Teaching Writing in an Age of Generative AI with Jen Roberts
Have you ever worried that your students are using an artificial intelligence (AI) program to cheat? To do homework? To do something else dishonest? Join me for this conversation with Jen Roberts as we talk about not only how to prevent academic dishonesty using AI, but actually how to use it in the classroom to actually enhance student writing!
Jen is a Nationally Board Certified (NBCT) high school English teacher, with over 25 years of experience teaching English Language Arts and Social Science in grades 7-12. She has been teaching English at Point Loma High School since 2006, where she still enjoys the daily pleasure of working with 140+ students. In 2008 Jen became one of the first teachers in her district to have 1:1 laptops for her students. She pioneered best practices in digital pedagogy, online classroom management, and supporting student empowerment with technology. In 2011 Jen became a Google for Education Certified Innovator. Her co-authored book Power Up: Making the Shift to 1:1 Teaching and Learning with Diana Neebe. Jen has worked as adjunct faculty at the University of San Diego supporting pre-service teachers and has mentored dozens of student teachers, presented at conferences locally, nationally, and virtually. CUE named Jen their Outstanding Educator in 2022. She serves her district as a 1:1 Ambassador, and on the Technology Advisory Committee. Her current interests include standards based grading practices and teaching students about generative AI.
Jen’s blog Lit and Tech
Jen Roberts and Diana Neebe Power Up: Making the Shift to 1:1 Teaching and Learning book website
Episode Transcript
Scott Lee: Want to learn more about social emotional learning? Check out our partner website SEL Resource at selresource.com
Greetings friends and colleagues, welcome to The Thoughtful Teacher Podcast the professional educator’s thought partner-a service of Oncourse Education Solutions. I am Scott Lee. In today’s episode we’ll share a conversation I had with Jen Roberts. Jen has been an English and social science teacher for over 25 years and currently is a National Board Certified English Teacher; teaching at Point Loma High School in California. She is the co-author with Diana Neebe of Power Up: Making the Shift to 1:1 Teaching and Learning and a Google for Education Certified Innovator since 2011. She has worked with pre-service educators as an adjunct professor and has mentored dozens of student teachers during her career. Her interests include literacy instruction, and leveraging technology to make her teaching more efficient and effective. She blogs at LitandTech.com.
Scott Lee: Welcome, Jen, to the Thoughtful Teacher Podcast.
Jen Roberts: Thanks, Scott. Happy to be here.
Scott Lee: So let's start, with learning a little bit about, your current role and how you became interested in education
Jen Roberts: technology. All right. My current day job is I teach 9th and 12th graders how to read and write better, which is the way I like to say I'm an English teacher.
I'm in San Diego, California, and I teach at Point Loma High School. I've been there for about 17 years. And before that, I spent 10 years teaching middle school. I became interested in educational technology. All the way back to the earliest days. I mean, there was my student teaching program in the 19 nineties emphasize technology.
Jen Roberts: And I chose that. That was a specific cohort that I chose. So I must have known that I was interested in educational technology, even 30 years ago. , and I'm not sure why I knew that, but I knew that. We were right. Coding in HTML as student teachers, building websites, when websites were a very new thing still, then I just kind of kept a finger on it and stayed in touch and my professors in that program were amazing people who led local organizations about educational technology and I've stayed in touch with those organizations.
And then, of course, that all really ramped up in 2008 when, Some folks in my district talked me into piloting one to one laptops in my classroom and I said, look, I've got an infant and a five year old. I do not have time for this. And they said, please. And they kind of talked me into it. And I said, okay, but I don't know how this is going to go.
And it was amazing. My classroom was paperless within almost a year. At the time, it was back in the days when they were rationing paper and they would give us a case of paper for the whole year. And I got to March and realized I'd only used one ream. And my colleagues were like, how did you do that?
And can I have some, educational technology to me has been the grease in my classroom since 2008. The first machines they gave me for the pilot were Linux laptops that didn't have Word. I'm like, I'm an English teacher. What am I going to do with that Word? And they said, there's this thing called Google Docs.
And I went to the training and I took my student teacher at the time with me and I remember looking at her and said, you are never emailing me a lesson plan ever again. We're going to do it all on Google Docs. And we did, we were off and running. And I think it really helped that I had that instant collaboration partner there to work through it with me.
Because I was one of the early folks to have one to one laptops, I learned things earlier, and I figured out solutions to problems that, have since been solved in other ways, but there was no learning management system available to me. So I invented my own learning management system using blogger and Google drive. Like that was what I bootstrapped together. And then I started going to conferences and presenting about, Hey, there's this thing called Google forums and here's how awesome it is.
