100% yes on the "students are using generative AI as part of/in lieu of doing complex reading." I've got a student in AP US History who started the semester totally bamboozled by the textbook, which was a major roadblock for him. Clearly he'd never had significant reading expectations prior to this, and the book was a major ramp-up that he didn't know how to address, which led to panic, magical thinking, and a lot of bargaining with himself. He admitted later to a strategy of feeding chapters through ChapGPT (he had to come up with a workaround given the copyright restrictions on the e-text), and then studying the results. We've had a lot of discussions about this - lots of ex post facto justifications, most of which I've shot down - but on the other side, he's mostly judging the attempt unfavorably because it wound up being more time-consuming than he had expected, not on any other basis. It's the wild west out there right now, and we're playing whack-a-mole, to wildly mix the metaphor.
Thank you for sharing! When I was at the high school level, I thought all the time about how 99% of my students didn't have good textbook reading skills (fundamentally different from narrative reading skills, I think). And I realized it was going to take a lot of time and effort on my part to teach those skills. I tried, but not with a wild amount of success. Your last point is so telling; I worry that too many schools are either implicitly or explicitly building altars to the Gods of Ultimate Efficiency at the expense of any and all other values, and too many students will grow up a) not knowing how to slow down, b) not knowing that they can or even what it feels like, and c) having no idea what is worth their extra time or attention.
It concerns me that students didn't apparently understand that entering someone else's book, or any intellectual property, into Chat GPT without explicit permission is intellectual theft. Did you discuss this point? Not all writers want their books chopped into AI slop. I don't. Would they consent to having their work chopped up and given away? Seems like a teachable moment.
Also seems like a moment to discuss the value of originality, critical and creative thinking. You might find this series of posts on originality relevant:
I agree with your point. However, in response to the question “Would they consent to having their work chopped up and given away?” the answer for most students is, “Sure. Why not?” I’ve had this conversation with undergraduate students that I teach. Unlike previous generations, current undergraduates seem to lack any sense that anyone’s work actually belongs to him or her. They are so accustomed to their own words and identities being used by social media for everything from advertisements to AI that—to paint with a somewhat too-broad brush—everything is perceived as public domain.
Fabulous start to The Important Work! And thanks for introducing me to another great Substack.
I especially like the conclusion. Helping students learn how to live in the world, which includes when and how to resist social forces pressing upon them, as well as ways to use new technology to respond to human challenges, seems like exactly the right frame.
Thanks for sharing these survey results. I actually shifted the assignment in my US history survey course from a short analysis paper to a book hunt assignment that requires them to locate, check out, and explore a book from the library. It can't be done by AI, and the goal is to make them more comfortable with both finding books and actually reading at least part of them. Here's a blog post I wrote about the assignment: https://cassandragoodhistorian.com/2024/11/12/the-book-hunt-assignment/
Glad to see someone raise the question of ways students might be using AI to read, not just to write. As a secondary English teacher, reading is an integral concern of mine, and I'm going to be very interested to follow this discussion.
Agreed. In some ways, my first instinct was a higher level of frustration about use for reading than for writing. I'm obviously frustrated by both, but I'll be thinking a lot about the reading issue and ways to combat this. It starts with assigning books at the K-12 level...
This is very interesting and I love to see the combination of Rhetoric with data-informed insights! On the writing side of things, the optimist in me wonders if AI, used judiciously, could help students get over the intimidating and often steep hill of getting started with a writing project. I spoke with a programmer last week who suggested this is a benefit of using AI when coding and his argument is making me think. As to reading, and using AI to summarize readings, well the hallucination problem remains very real and for now if students rely on it they may end up turning things in with very mistaken arguments. Something to monitor, perhaps.
Something that worries me about how AI might be used to get started: I want my students to be writing things because they care about them and want to think about them. But it's so easy to just get AI to "brainstorm" an idea--which feeds into the writing-as-jumping-through-a-hoop-to-a-product framework. That is not to say it *can't* be used to get started but that this is another area where how it can be used seems at odds with how it is likely to be used. Which leads back to what we're trying to do in the first place, etc., etc.
I agree, and as someone who spent eight years teaching high school students, I think there is something fundamentally important about inhabiting that feeling of "blankness," learning to pay attention to the fact that ideas seem to spring out of nowhere if you allow yourself to get comfortable with the discomfort for long enough. I think it is probably *the* key moment/method by which students learn to trust themselves as thinkers, and believe in themselves as idea-havers.
Agreed that this is worrying. I feel like writing skills help you (1) come up with interesting ideas and (2) articulate them well. But outside of the classroom the former seems like a necessary prerequisite to the latter. If you aren’t able to really develop thoughts on the world happening around you, it’s not much use being able to articulate those lack of thoughts to others. It’s the kind of thing that probably makes life richer even if you don’t share these observations with anyone else and it’s a shame if students will have less practice refining those skills. This was just a small sample of mixed data but yes hopefully building the skill of “starting” can be protected
Just to be provocative, did you ask the students to see the chapter summaries and explanations of certain parts of the readings they didn't understand? Did you ask to look at the questions they asked, responses, and follow ups in the chats? I'm just curious if you evaluated the overall quality of those interactions. From what you are saying, many of the students actually did the reading but used the AI as a companion or helper with sections they wanted to get more out of. Is your point that they should have highlighted those questions and come to office hours instead?
