I'm very much with you on relying on version histories, but typically just when something feels off and to initiate a conversation rather than to "prove" something. And your note about showing them this at the front end of the course and explaining how it will be used is really helpful to me—definitely something I'll be doing in the fall.
I also worry, though, that we're drifting into a lose/lose landscape in the years ahead: where there will be an incredibly narrow path to value trust along with accountability, and I worry that both writing and relationships will suffer as a result.
I agree completely about the intrusiveness of process tracking, but I am also fully daunted by the workload implications. I have 100 students every semester in first-year writing courses. I can barely keep up with reading their stuff as it is.
Great post and others made this point as well -, it's not enough now to grade student work but we have to become writing sleuths in order to determine whether something as vague and idiosyncratic as someone's writing process is authentic? As one of my students opined when I explained the school was employing process tracking software, "sounds Orwellian." I think it's a complete invasion of one of the most intimate and vulnerable acts of expression and I implore every teacher who is in favor to undergo the exact process you did. I tell students my default position is I won't use it but they need to be aware it exists.
Appreciated this in that it balances a question of 'what works' with 'what effects might it be having' with 'what is the experience like'? Thanks for exploring this, as the 'try it yourself' ethos might chasten a lot of of the tracking-curious... how does it feel to do this yourself?
It's interesting to read about this in a writing context. I'm an illustrator, and it's pretty common for artists/students to make screen recordings of their artmaking process to use on social media. Those kinds of posts are super popular and can get them lots of likes or attention online. I haven't asked for them for grading, but a development portfolio of all their work along the way to a final piece (including bad drawings or artwork!) are really important for marking assessments. Thanks for sharing!
It's massively intrusive. And as you yourself said, some students handwrite their first draft (or multiple drafts) on paper. BTW, some non-native English speakers might write their first draft in a language other than English.
This has concerned me since AI hit the English classroom. Thank you for raising these issues and complications here. I, too, am stuck with many of these questions.
In addition, if you do develop your reflective assignment, I would be very interested in reading about it!
Thank you for sharing your reflections on your experiences! I wonder to what extent it might be helpful to think about athletes reviewing game footage as a model for reflecting on process. Or using video to help athletes perfect key movements (like a golf swing). The comparison is imperfect, certainly, but might have some utility.
More seriously: One of my main goals is to help students experience the writing process in terms of learning and creativity. For me, personally, the kind of process tracking described here feels like it runs counter to what I'm trying to do.
What works better, for me, is to simply sit down and talk with my students about what they're working on. Sometimes this involves looking at their work in progress, sometimes it's just talking about how things are going. A significant part of our class time is organized something like a studio art class, where students have time to work on their projects and I have time to check in with them individually, which provides a framework for having these conversations on a regular basis. It's not necessarily easy - I'm definitely tired at the end of the day - but in my experience it works really well.
While I didn't start doing this in response to AI, I think it's one reason why I've been getting comparatively low levels of stuff that concerns me on that score. (I do see some, and I'm sure I miss some.) Combined with ungrading and significant space for student choice, the ongoing personal interaction seems to foster an environment where students are mostly making good decisions.
For context, I teach writing at a technical community college, and I typically have 100+ students a semester, with fairly limited contact time. I could wish for smaller classes and more time with them, but I'm able to make it work. With bigger numbers, or more subject-knowledge-oriented classes, it would definitely be more difficult.
I really appreciated this reflection; I wouldn't have thought of all the issues that might come up for students until I read a description of the experience of process tracking. What really strikes me is how it turns writing into a performance -- which isn't inherently bad, but (1) it's a different thing from what most of us think we're doing when we assign writing; and (2) so much of students' lives is already about performance.
I think that's one of the main things I'm taking away. We've set up a system in which students are already trying to perform writing instead of actually writing. I'm worried that process tracking might exacerbate this.
I’ve been using Google Docs in my teaching for years, long before AI, because Google Docs is a free cross-platform software that “just works,” which makes teaching students how to use it a lot easier, and so I can see the changes students make in revision. I usually allow students to make revisions to improve a final grade, and looking at the version history allows me to see what they did (or didn’t do) to revise quite clearly. Using it now to try to detect AI is, for me, an added benefit.