And then Twitter came along and I got on that and learned about the Google Teacher Academy, applied for that. And they accepted me. So I've been a Google for Education Certified Innovator since 2011.
, then in 2015, well, actually it was probably 2012 when I was at a conference, teaching conference with a few colleagues and one of my colleagues had, Diana Nibi, amazing teacher out of, Silicon Valley.
Her school had one to one laptops, and that was that was brand new for them. So she and I spent this whole conference just talking about what we were doing with digital pedagogy. And finally, another friend, our good friend, Heather Latimer leaned in and said, you two should just write a book about this.
And we went, No, we're not doing that. But we kept talking. And we wrote a proposal to Stenhouse at the time, and they accepted it. The book that came out in June of 2015. And, since then, I just keep going to conferences and presenting and talking about what's going on in my classroom and.
People ask me all the time, like, you know, when are you going to leave the classroom and go become a resource teacher or something like that? And I'm like, I love the classroom. I love doing this every day with kids. I think that, showing them how this stuff works is what's going to make their lives better.
Scott Lee: One of the things that, you've written about is your, academic honesty unit. . I really thought that was interesting. So, , tell us first off, how you did that, how, how did you structure it? I was about to ask why, and pretty obvious why any teacher would want to teach that.
Tell us about the structure and then what were some things that the students learned, from the academic honesty unit in particular.
Jen Roberts: So since 2011, I've been on a really strong, 9th grade teaching team, and one of our colleagues on the 9th grade teaching team at the time was also teaching some seniors, and he brought us a short article he was reading with his seniors.
That was about academic honesty. It was all of three or four paragraphs, about the zeitgeist of academic honesty and how that was affecting kids. And we read, that and went, this would be , a nice piece to read with our ninth graders.
We had noticed, and again, this was probably in, I want to say we started this around 2014, 2015. , we know that we had noticed by then that academic honesty was sometimes an issue for our ninth graders. We were all using Google drive. They were all submitting all their assignments .
I don't even know if it was Google Classroom yet, but, it, it meant that when a student turned in something that we were a little suspicious of, that was back in the day when they were usually getting their suspicious content from another student, not from the internet necessarily. But , my colleague would email me a paragraph and say, can you search some of these sentences in your Google Drive?
And I would put those sentences into my Google Drive. And sure enough, one of my students papers would come up and we would figure out that my student had given her student. This work and that they turned it in for her. So that was really great because , we were smart enough to know that, that we could search our own Google drives and figure out that kids returning and work in other places.
And that wasn't very common. The frequency of students who are willfully and intentionally cheating, , is actually, in my experience, I think, fairly low. It's under 5%. It might even be under 2%. But when you have 140 students, that means it's three or four kids per assignment.
That, that adds up.
Scott Lee: Yeah. Well, the interesting thing I think about is usually when I have, Had a student that I've realized was being dishonest and turned in something that wasn't theirs the the excuse And I think the kid was telling me the truth was well, I didn't have time to do it myself the kids , are rationalizing it in a real world kind of way.
Jen Roberts: Absolutely. As an English teacher, I have made an extensive personal study of plagiarism and academic honesty. There's several key reasons why students cheat. First of all, kids take risks, teenagers take risks. It's literally hardwired into their brain. That they take risks.
It's part of what helps teenagers eventually separate from their home family. It's part of what, literally evolutionary, trait is when you are this age, you will try new things. And, , sometimes that means jumping off a roof into a swimming pool. And sometimes that means turning in work that's not your own.
And the Venn diagram of those kids is kind of a circle. So risk taking is normal for teenagers, and we get a new crop of teenagers every year. And so even though we taught the previous crop of teenagers that we would know if they cheated, this new group doesn't know that, right? They have not passed that knowledge down to each other.
And so the risk taking is number one. They are willing to take a risk. We know that biologically, their brains, favor reward over risk, right? So the reward of getting away with it is weighed way more heavily in their decision process than the risk of getting caught. And for some of them, it's just a challenge, like, Ooh, let's see what I can get away with, right?
It's just being sneaky. It's just like, I think I'm smarter than that teacher and they won't catch me, it's that adversarial relationship I work really hard to break that down with my students, but there's always the new kid who comes in and still carries over that and doesn't doesn't know yet.
, so there's that then there's, , the panic of I've been absent for many, many, many, many, many days and I still have to turn this work in. I'm not going to graduate. I'm not going to pass the class. I'm going to panic and I'm going to turn in something. I don't understand the assignment. I don't know what to do.
So I'm going to result to borrowing a paper from a friend or using chat GPT or using something else, copying, just blatantly copying, pasting off the internet even. And I think I'll get away with it. . And they're just desperate. That's just, that's not risk taking, that's just desperation.