I read—and grade—all of their discussion posts, so I see the way they interact with every part of the book. My point is not that students never do the reading; most of them did do it. My point is that among the students who are at least occasionally using AI, they're rarely using it for generative writing, and more frequently using it for reading assistance. I'd prefer that students read on their own, respond to each prompt to the best of their ability, and then a) come to office hours to discuss if my feedback shows they significantly missed the mark, or b) pay attention to the smaller issues that may have cost them a point or two and keep those in mind as they proceed.
I think the cart is being put before the horse here. Students need assistance with reading . . . yes . . . even at the college level. Read alouds in class are great . . . teacher-created reading guides with oh so specific questions that require evidence from the text (something AI is not great at doing) help as well. Then discussion to follow as a group and then on their own. Students cannot respond to text they do not understand.
Asking students to refrain from using AI as a reading assistance tool is not the same as expecting them to meet the demands of the course with zero assistance. All of the above pedagogical methods exist in my classroom. Understanding of a text is also never an all-or-nothing phenomenon. Students always understand *something* about a text and can begin to work from there. (If most/all students truly understand nothing—and I've never seen this happen—that just means the text is ill-chosen for the course.) My larger goal is here is to keep students attentive to what they understand and what they don't understand, and then use their relationships with me and with each other (rather than AI) to pursue what might be beyond their grasp at any particular moment.
Understood. What I'm wondering is if any student shared with you their actual exchange with AI - what they asked and how it responded. Having read Postman's work, I can see ways in which a curious student could have a productive conversation with AI that pushed them deeper into engaging with and challenging the text. However, I can also see how, if they are doing that prior to a discussion post, their response may be informed more by their conversation with AI than their initial reaction to the text, but I just don't see how that is not going to become the norm for students going forward. It might be worth engaging with AI yourself and probing aspects of Postman's work to see the quality of answers so you can point out to students why you think it is either misleading, superficial, or otherwise not helping them. Are you familiar with Google's Notebook LM tool? You can put up to 300 PDF's, documents, Youtube videos, etc... and use these as the basis for a chat which will comb the material and provide footnotes in its responses taking you precisely to the text that supports its analysis. This is just what is going to be available to future generations of students and something teachers will have to be prepared to discuss in their classes.
100% yes on the "students are using generative AI as part of/in lieu of doing complex reading." I've got a student in AP US History who started the semester totally bamboozled by the textbook, which was a major roadblock for him. Clearly he'd never had significant reading expectations prior to this, and the book was a major ramp-up that he didn't know how to address, which led to panic, magical thinking, and a lot of bargaining with himself. He admitted later to a strategy of feeding chapters through ChapGPT (he had to come up with a workaround given the copyright restrictions on the e-text), and then studying the results. We've had a lot of discussions about this - lots of ex post facto justifications, most of which I've shot down - but on the other side, he's mostly judging the attempt unfavorably because it wound up being more time-consuming than he had expected, not on any other basis. It's the wild west out there right now, and we're playing whack-a-mole, to wildly mix the metaphor.
Thank you for sharing! When I was at the high school level, I thought all the time about how 99% of my students didn't have good textbook reading skills (fundamentally different from narrative reading skills, I think). And I realized it was going to take a lot of time and effort on my part to teach those skills. I tried, but not with a wild amount of success. Your last point is so telling; I worry that too many schools are either implicitly or explicitly building altars to the Gods of Ultimate Efficiency at the expense of any and all other values, and too many students will grow up a) not knowing how to slow down, b) not knowing that they can or even what it feels like, and c) having no idea what is worth their extra time or attention.
It concerns me that students didn't apparently understand that entering someone else's book, or any intellectual property, into Chat GPT without explicit permission is intellectual theft. Did you discuss this point? Not all writers want their books chopped into AI slop. I don't. Would they consent to having their work chopped up and given away? Seems like a teachable moment.
Also seems like a moment to discuss the value of originality, critical and creative thinking. You might find this series of posts on originality relevant:
https://open.substack.com/pub/janetsalmons/p/encourage-originality-create-a-culture?utm_source=share&utm_medium=android&r=410aa5
I agree with your point. However, in response to the question “Would they consent to having their work chopped up and given away?” the answer for most students is, “Sure. Why not?” I’ve had this conversation with undergraduate students that I teach. Unlike previous generations, current undergraduates seem to lack any sense that anyone’s work actually belongs to him or her. They are so accustomed to their own words and identities being used by social media for everything from advertisements to AI that—to paint with a somewhat too-broad brush—everything is perceived as public domain.
Fabulous start to The Important Work! And thanks for introducing me to another great Substack.
I especially like the conclusion. Helping students learn how to live in the world, which includes when and how to resist social forces pressing upon them, as well as ways to use new technology to respond to human challenges, seems like exactly the right frame.