But look, the writing we ask students to do for school is already a performance that is quite different than the self-initiated/motivated writing I’m doing right now, different than some kind of writing I do on my own that isn’t part of some kind of assignment. Yes, Google Docs does enable “surveillance,” but so does the very act of requiring students to write in a way that meets the outcomes of an assignment, and that ends up being evaluated in terms of a grading scale. Schooling (especially when it involves grading and testing) is surveillance, full stop. Part of teaching writing as a process is asking students to show their work and demonstrate that process. I used to do that kind of surveillance by looking at paper copies of rough drafts. How is this different?
I think you're right that grades and testing turn school itself into surveillance. But I don't think that's how it has to be! I want to minimize the performance/outcomes/grades part so that we can actually do the "self-initiated/motivated" writing that is better at facilitating learning. Because we live in a system that requires grades, I have to make some concessions. But ideally, I wouldn't grade at all.
I get that, and I sympathize. I also don’t think there is a clear correlation between grades and learning in that I have a lot of students in a class like first year writing who scrape by with a C who end up learning A LOT more than students who get an A, especially with the A student who came into the class already knowing a lot about how to research and write well. I wish all of my students were self-initiated/motivated to learn for learning’s sake, which is one of the reasons why I always liked working in writing centers. Students come in— generally encouraged, but voluntarily— because they’re self-motivated to improve. Consultants/tutors get to work with those students with no judgment, just to help. Plus grading writing takes A LOT of work, and it is certainly not nearly as much fun as it is to lead students through class activities and discussions.
But that’s how it works.
Students attend colleges/universities to earn degrees, which are widely recognized credentials that are supposed to represent a certain level of expertise and experience in a particular subject matter. Degrees are made up of a series of individual courses students complete based on some kind of instructor assessment: that is, if a student passes the specific courses required to graduate with a degree in English or business or whatever, then they are awarded that degree.
Now, there are all kinds of ways to try to make students focus on the learning rather than the grading: so-called “ungrading” approaches to make the process more humane, such as contract grading, considering students’ self-assessment of some parts of the course as part of the process, and so forth. I see it as part of my job to get students to not worry too much about the differences between different grades and to focus instead on what they’re learning. But the reality is that most of us are not inherently self-motivated to learn things in a systematic way, and that is certainly true with subjects/classes we are not that interested in.
And while grading and evaluation don’t always correspond to learning, I think they generally correspond to competence, which I think is important. I mean, I’m glad that my doctor had to pass all of her courses and tests before they let her go out and practice medicine, and the same goes for a lawyer I might hire or the engineer who designed the car I drive in, etc.
I have students enable track changes during the writing process and keep it on until the "final" version is submitted. The courses align with the labor-based grading model. I show all students how to use the editing tracker and provide YouTube videos as support. I score for process over product. Students who do not follow directions, regardless of their writing competency, are docked points. If the student uses AI as a research and/or writing partner and cites it correctly, this is not considered plagiarism.
Great point. The effect of having your original writing suspected of being done with AI is a reality that many teachers and students need to face. Culturally, I believe we are at the point now where all writing is first suspect of being written, or at least modified, with AI. For practical purposes, people would be foolish to NOT use AI as a tool. For students preparing for the workplace, the need to be familiar with how to leverage AI as a productivity tool is growing in demand by employers and industry. So I am trying to use process tracking as well as using AI in the classroom as an interactive tool WITH the students so we can discuss limitations, challenges, and how to more effectively use it.
I have been using version history to help me when I suspect AI use too, as it gives a springboard to open up the conversation, and I really appreciated this post. I haven't heard of Authorship and while I see the advantages it also does make me feel a bit uneasy. I feel better about version history than I would about a full screen recording for a typical essay (for a test, lockdown browser has screen recording, which is fine to me, but tests are doing something different).
I love your commitment to be more transparent about your use of version history, I think I could do this too. I have only recently starting using it and have vaguely been saying that it's helpful for me to evaluate process, but this might be sort of disingenuous since the real reason I use it is primarily for academic honesty/ AI reasons. Part of me worries that the student will read this as instructions on how to cheat and get away with it - like if they use AI, but type it in themselves, then they can use the version history to "prove" it was typed in themselves? Maybe it is an unfounded worry, and in any case, not being fully transparent doesn't do anything to help with that worry.
I also love your idea to create a reflection assignment where students look at their own version histories. Thank you for sharing your reflections and ideas!!