And then there are the kids who legitimately have other time crunches, and there are the kids who just don't take, the, English is never going to be their priority, and they don't think it's important, and, and so they don't have any investment in learning it or doing it right yet. They may eventually in life realize that that was a mistake, .
, there's always that kid. But again, the sum total of all of those examples is still somewhere between 3 and 5 percent of kids. So it's not a lot of kids, but it's always a different 3 to 5 percent of kids often too , there's different kids try it on different things. Our academic honesty unit was structured so that first we surveyed our students. And we surveyed them at the time of year that they're also being surveyed about all kinds of other things. They're taking the California healthy kids survey and they're taking a survey about, what classes they want to take next year.
And they're doing a survey about something else for the district. And they're getting surveyed to death in the spring and. So we just throw in there this survey about academic honesty and we don't even , tell them that it comes from us. And we ask them, about whether or not they've cheated on things and what kinds of things they've cheated on.
And we collect all that data and we give the same survey to almost 500 ninth graders. So we get to see what they bring back That's really informative for us. Like, we learn a lot. And the vast majority of them are cheating on things like homework, especially math homework, where they can borrow a kid's paper and write down what the other kid did, or, , they're, they're texting each other answers to things.
They're collaborating. There's a lot of collaboration that goes on in high school. Some of it's authorized, some of it's unauthorized, but, in many ways, Academic success in high school is not as much about your academic skills as it is about your social skills, right? , you're leveraging who you know to and who you can get to help you with things so that you can be successful.
So after we survey them, then we start to feed them articles about other examples of cheating, other places where Oh my gosh, this group of kids here had this group text going on about their math class and this group of kids here did this and this group of kids here did that. We give them these other articles.
We revamp those articles every few years to keep them current, and to make them relevant and we show them, , pictures of kids in China who have built a fake arm into their jacket so they can have a phone under their table to take a test with, right? , third arm cheating, that's called.
So we show them all this and then. We feed them back their own data. We give them back the aggregate data from their survey and say, well, here's what's going on at our school. And they kind of freeze because they're like, Oh my gosh, that's exactly what we just read about. Like, so what's the difference between those kids in Georgia who got caught with that, math ring going on, their group text.
And and what's happening here? Well, nothing like you're doing the exact same thing that they got busted for. So, they they start to read that. And then they have to write a position paper on whether or not cheating is an issue issue at our school. And what should, if anything, should be done about that.
And we wait and we do this unit in the spring. And part of the reason we do it in the spring is because we want to at least make it appear to be responsive. We are responding to what we've seen you doing. Hey, we've noticed that these same kinds of things are an issue, right? Also, we need our ninth graders to trust us when we bring them this unit, that we're not playing gotcha, that we're trying to say like, Hey, this is something we want, we think you need to know before you get to 10th grade.
Whereas if we do it in the fall. It's, it's prescriptive. If we do it in the fall, it's, it's us saying, darn it, you kids, you should not cheat, right? But when we do it in the spring, it's educational. And it feels more like we're saying, hey, we've noticed. So that means that they are more prepared to listen, more prepared to hear it.
More ready for it. And then when they get to 10th grade, their teachers know they've had that unit, especially, and, and they know what's going on now it does not stop the kids. who are really desperate. It does not stop the kids who are risk takers. I had one ninth grader once I busted him for cheating on something and it was pretty obvious to me.
And I said, here's what I noticed was going on here. This was 15 years ago. I don't remember what he was doing, but I do remember that he looked at me and said, nobody else has ever caught that. I went, all I did was read your work. Um, , there is, I think, Somewhere out there, this idea that kids have that teachers aren't really reading their stuff, and that as long as they turn in something, they'll get credit for it, and they're flabbergasted when I catch them cheating just by reading it.
It's so obvious to me when I read it, like, wow, I remember these exact same mistakes in another paper, that I read an hour ago. They don't, they don't think that's possible. They don't realize that my brain can do that because their brain probably wouldn't do that. They think they wouldn't catch it.
So I wouldn't catch it, but I catch it.
Scott Lee: . So, tell us a little bit more about, the, generative AI programs like chat GPT. I think it's safe to say that many, if not most teachers, have concerns, about, programs like that primarily around cheating, but maybe in other ways as well.
What do you think those concerns are? And then we'll talk a little bit more about some strategies you have. For that.
Jen Roberts: So first of all, I will come out with a disclaimer that I am an English teacher, not a cybersecurity expert. I'm not a programmer. , I wouldn't say that I'm an expert in large language models or how they, compete with each other to come up with better answers.
There's a whole absolute science, multiple PhD field that goes into that. I know there are multiple large language models that have been trained on the average of the internet, right? They have been fed. I think ChatGPT 3 had 1. 75 billion parameters and ChatGPT 4 has, , it's measured in the trillions.
