ve made.
What first stood out for me in the example you gave where process needs documenting which led to ballooned output is how education in general needs and wants to treat the external. So much of learning is an internal process but the formal education system is based on capturing that in some way to then mark.
The second that that came to mind is indeed how prescribed the process is and whether it calls for students to learn how to define their own processes.
This makes a ton of sense. I’m in the middle of creating an online course for screenwriters using AI and have come to the realization that — if I want them to preserve their creative voice — they need to constantly improve their process and THEN use AI (product). You’ve stated this very very well.
Brilliant push back on the 'process over product' silver bullet. We can't let 'process' become just another compliance checkbox. Your point about emergence in creativity is essential. Thanks for the nuance!
You’re welcome! And I’m very much worried about that. I see “grade the process” and “grade the thinking a lot.”
My immediate response is usually, “please don’t.”
That '6 assignments under a trench coat' line is so spot on. With 20+ years in the ELA classroom, I’ve been deep in thought lately about how we need to reshape the writing process we’re teaching to better fit this new reality.
I just finished Eric Hudson’s series on AI skills (I'm late finding the series. He posted it almost a year ago), and his take on 'playfulness' and the 'extended mind' feel like such a helpful bridge here. They offer a way to move toward a more creative, expanded partnership with these tools rather than just turning thinking into another compliance checkbox. Thank you for this piece, which has further pushed my thinking.
https://erichudson.substack.com/p/ai-skills-that-matter-part-3-playfulness
From the k12 side of things, i’m seeing a torrent of products that are required to prove process.
You’ve written about it very well here - and I’m not sure you’ll agree with the conclusions I’ve come to based on my observations.
But- when you have more than a few students, it is nearly impossible to grade anything but Product.
You can break that product down into a little bits but anybody who studied agents knows that those little bits can be defined and created.
I sound like a fatalist - I’m actually really excited about K-12 AI, and sending those students on to you.
But oversimplification is not the answer .
Really looking forward to the rest of the series .
It’s true!
I don’t really have an answer for agents yet, and I’m not sure anyone does.
For me, the hope is to design better learning experiences.
And I will say that shifting to process hasn’t ballooned my work at all, especially with the shift to student agency and alternative assessment a bit more. I just spread out the work differently.
Nice piece, Jason. This feels important because the tendency you identify of treating process as a step-by-step formula and opposing it to product is, as Owen on LinkedIn says, merely recreating the problem. That's especially true if you're using a product to structure and track the completion of the process.
The point is to collapse the distinction and embrace the ungraded messiness that occurs when you turn a group of students on to a problem with no clear outcome or assurance of success.
That "third step" you outline is the key. It cuts against the systems we have set up that expect replication and control. The problem is not so much theoretical as our habits of mind about what we expect students to do. The older students get, the more their habits of mind are a factor. My own hot take is that students are the biggest barrier to adopting active learning approaches because it feels so alien and strange, and they have become addicted to being ranked and measured.
I can see the process model being appropriate in some contexts, but at some point, we assume the student knows the process. If we're still wanting to see how the student is making sausage in a 400-level course, we clearly are into surveillance. For a 100-level class, it makes sense. "This is how you do things in this field...."
As a doctoral student, my first class was a process class. For a 15-20 page research paper we had to submit an initial 20-source bibliography, then an annotated bibliography, then an outline, then the final product. IN other words, 4 total products. Fine. Still felt a bit elementary for a doctoral program, to be honest. But now, while those scaffolded steps are still available for most major papers, (if the student desires feedback) the only graded piece is the final product.
Interesting. I abandoned product long ago in favour of process. In recent years the first question I ask students is to outline their writing process: the answers are not surprisingly fairly basic but it lays down a benchmark before we work through the steps. I have a blog from 2014 describing my thoughts at that time, which added to my colleagues conviction that I was a crazy loon! https://mikecosgrave.com/blog2006/random/why-i-never-set-essays-anymore/
I really appreciated this thoughtful post, Jason.
I think you and I are exploring two sides of the same coin: you’re looking at process-based assessment, and I’m looking at AI-enhanced processes to support cognition. It would be a fun conversation. 😊
One nuance worth naming is that the learning processes we ask of adults (like university students) are often quite different from what’s developmentally appropriate in K-12 contexts. For adult learners, it makes a lot of sense to deliberately require the use of AI and ask students to intentionally design their own learning processes (like your SEWPS).
In K-12 environments, however, I think more scaffolded processes are still essential, especially when the goal is to ensure AI-use supports effort rather than replaces it. As students mature and demonstrate the ability to manage their own learning with this emergent technology, it then makes sense to gradually shift toward having them design their own processes. That very well may include high school students, but I would say that we generally aren't there yet.
From my experience, most young people are not necessarily fluent in using AI to support learning (as opposed to just output). Neither are most teachers yet. I wonder if we will eventually get to a point where young people are generally more savvy with their use of AI and can use it effectively without as much scaffolding.
I also want to mention that the idea that process design should be explicitly about supporting different forms of thinking. Everything leading up to the creation of a product, whether that’s an artifact of thinking or a polished final outcome, is part of the "process." As you point out, not everything needs to be documented, but when documentation supports thinking for the learner and/or provides visibility for the teacher, it is worth doing. There are many things that could be an artifact of thinking too: screenshots, chats, sketches, etc. You get the point. I think I go back to Project Zero and how thinking routines can make thinking visible. I would say processes documentation can and should be in that line. I think one of the most powerful things process documentation can lead to, as your Transparency Statement document implies, is the metacognition that can come from it. I love hearing kids have a cognitive shift when they realize that they learned quite a bit and see their own growth. I imagine you and I are generally in agreement-- process should be about thinking not about forcing students to prove the didn't use AI.
I’d love to talk shop with you on your/my podcast sometime. I'm also writing a book called AI-Enhanced Processes (currently working on second edition). If you're keen, I'd love to be a critical friend or collaborator with you since we're both working on similar projects with what seem to be similar guiding questions.
Anyway, I'm going on and on. Thanks again for the great article and hope to speak with you soon!
Alex
This is a very reflective post that has me also thinking about a lot of assumptions I’ve made.
What first stood out for me in the example you gave where process needs documenting which led to ballooned output is how education in general needs and wants to treat the external. So much of learning is an internal process but the formal education system is based on capturing that in some way to then mark.
The second that that came to mind is indeed how prescribed the process is and whether it calls for students to learn how to define their own processes.
Right! I think that’s what we need to be careful of!
It’s fine to say that students should submit process documents and be able to resubmit anything to prove their growth. But that balloons quickly!