Thanks for this reflection. I found myself reflecting on the importance of trust to learning, the way teaching in every field is moving toward coaching, and the lifelong cognitive, social, and emotional value of helping children to learn early not to fear sharing their work. I especially appreciate the nuanced take on neurodiversity; I think it's important to acknowledge the likelihood of elevated fear/shame re process monitoring among neurodiverse students while also recognizing that neurotypical folks often feel both as well. And if we taught children to share process early, we'd catch more neurodiverse kids earlier and help reduce that fear/shame differential.
As someone whose handwriting once brought down my entire class's average and caused grading issues for teachers (and therefore for me), I'm less persuaded that AI is a major threat to human cognitive competence (as manifested in writing) than some folks. AI as a writing tool has raised sharply the bar for minimal effort. When it's necessary to judge, former Fs are now AI-enhanced Ds or Cs--but former As are still AI-As, because average AI use/competence produces average AI-work (and as average competence improves, we can raise the bar again). If every student uses AI, and we shift pedagogy to teaching what students need to know and do produce superior-to-AI text as adults in an AI world, how is that a loss?
Getting an AI to produce *superior* writing is an intensely interactive, multi-step act of creativity and rigor using all the same cognitive, social, and emotional tools as writing with pen-and-paper or keyboard. The output may be better (or just bigger), but the process, which is heavy on reflection and revision and requires original drafting in both cases, looks more different than it is. Assessing AI-supported writing does differ--but does it differ importantly, or are we just a bit stuck in our own ways of doing and thinking?
I'm hopeful that something closer to the above is where we end up. That's especially true because I don't see process monitoring as working for long: it's a variant of AI-detection, and educational institutions cannot win that arms-race. It's already possible to get an AI agent to write a paper simulating the way a human would write, even while being watched by a screen recorder, and while it's easier today to write the paper yourself, that won't be true much longer. We'll have to go to biometric monitoring, then human proctoring.
I'm very much with you on relying on version histories, but typically just when something feels off and to initiate a conversation rather than to "prove" something. And your note about showing them this at the front end of the course and explaining how it will be used is really helpful to me—definitely something I'll be doing in the fall.
I also worry, though, that we're drifting into a lose/lose landscape in the years ahead: where there will be an incredibly narrow path to value trust along with accountability, and I worry that both writing and relationships will suffer as a result.
I worry about this too. Lots of fraught decisions for teachers to make.
I agree completely about the intrusiveness of process tracking, but I am also fully daunted by the workload implications. I have 100 students every semester in first-year writing courses. I can barely keep up with reading their stuff as it is.
Great post and others made this point as well -, it's not enough now to grade student work but we have to become writing sleuths in order to determine whether something as vague and idiosyncratic as someone's writing process is authentic? As one of my students opined when I explained the school was employing process tracking software, "sounds Orwellian." I think it's a complete invasion of one of the most intimate and vulnerable acts of expression and I implore every teacher who is in favor to undergo the exact process you did. I tell students my default position is I won't use it but they need to be aware it exists.
Appreciated this in that it balances a question of 'what works' with 'what effects might it be having' with 'what is the experience like'? Thanks for exploring this, as the 'try it yourself' ethos might chasten a lot of of the tracking-curious... how does it feel to do this yourself?
It's interesting to read about this in a writing context. I'm an illustrator, and it's pretty common for artists/students to make screen recordings of their artmaking process to use on social media. Those kinds of posts are super popular and can get them lots of likes or attention online. I haven't asked for them for grading, but a development portfolio of all their work along the way to a final piece (including bad drawings or artwork!) are really important for marking assessments. Thanks for sharing!
Interesting! Thanks for sharing this. I'm now thinking about the similarities and differences between the verbal and visual mediums here...
It's massively intrusive. And as you yourself said, some students handwrite their first draft (or multiple drafts) on paper. BTW, some non-native English speakers might write their first draft in a language other than English.
Yes, thanks for raising this!
This has concerned me since AI hit the English classroom. Thank you for raising these issues and complications here. I, too, am stuck with many of these questions.
In addition, if you do develop your reflective assignment, I would be very interested in reading about it!
Thank you for sharing your reflections on your experiences! I wonder to what extent it might be helpful to think about athletes reviewing game footage as a model for reflecting on process. Or using video to help athletes perfect key movements (like a golf swing). The comparison is imperfect, certainly, but might have some utility.