They're working on ChatGPT 5 behind the scenes, yesterday, well, yesterday in my world, and so it's, , February, but, Open AI released, , the website for something called Sora, which is their text to video option, right? Like this is an area that is accelerating by leaps and bounds, a month in the world of AI is like a year and any other kind of development.
But what these models do, especially what, what English teachers are concerned of, I mean, they can make art, they can make music, they can make apparently now video, , and they can make And that's the thing that got English teachers to. To pay attention. , chat, GPT by OpenAI was released in November of 2022.
I saw my first student work submitted using AI in December of 2022. And what these things do is they generate text, right? Everybody probably knows that by now. You put in a prompt, you ask it a question, you tell it what you need, and it spits out text. And that is just, it is predictive text. And so I like to say it is the average of the internet.
, it absolutely contains bias. If you ask for a story about a doctor, the doctor will almost always be male. If you ask for a picture of someone reading, the person reading will almost always be female. Because that is The examples that it has seen, , I think somewhere along the line, it was trained with a whole photo set of abnormal human anatomy. And that's why all the hands are messed up when you look at images. But, . When a student says to ChatGPT hey, can you write my essay? ChatGPT says, sure, what do you want me to write about? I love to write.
I'll write whatever you want. But it's bland and it's formulaic and it has some, some signal words that teachers have learned to look for. It is, it is spitting back the text. And the problem is that when it first came out, I guess to problematize this, it was presented to kids by other kids on the Internet, Snapchat, what have you, that this will write your paper for you.
All you have to do is paste to the prompt and then take what you get out and turn it in. That sounded great to a lot of kids. And that's as far as they went with it. They have never been taught and they haven't figured out for themselves that ChatGPT can help them in ways that are both ethical and helpful without being cheating, without being plagiarism, without taking all their ideas out of their hands, without doing the writing for them.
And the problem is students are not sophisticated enough yet as writers to recognize when the output is not helpful. Or they're, just being lazy enough to copy and paste the result either because they didn't bother to read it or because they didn't understand what they read. And they assumed that because it had some big words in it that it would be great and get them an A. The major concern of teachers is, Oh, my gosh, kids are going to use this to cheat. Well, that's what they said when Google existed, right? Oh, my gosh, kids are going to Google the answers. And Google Translate had world language teachers up in arms. We still teach world language classes.
And math teachers protested calculators in the 70s. Right. So, This is not a new thing. Education will survive. I'd like to definitely say that loud and clear. Kids absolutely still need us. But it's possible that what they need us for is to teach them how to work with these models in ways that are actually helpful and and move things along faster.
AI language models can do things for students and for teachers that, Carry some of the cognitive load at both the, the low end and the high end to make our, our work faster and more efficient. Right? Mm-Hmm. , let me spend my time where it matters most. Let my students spend their time where it matters most.
Yeah. And, and I can talk about some of the examples .
Scott Lee: Well, and that, and that's what I was just about to ask. I know you have. Specifically taught your ninth grade students how to use chat GPT in particular, in completing assignments. Tell us about that as a, as an
Jen Roberts: example. So I have modeled for my students how I would use chat GPT to help them with their assignments.
I want to make it really clear that if you're just. a teacher listening to this, , the usage, guidelines for ChatGPT say that students between 13 and 18 have to have parent permission to have an account. And I am not willing to go through, , the hurdle of, making sure every single one of my 140 students has parent permission to use ChatGPT.
So I never require kids to use this. But I do show them some ways they can use it ethically. And this started last year when I had just taken my ninth graders to, to do an activity we've done before. There's a, there's a, my blog is litintech. com and there's a post on there about how to do this assignment both with and without Chat GPT.
And
Scott Lee: we'll have a link to that on our website.
Jen Roberts: So, so there's a post about comparative analysis with art, and our school has a fantastic art program and a wonderful art teacher, Sean Wells is amazing. And, he sets up a gallery every spring with some of the fabulous examples of student art that have been created throughout the year.
And every spring I take my students down there with clipboards and, and a graphic organizer. And they choose two pieces of student art and they write down what they think, is the symbolism in the art and they describe the art and they write down what they think the message of the art is. And we talk about how that could be like almost anything you think it is, right?
, we're, we're taking some of the same concepts we've applied to literature and applying it to art, right? So symbolism, mood, theme, those message, those kinds of things. And they're allowed to take a picture of the artwork, but only for their own personal use. And they learn about the artist's copyright and that kind of thing.
