Appreciate this sharing, and agree wholeheartedly regarding the importance of process and insisting upon meaningful writing experiences.
A wondering I have: in an ideal classroom with reasonable numbers of students and adequate planning time, this all sounds incredibly feasible. But in many of our situations (at least in the US) with 30+ students per class, a writing task involves going through 100+ writing samples—which here would mean watching video playbacks 100+ times, reviewing conversations via the Flint dashboard 100+ times, etc.
My concern, then, is that the thoughtful review and intentionality would get pushed out and it would end up, at least in a significant number of situations, with students essentially just writing for the AI tool and teachers recording the scores. While I don't think we should design classrooms for the worst-case scenarios, either, a repeated question I find myself asking is what any of these tools or practices look like at scale?
Very much appreciate your reflection and sharing—and especially the learning it avails to others on this journey, too!
Great questions. I also worry about these things at scale--and also, in the case of the Brisk tool, about students avoiding trying things out/erasing/starting over if they think it will look "suspect" or that they will look like they don't know what they're doing--when, of course, we want the to not know/to try things. I would not want someone looking at my writing process in this way, so that concerns me. Warren, I know you plan to follow up with some insights from teachers, but have you had any pushback that you can share here on these questions?
I have a meeting with some teachers and students planned in March to dig into exactly these questions. I know that I personally don't like typing in a Google Doc when I see another collaborator has it open, since I feel like they're watching over my shoulder. I wonder if younger people have that same self-consciousness and if there's a chilling effect on their open thinking when they worry that their early drafts are being judged.
I'm also thinking about scalability. If we design a wonderful system that takes too much time to implement, then we'll fall back on old processes or just give up. For now, I think the optimistic outcome is to reduce time spent on administrative tasks that don't improve learning, determine which tasks are best done by a human teacher and lean into those. We have a finite amount of time to spend on each student, so any strategy that can make that time more effective is worth working towards.
Marcus, I share your concern about what's been called the "AI Death Spiral." Teachers use AI to design an assessment, students use AI to complete it, teachers use an AI to assess it, and humans are out of the loop. At least recognizing that it's both a possible outcome and an undesirable one helps us avoid it. And hopefully helps EdTech vendors from designing towards it.
This is excellent and so agree and just reminds me of my own time in Japan and learning about Shokunin and how the process of continuous improvement and the process is so more meaningful than the end product, thank you for this insights and reflections
Thank you for the tips in using AI in the writing process. I am most interested in the “meaningful engagement” piece of this question. How do we help students see/understand that the process of writing has value? What kinds of writing tasks are you assigning that students are excited to engage in? Writing an analysis of a poem for the teacher to read (I’m assuming to help the teacher see if you understood the poem or can use skills of analysis) doesn’t seem like it will engage many students. Would love to get ideas on how to help students WANT to engage in writing. 😃
Yes! Totally agreed. This article is a briefer version of a keynote I'm doing for an AI conference next week where I go into more examples. One of my favorites is a writing assignment in grades 11/12 right now. The students are reading In Cold Blood and they're writing in a crime genre. They get to choose whether they write a psychological thriller, a true crime story, a police procedural, or a whodunit. It's a much more engaging piece of writing than a poem analysis essay. We also have a class called Creative Non-Fiction where students are really pushing the boundaries of the format - then they choose their best work for publication in a physically-printed book. I really like having physical publication as an end goal for a high school class, since it gives kids an audience that extends beyond the school (even if it's a very very small audience.)
Glad you liked it! I've compared Draftback and Brisk, and the reason I prefer Brisk is that it makes it easier to jump through the timeline to both Comments and to large copy-paste actions. Aside from that, they do have similar features.
Appreciate this sharing, and agree wholeheartedly regarding the importance of process and insisting upon meaningful writing experiences.
A wondering I have: in an ideal classroom with reasonable numbers of students and adequate planning time, this all sounds incredibly feasible. But in many of our situations (at least in the US) with 30+ students per class, a writing task involves going through 100+ writing samples—which here would mean watching video playbacks 100+ times, reviewing conversations via the Flint dashboard 100+ times, etc.
My concern, then, is that the thoughtful review and intentionality would get pushed out and it would end up, at least in a significant number of situations, with students essentially just writing for the AI tool and teachers recording the scores. While I don't think we should design classrooms for the worst-case scenarios, either, a repeated question I find myself asking is what any of these tools or practices look like at scale?
Very much appreciate your reflection and sharing—and especially the learning it avails to others on this journey, too!
Great questions. I also worry about these things at scale--and also, in the case of the Brisk tool, about students avoiding trying things out/erasing/starting over if they think it will look "suspect" or that they will look like they don't know what they're doing--when, of course, we want the to not know/to try things. I would not want someone looking at my writing process in this way, so that concerns me. Warren, I know you plan to follow up with some insights from teachers, but have you had any pushback that you can share here on these questions?
I have a meeting with some teachers and students planned in March to dig into exactly these questions. I know that I personally don't like typing in a Google Doc when I see another collaborator has it open, since I feel like they're watching over my shoulder. I wonder if younger people have that same self-consciousness and if there's a chilling effect on their open thinking when they worry that their early drafts are being judged.
I'm also thinking about scalability. If we design a wonderful system that takes too much time to implement, then we'll fall back on old processes or just give up. For now, I think the optimistic outcome is to reduce time spent on administrative tasks that don't improve learning, determine which tasks are best done by a human teacher and lean into those. We have a finite amount of time to spend on each student, so any strategy that can make that time more effective is worth working towards.
Marcus, I share your concern about what's been called the "AI Death Spiral." Teachers use AI to design an assessment, students use AI to complete it, teachers use an AI to assess it, and humans are out of the loop. At least recognizing that it's both a possible outcome and an undesirable one helps us avoid it. And hopefully helps EdTech vendors from designing towards it.
This is excellent and so agree and just reminds me of my own time in Japan and learning about Shokunin and how the process of continuous improvement and the process is so more meaningful than the end product, thank you for this insights and reflections
Thank you for the tips in using AI in the writing process. I am most interested in the “meaningful engagement” piece of this question. How do we help students see/understand that the process of writing has value? What kinds of writing tasks are you assigning that students are excited to engage in? Writing an analysis of a poem for the teacher to read (I’m assuming to help the teacher see if you understood the poem or can use skills of analysis) doesn’t seem like it will engage many students. Would love to get ideas on how to help students WANT to engage in writing. 😃
Yes! Totally agreed. This article is a briefer version of a keynote I'm doing for an AI conference next week where I go into more examples. One of my favorites is a writing assignment in grades 11/12 right now. The students are reading In Cold Blood and they're writing in a crime genre. They get to choose whether they write a psychological thriller, a true crime story, a police procedural, or a whodunit. It's a much more engaging piece of writing than a poem analysis essay. We also have a class called Creative Non-Fiction where students are really pushing the boundaries of the format - then they choose their best work for publication in a physically-printed book. I really like having physical publication as an end goal for a high school class, since it gives kids an audience that extends beyond the school (even if it's a very very small audience.)
Great ideas! Thank you! Will you publish full keynote speech?
Helpful article. Thank you for the focus on process. We have used Draftback for similar features.
Glad you liked it! I've compared Draftback and Brisk, and the reason I prefer Brisk is that it makes it easier to jump through the timeline to both Comments and to large copy-paste actions. Aside from that, they do have similar features.
Brisk sounds interesting. I will check it out.