I've seen the sports metaphor coming up a lot! This is another interesting application of it.
I can't help thinking of this Monty Python skit involving Thomas Hardy:
https://www.youtube.com/watch?v=ogPZ5CY9KoM
More seriously: One of my main goals is to help students experience the writing process in terms of learning and creativity. For me, personally, the kind of process tracking described here feels like it runs counter to what I'm trying to do.
What works better, for me, is to simply sit down and talk with my students about what they're working on. Sometimes this involves looking at their work in progress, sometimes it's just talking about how things are going. A significant part of our class time is organized something like a studio art class, where students have time to work on their projects and I have time to check in with them individually, which provides a framework for having these conversations on a regular basis. It's not necessarily easy - I'm definitely tired at the end of the day - but in my experience it works really well.
While I didn't start doing this in response to AI, I think it's one reason why I've been getting comparatively low levels of stuff that concerns me on that score. (I do see some, and I'm sure I miss some.) Combined with ungrading and significant space for student choice, the ongoing personal interaction seems to foster an environment where students are mostly making good decisions.
For context, I teach writing at a technical community college, and I typically have 100+ students a semester, with fairly limited contact time. I could wish for smaller classes and more time with them, but I'm able to make it work. With bigger numbers, or more subject-knowledge-oriented classes, it would definitely be more difficult.
Thanks for sharing this. I'm having similar experiences, but I teach *far* fewer students. I'm really happy to know that it can be done with 100+!
I really appreciated this reflection; I wouldn't have thought of all the issues that might come up for students until I read a description of the experience of process tracking. What really strikes me is how it turns writing into a performance -- which isn't inherently bad, but (1) it's a different thing from what most of us think we're doing when we assign writing; and (2) so much of students' lives is already about performance.
I think that's one of the main things I'm taking away. We've set up a system in which students are already trying to perform writing instead of actually writing. I'm worried that process tracking might exacerbate this.
I’ve been using Google Docs in my teaching for years, long before AI, because Google Docs is a free cross-platform software that “just works,” which makes teaching students how to use it a lot easier, and so I can see the changes students make in revision. I usually allow students to make revisions to improve a final grade, and looking at the version history allows me to see what they did (or didn’t do) to revise quite clearly. Using it now to try to detect AI is, for me, an added benefit.
But look, the writing we ask students to do for school is already a performance that is quite different than the self-initiated/motivated writing I’m doing right now, different than some kind of writing I do on my own that isn’t part of some kind of assignment. Yes, Google Docs does enable “surveillance,” but so does the very act of requiring students to write in a way that meets the outcomes of an assignment, and that ends up being evaluated in terms of a grading scale. Schooling (especially when it involves grading and testing) is surveillance, full stop. Part of teaching writing as a process is asking students to show their work and demonstrate that process. I used to do that kind of surveillance by looking at paper copies of rough drafts. How is this different?
I think you're right that grades and testing turn school itself into surveillance. But I don't think that's how it has to be! I want to minimize the performance/outcomes/grades part so that we can actually do the "self-initiated/motivated" writing that is better at facilitating learning. Because we live in a system that requires grades, I have to make some concessions. But ideally, I wouldn't grade at all.
I get that, and I sympathize. I also don’t think there is a clear correlation between grades and learning in that I have a lot of students in a class like first year writing who scrape by with a C who end up learning A LOT more than students who get an A, especially with the A student who came into the class already knowing a lot about how to research and write well. I wish all of my students were self-initiated/motivated to learn for learning’s sake, which is one of the reasons why I always liked working in writing centers. Students come in— generally encouraged, but voluntarily— because they’re self-motivated to improve. Consultants/tutors get to work with those students with no judgment, just to help. Plus grading writing takes A LOT of work, and it is certainly not nearly as much fun as it is to lead students through class activities and discussions.
But that’s how it works.
Students attend colleges/universities to earn degrees, which are widely recognized credentials that are supposed to represent a certain level of expertise and experience in a particular subject matter. Degrees are made up of a series of individual courses students complete based on some kind of instructor assessment: that is, if a student passes the specific courses required to graduate with a degree in English or business or whatever, then they are awarded that degree.