The artist has given you permission to photograph this so that you can write your essay, but not so you can put it on your Instagram. We talk about that. And then we usually just before ChatGPT, we just came back up to the classroom and I would give them a suggested outline and they would work on writing their comparative analysis .
piece about art. And this was sort of like a little precursor to a larger comparative analysis they were going to do about literature later, but art was the way to warm them up to that. So in the spring of 23, I took them down there and we looked at the art and we wrote our notes down and then we came back up to the classroom and I said, so, , can ChatGPT write this essay for you? And they all kind of froze because I hadn't really mentioned ChatGPT to them before then, not to my ninth graders.
And they all kind of like paused and they were afraid to say anything. And, and finally one kid said, I think they were afraid to admit they knew what ChatGPT was that, that was the problem. And finally one kid said, well, no, because it hasn't seen the art. I said, oh yeah, I guess that's true.
Does that mean it can't help you at all? And they were like, no, it hasn't seen the art. So I opened ChatGPT, and I told it what the assignment was, and I said, Can you give me an outline to help me write this essay comparing these two pieces of student art? And ChatGPT happily spit out an outline. And my students and I looked that over, and we read it carefully.
I modeled for them how you have to read this carefully. And I said, Would this help you if I gave you this outline? Now, keep in mind, in years past, I just gave them an outline. This time, I let them think that ChatGPT gave them an outline, right? So they said, Yeah, that would actually help. Because this is their first time writing a comparative analysis essay, so I gave them an outline all the time.
And now I'm just asking ChatGPT to do it. So we looked it over and decided it would be helpful, so I copied and pasted that onto Canvas where they could find it. And they went to work, started working on their essays. This does take place over several days. On day one, we go look at the art.
On day two, we start doing the writing. On day three, we do a peer review, etc. And they've been writing for a while. When I said, you know, your essays due Friday. And today's Tuesday. So what do you need to do every day between now and Friday to make sure you have it done on time? And they're like, I don't know.
They're ninth graders. They have no idea. I said, let's ask ChatGPT so we asked ChatGPT to give us a time management plan for writing this essay. And it did a beautiful job breaking down what they should do every day.
As a writing teacher, I really liked what ChatGPT came up with for them because it had, sort of broken it down into, a planning day, a writing day, and a review and editing day, which was great. It wasn't write a little bit each day, it was, it actually had that time built in for reviewing and editing.
And I said, would this be helpful? And they said, yeah. And for my students who have IEPs, that kind of thing is super helpful. If you're a student with executive functioning issues or things like that, having someone break it down for you and say, do this on Monday, do this on Tuesday, do this on Wednesday.
Um, was, was really, really, they don't know how to break down a task or they have struggled with breaking down a task. So that was really useful to them. So they did a little writing for a while, and then finally I said, Is anybody, done with their first paragraph? Want some feedback? And a couple kids raised their hands.
I said, Are you okay with getting feedback from ChatGPT? And they said, Yeah, that'd be cool. Let's do that. So I went to their Google Doc, and I grabbed the paragraph they had written already, and I pasted it into ChatGPT. And I said, Please do not rewrite this for me, but give me feedback on how I could improve it.
I gave it the paragraph, and it spit back a bunch of feedback, and we read the feedback together, and I said, would that be helpful feedback? And some of it was kind of generic, and some of it was very specific to what the student had written and very helpful. And the, the, the kid who I'd gotten the paragraph from was like, yeah, can you, can you take that feedback and put it on my Google Docs so I can work with it?
I'm like, absolutely. And other students were noticing this and saying, actually, what it told him would help me. I'm like, great. That's awesome. Right. I said, if you ask ChatGPT for feedback, and then you take that feedback and use it to revise your writing, is that cheating. And they said, no, and I said, I agree.
That's not that's not cheating. That's just like asking someone else to read your writing. It's just like asking a parent to read your writing. It's just like asking a peer to read your writing or asking a teacher to read your writing. So, with that in mind. They, realized that there were now three ways they could use ChatGPT that would be helpful to them that wouldn't be academically dishonest, right?
They could ask for an outline. They could ask for feedback. They could ask for a time management plan. And part of what I wanted them to learn in that lesson is you do have to read the results, right? You have to read what comes back. My ninth graders haven't seen ChatGPT fail yet, but my seniors certainly have.
So the seniors are reading an essay about the value of life from 2002. Amanda Ripley is part of our value of life unit and I asked chat GPT to summarize it for us. And it got it entirely wrong.
We didn't give it the article. We just said, are you familiar with this article by Amanda Ripley? And it said, yes. And it spit back a summary. And my students read that and went, Oh my gosh, that's all bogus. And I said, yeah, it is. But it was my bad. I didn't tell it what year she wrote the article. Let's try that.