Thanks for sharing these survey results. I actually shifted the assignment in my US history survey course from a short analysis paper to a book hunt assignment that requires them to locate, check out, and explore a book from the library. It can't be done by AI, and the goal is to make them more comfortable with both finding books and actually reading at least part of them. Here's a blog post I wrote about the assignment: https://cassandragoodhistorian.com/2024/11/12/the-book-hunt-assignment/
Glad to see someone raise the question of ways students might be using AI to read, not just to write. As a secondary English teacher, reading is an integral concern of mine, and I'm going to be very interested to follow this discussion.
Agreed. In some ways, my first instinct was a higher level of frustration about use for reading than for writing. I'm obviously frustrated by both, but I'll be thinking a lot about the reading issue and ways to combat this. It starts with assigning books at the K-12 level...
This is very interesting and I love to see the combination of Rhetoric with data-informed insights! On the writing side of things, the optimist in me wonders if AI, used judiciously, could help students get over the intimidating and often steep hill of getting started with a writing project. I spoke with a programmer last week who suggested this is a benefit of using AI when coding and his argument is making me think. As to reading, and using AI to summarize readings, well the hallucination problem remains very real and for now if students rely on it they may end up turning things in with very mistaken arguments. Something to monitor, perhaps.
Something that worries me about how AI might be used to get started: I want my students to be writing things because they care about them and want to think about them. But it's so easy to just get AI to "brainstorm" an idea--which feeds into the writing-as-jumping-through-a-hoop-to-a-product framework. That is not to say it *can't* be used to get started but that this is another area where how it can be used seems at odds with how it is likely to be used. Which leads back to what we're trying to do in the first place, etc., etc.
I agree, and as someone who spent eight years teaching high school students, I think there is something fundamentally important about inhabiting that feeling of "blankness," learning to pay attention to the fact that ideas seem to spring out of nowhere if you allow yourself to get comfortable with the discomfort for long enough. I think it is probably *the* key moment/method by which students learn to trust themselves as thinkers, and believe in themselves as idea-havers.
Agreed that this is worrying. I feel like writing skills help you (1) come up with interesting ideas and (2) articulate them well. But outside of the classroom the former seems like a necessary prerequisite to the latter. If you aren’t able to really develop thoughts on the world happening around you, it’s not much use being able to articulate those lack of thoughts to others. It’s the kind of thing that probably makes life richer even if you don’t share these observations with anyone else and it’s a shame if students will have less practice refining those skills. This was just a small sample of mixed data but yes hopefully building the skill of “starting” can be protected
Just to be provocative, did you ask the students to see the chapter summaries and explanations of certain parts of the readings they didn't understand? Did you ask to look at the questions they asked, responses, and follow ups in the chats? I'm just curious if you evaluated the overall quality of those interactions. From what you are saying, many of the students actually did the reading but used the AI as a companion or helper with sections they wanted to get more out of. Is your point that they should have highlighted those questions and come to office hours instead?
I read—and grade—all of their discussion posts, so I see the way they interact with every part of the book. My point is not that students never do the reading; most of them did do it. My point is that among the students who are at least occasionally using AI, they're rarely using it for generative writing, and more frequently using it for reading assistance. I'd prefer that students read on their own, respond to each prompt to the best of their ability, and then a) come to office hours to discuss if my feedback shows they significantly missed the mark, or b) pay attention to the smaller issues that may have cost them a point or two and keep those in mind as they proceed.
I think the cart is being put before the horse here. Students need assistance with reading . . . yes . . . even at the college level. Read alouds in class are great . . . teacher-created reading guides with oh so specific questions that require evidence from the text (something AI is not great at doing) help as well. Then discussion to follow as a group and then on their own. Students cannot respond to text they do not understand.
Asking students to refrain from using AI as a reading assistance tool is not the same as expecting them to meet the demands of the course with zero assistance. All of the above pedagogical methods exist in my classroom. Understanding of a text is also never an all-or-nothing phenomenon. Students always understand *something* about a text and can begin to work from there. (If most/all students truly understand nothing—and I've never seen this happen—that just means the text is ill-chosen for the course.) My larger goal is here is to keep students attentive to what they understand and what they don't understand, and then use their relationships with me and with each other (rather than AI) to pursue what might be beyond their grasp at any particular moment.
Understood. What I'm wondering is if any student shared with you their actual exchange with AI - what they asked and how it responded. Having read Postman's work, I can see ways in which a curious student could have a productive conversation with AI that pushed them deeper into engaging with and challenging the text. However, I can also see how, if they are doing that prior to a discussion post, their response may be informed more by their conversation with AI than their initial reaction to the text, but I just don't see how that is not going to become the norm for students going forward. It might be worth engaging with AI yourself and probing aspects of Postman's work to see the quality of answers so you can point out to students why you think it is either misleading, superficial, or otherwise not helping them. Are you familiar with Google's Notebook LM tool? You can put up to 300 PDF's, documents, Youtube videos, etc... and use these as the basis for a chat which will comb the material and provide footnotes in its responses taking you precisely to the text that supports its analysis. This is just what is going to be available to future generations of students and something teachers will have to be prepared to discuss in their classes.