Now, there are all kinds of ways to try to make students focus on the learning rather than the grading: so-called “ungrading” approaches to make the process more humane, such as contract grading, considering students’ self-assessment of some parts of the course as part of the process, and so forth. I see it as part of my job to get students to not worry too much about the differences between different grades and to focus instead on what they’re learning. But the reality is that most of us are not inherently self-motivated to learn things in a systematic way, and that is certainly true with subjects/classes we are not that interested in.
And while grading and evaluation don’t always correspond to learning, I think they generally correspond to competence, which I think is important. I mean, I’m glad that my doctor had to pass all of her courses and tests before they let her go out and practice medicine, and the same goes for a lawyer I might hire or the engineer who designed the car I drive in, etc.
I have students enable track changes during the writing process and keep it on until the "final" version is submitted. The courses align with the labor-based grading model. I show all students how to use the editing tracker and provide YouTube videos as support. I score for process over product. Students who do not follow directions, regardless of their writing competency, are docked points. If the student uses AI as a research and/or writing partner and cites it correctly, this is not considered plagiarism.
Great point. The effect of having your original writing suspected of being done with AI is a reality that many teachers and students need to face. Culturally, I believe we are at the point now where all writing is first suspect of being written, or at least modified, with AI. For practical purposes, people would be foolish to NOT use AI as a tool. For students preparing for the workplace, the need to be familiar with how to leverage AI as a productivity tool is growing in demand by employers and industry. So I am trying to use process tracking as well as using AI in the classroom as an interactive tool WITH the students so we can discuss limitations, challenges, and how to more effectively use it.
I have been using version history to help me when I suspect AI use too, as it gives a springboard to open up the conversation, and I really appreciated this post. I haven't heard of Authorship and while I see the advantages it also does make me feel a bit uneasy. I feel better about version history than I would about a full screen recording for a typical essay (for a test, lockdown browser has screen recording, which is fine to me, but tests are doing something different).
I love your commitment to be more transparent about your use of version history, I think I could do this too. I have only recently starting using it and have vaguely been saying that it's helpful for me to evaluate process, but this might be sort of disingenuous since the real reason I use it is primarily for academic honesty/ AI reasons. Part of me worries that the student will read this as instructions on how to cheat and get away with it - like if they use AI, but type it in themselves, then they can use the version history to "prove" it was typed in themselves? Maybe it is an unfounded worry, and in any case, not being fully transparent doesn't do anything to help with that worry.
I also love your idea to create a reflection assignment where students look at their own version histories. Thank you for sharing your reflections and ideas!!
To be honest, I worry about that too. It's difficult to keep the faith! Thanks for your kind words about the post!
Thanks for this reflection. I found myself reflecting on the importance of trust to learning, the way teaching in every field is moving toward coaching, and the lifelong cognitive, social, and emotional value of helping children to learn early not to fear sharing their work. I especially appreciate the nuanced take on neurodiversity; I think it's important to acknowledge the likelihood of elevated fear/shame re process monitoring among neurodiverse students while also recognizing that neurotypical folks often feel both as well. And if we taught children to share process early, we'd catch more neurodiverse kids earlier and help reduce that fear/shame differential.
As someone whose handwriting once brought down my entire class's average and caused grading issues for teachers (and therefore for me), I'm less persuaded that AI is a major threat to human cognitive competence (as manifested in writing) than some folks. AI as a writing tool has raised sharply the bar for minimal effort. When it's necessary to judge, former Fs are now AI-enhanced Ds or Cs--but former As are still AI-As, because average AI use/competence produces average AI-work (and as average competence improves, we can raise the bar again). If every student uses AI, and we shift pedagogy to teaching what students need to know and do produce superior-to-AI text as adults in an AI world, how is that a loss?
Getting an AI to produce *superior* writing is an intensely interactive, multi-step act of creativity and rigor using all the same cognitive, social, and emotional tools as writing with pen-and-paper or keyboard. The output may be better (or just bigger), but the process, which is heavy on reflection and revision and requires original drafting in both cases, looks more different than it is. Assessing AI-supported writing does differ--but does it differ importantly, or are we just a bit stuck in our own ways of doing and thinking?
I'm hopeful that something closer to the above is where we end up. That's especially true because I don't see process monitoring as working for long: it's a variant of AI-detection, and educational institutions cannot win that arms-race. It's already possible to get an AI agent to write a paper simulating the way a human would write, even while being watched by a screen recorder, and while it's easier today to write the paper yourself, that won't be true much longer. We'll have to go to biometric monitoring, then human proctoring.