No, we're looking for the one she wrote in 2002 in Time Magazine. And it spit back another summary that was also still mostly wrong and I looked at them and said, you know, if you were a freshman in college. And you were trying to get away from reading something your professor assigned to you and use ask ChatGPT just to give you a summary of it.
And then you use that. What would happen? And they said, we would fail. I'm like, exactly right. You can't trust the AI. You have to be it. the human brain who reads the answer and says, is this actually accurate or not? And if you're not in a position to judge the accuracy of it, you're not in a position to use it.
And I think, for that group of seniors, that was actually a pretty powerful lesson. Because I've seen the examples on the internet where, the student asks the ChatGPT to write their essay on Midsummer Night's Dream. And the first sentence of the essay that the student turned in is as an AI language model, I cannot do that.
My students haven't done that, but, but it, it's happened, right? Kids copy and paste blindly without reading it. And that's why I tell other English teachers that ChatGPT might just be the best tool we've ever had for teaching critical reading. Because you have to read the output and you have to read it critically and you have to know what you're talking about.
If you're not already an expert in your field, it's going to give you stuff that's not helpful. And you won't
Scott Lee: know it. Exactly. Yes, , it's very interesting. I've, I've played around with it a little bit just out of curiosity. And, it's told me some things that were, some things that were useful and some things that absolutely were not true.
So, you've covered a lot about how you use generative AI, in the writing process. And one of the things that I noticed, your, January 6th of 2024, preserving authenticity in student writing, in the age of generative AI.
And the first thing I thought of when I'm reading that was, well, you know, a lot of what you're doing and a lot of what you're talking about doing is best practice for teaching writing anyway. I mean, one of the worst things that teachers do and, being a social studies teacher, I have cringed when I've seen colleagues do it is.
Not really follow the best practice on teaching writing assume the English teacher has taught my students how to teach writing and I just let them loose on a on an essay That they have to do totally out of class And what I noticed was, you're using it to teach the writing process itself so, what are your comments about that or what are your thoughts on, your, approach to the writing process is and why it works so well with chat GPT or generative AI.
Does that make sense?
Jen Roberts: Okay. So one of the, the evolutions of modern education is. We use writing now to find out what's happening in someone else's brain. , it, it would be better if we could use speaking, probably. , go back to Aristotle and Socrates and things like that. But we I have gotten to the point where we have established that universal literacy is the goal, and therefore, if you want to know what someone thinks about something or what they understand about something, you can ask them to write about it.
And so writing has become a method that is, at least until recently, was a fairly effective way of finding out, when you have 140 or 180 students and you want to know what they all know, you can ask them to write something. And that's not invalid, right?
Writing is something that everyone does because it's how we show what's happening in our brains, right?
It's become our modern method of communication. We write emails. We write text messages. We'd rather write each other text messages and call each other even. So writing is, is absolutely critical. And it's, it's not inaccurate or unfair to say that history teachers or science teachers ask students to write things and they're not teaching them how to write.
I mean, yes, all teachers need to be teachers of literacy and all teachers need to expect literacy from their students. But
It is still a writing teacher or an English teacher's place to be actually teaching that writing process, right? We expect students to write longer process pieces that they work on over time, whereas a science teacher or history teacher might be expecting them to write something that they write in a shorter period of time.
And, and they're trying to assess the student's understanding of a historical period or assess their understanding of a scientific concept. And that's a different kind of writing than when we ask them to, you know, build an argument or write, write a narrative or something like that.
Scott Lee: And I think there's different jobs there.
Yeah. And I think that's kind of what I was getting at it. I don't want to diss a lot of my colleagues, history teacher colleagues, it may have sounded a little unfair. One of the things , is there are some differences that a history teacher is doing, or a science teacher is doing, but a lot of times, I've witnessed.
Not really a good job of teaching enough about how that writing should be in history content or in science content.
Jen Roberts: Teaching the literacy of your subject area is an important skill for all content area teachers to be doing. I'll address though for English teachers, right? If you're, if you're assigning writing and not teaching writing, there's a difference for that.
We all know that. And probably most of the people listening to your podcast are people who are on, fall on the side of teaching writing rather than assigning writing. So I'm probably preaching to the choir here. But we know that the writing process is something that students can sometimes be reluctant to engage in or have trouble accessing.
What I was trying to address in that blog post you're referring to is the things we can do as teachers to give students the agency to do the authentic writing that they, that they can and should be doing, right? So when students have choices, and that doesn't mean they have unlimited choice, But it means they have some choice.
Right? You're, you're going to be writing, a narrative, but you get to choose the person that you interview for your narrative. Right? You're going to be writing, an essay, but you get to choose the topic you're arguing about. Right? So when students get to have a little bit of ownership of what's going on there, that's the first piece of giving them something to hold on to.
The next thing is really just writing over time. I have all my students do all their writing in a Google Doc and the beautiful thing about Google Docs is the version history, right? I can see what they wrote and when. And I'm sure Microsoft Word probably has something similar to that now at this point, too.
Jen Roberts: I just don't, I don't live in that environment anymore. So when you can see what they're writing over time, you can judge the authenticity of that writing. Now that doesn't mean you have to go look at the version history of every single piece of student work, but it's one tool. At this point, I have high achieving students who are legitimately worried about being accused of using AI, even when they're not. And I teach them that the best tool you have to defend yourself against that accusation is your version history. You're going to be able to show how your work developed over time.
Now, if it developed because it just looks like you were typing it in without ever stopping and without ever revising, I'm still going to believe you used ChatGPT to write it. I had a senior try that. Because risk taking, like, and again, I think he would also jump off a roof. My point is I'm looking at that over time.
I'm also looking at my students writing while they're working on it. So just because. It's not due till Friday doesn't mean I have to wait till Friday to look at it and looking at it doesn't mean I have to give extensive feedback to every kid all those Google Docs are in one folder. I can open up the preview button and flick through it with my arrow keys just to see where we are, who's making progress, who's not making progress.
I can regroup my students like, oh, these are the people who haven't started yet. Let's put them all at one table group today and see I will make sure that they all get started right so that they have something to work with because when students. Panic. That's when they resort to copying from a friend or, going to ChatGPT or copying off the Internet, right?
So let me make sure they are empowered to start because self efficacy I have found is one of the greatest predictors of whether or not students will cheat when they don't feel like they know what to do when they don't know how to start when they don't know how to do it. That's when they're going to panic and turn to some other tool.
But when they feel like when they have, maybe you need an outline, maybe you need a sentence frame, maybe you need to see what somebody else has written so that you see a model, not just the model I gave you, but the model from another student. Maybe you need to have a conversation with a classmate about what you're going to write.
That's going to help you more than me, so what are the tools I can give to those students who are struggling early on before they get to Thursday night, and my essays aren't always due on Friday, but it's the week makes a good example here. , I don't want them to get to the 11th hour and find out that they have nothing and they panic.
So, writing over time bringing your writing to a writing group. Has been, really helpful because, my students don't want me to know that they cheated with AI, but they really don't want their teammates or their, their classmates to know, right? They're not willing. They're, they may be willing to take a risk that I won't really read it.
Right, but they know that they can't bring it to a writing group and read it out loud to their table group and, and not have one of their classmates say, dude, I don't think you wrote that. Right? So. That has been has been huge. So when I finally get that student paper, and it's ready for me to grade, I'm looking at that Google Doc, and I've got 3 or 4 different, maybe 5, 5 different indicators that I'm looking at authentic writing, plus 6, my own teacher brain, right?
There's a revision history Chrome extension that has a little bar across the top of their Google Doc that tells me how many edits they made, how many hours they worked on it, how many deletions they did, how many copy paste they did. There's a Brisk Chrome extension that I can do a little, I can run a little thing that says detect AI writing.
I don't necessarily believe it, but it can be an interesting indicator. When I've already decided that it was written by AI, and Brisk confirms that, that's kind of handy. There are the comments from their peers going down the side of the, of the, of the dock. There's what, what I'm seeing in my own brain.
And if I really want to, there's the version history, right? So I can pretty clearly and within just a few seconds say, yes. David wrote this. I'm sure David wrote this. There's no doubt in my mind that David wrote this. Or, hmm, some of these things are a little suspicious. Let me go investigate further, right?
So for those four or five, , one to three percent, for those essays, I think I need to investigate further. Doing that is also pretty easy, easy because of those tools that are built in. And over time, my students have learned this right that they're that they're going to get called out that they're going to get, awareness of it.
And every kid is different. And the way I handle plagiarism with every kid is different. And I fully support them using AI and all the ways I taught them. Use it for an outline. Use it for feedback. Use it for a time management plan, etc, etc. But if you give me an essay that you didn't write, we're gonna have words.
And they know that. And, sometimes I call them out quietly, and sometimes I call them out a little more publicly. It depends on the kid. One, the first time I ever saw AI in my own classroom was, like I said, December of 22. And, I came in the next day, and the kids had submitted work in a Google form, and it was, , like, paragraph, paragraph, paragraph, little teeny paragraphs, and then this kid had written, like, practically an essay for a short answer question.
And it was ChatGPT. And it was right after one of the first detectors had been released. And I came in, I said, you know, I found this cool thing. Let's play with it. And we pasted in some kids work and we pasted in some other kids work. And it's a hundred percent human, a hundred percent human. And then we get to the one that I know is already is chat GPT.
And we paste that in as a, you know, a hundred percent AI. And I just, I watched the kid across the room, just drop his head. Right? Like I didn't even have to say it was his, I wasn't showing any names. I was just putting in some examples of student work. Yeah., And he got it like, Oh, right.
With, with my freshman this year, it was hysterical because, I had a student turn in a paragraph that had been written by ChatGPT. I took that paragraph and I put it on a piece of paper then I went and found two other paragraphs from that same assignment from the year before.
So what I ended up giving my ninth graders was a piece of paper that had three paragraphs on it and at the top it said, how would you score these? And I told them this was about standards based grading. I told them, we're just trying to figure out what's a three, what's a four, what's a two, because I'm using a one to four grade scale.
And so they dive in with their partner. I passed out 18 copies, so they had to share. And I said, you need to read this. Tell me which one would score him how. And they spent about 10 minutes reading through all of them really carefully and writing numbers down.
And then they want to know if they're right. And I said, actually, I just want you to tell me which one was written by ChatGPT. But don't say it out loud, just write it down, like mark it on your paper. Within 10 seconds, every team in the room had identified the middle paragraph as being written by ChatGPT.
Every single one. And I walked around and I just looked at that, but I didn't say anything. And finally they want to know, again, they want to know if they're right. I go and I said, , you all got it. You all nailed it. You all knew it was the second paragraph. Yeah. Good job. You don't need AI detectors.
You found it. And they're all happy and they're all proud of themselves. And finally I said, if you can tell, I can tell. And they went, Oh, right. And the best part was. the kid who had turned in the middle paragraph was in the room. And I told them that. I said, by the way, the person who submitted that ChatGPT answer is here right now.
And they don't even know that that's their paragraph that they submitted. And pretty soon they're going to go open their Google doc and double check. And they're going to realize that that was their work and that I've already busted them and they don't even know it. Oh my gosh. To ninth graders, this was like candy, they thought this was hysterical and.
Sure enough, the kid who had done that, and trust me, I, I know my children. I, I knew this would be okay to do, right? You don't do this to the vulnerable child, right? This is, this kid was fine. He came up to me after class and, , and admitted he's a Miss Roberts. I'm so sorry. It was me. You were right.
I shouldn't have done that. , I, I very much apologize. I'll never do it again. And he has been going around telling other kids that this happened to him. And when, when someone mentions ChatGPT in our class, he's like, don't do what I did. He is the poster boy for don't use chat GPT at your work.
It became kind of like. team building thing, actually., yes, kids are going to use it. Yes. There are lots of ways that teachers can promote authentic writing for kids, right? When you give them choices, when you give them options, when you give them a chance to write over time, when you give them adequate support, right?
I mean, remember kids cheat because they don't think they can do it on their own, but when you raise their self efficacy and give them the scaffolding they need. To do that writing independently, they don't feel the need to go try something else.
Scott Lee: So what are some final thoughts or advice that you would have for teachers?
Jen Roberts: , I'm, I'm very aware that there is a lot of moral panic about chat GPT and students using AI for writing and things like that. And. I want to emphasize that absolutely I want my students to know how to write.
I want them to know how to build a strong argument. I want them to know how to use evidence. I want them to know how to do lateral reading and, qualify the sources they're looking at, those kinds of things. But I also do firmly believe that the advantage goes to students with access, the students who are using AI now to make their writing better.
Even if it might, in an academic world, be considered cheating. They are still learning a skill that's going to serve them better later in life because outside of academic writing outside of academic spaces, being able to write effectively and communicate effectively is the goal and how you got to that effective communication as long as it says what you want it to say in the way you want to say it.
It's not cheating. If I ask AI to help me write a grant and I get that grant, which I did, then that's not cheating. If I ask AI to help me rephrase, an, email to a parent so that it sounds more the way I want it to sound, that's not cheating. I've communicated better. If my students ask AI to help them write an email to their landlord or, something like that, they are not cheating.
They are effectively communicating. And the goal is to grow effective communicators. And therefore, I think that the, the skill we need to be teaching our students is Yes, how to use act AI academically and honestly in academic environments, but more importantly, I think also how to use it for communication that says what they want to say, and therefore, I want to emphasize the critical reading aspect of it.
Does this really say what you want it to say? Or does it just sound good? Does this communicate what you hope to communicate? Or does it just sound good? So no. I worry about the places that are blocking chat GPT, or they're, trying to prevent kids from using it because I think that's inequitable.
I do see this very much as an equity issue. I want students to have access to it because even if they're using it in a moment that is academically dishonest, in the long run, they are learning skills that are going to serve them better in life. Just like kids who used to Google the answers to questions and courses they didn't care about.
Now those students know how to effectively search for things, right? So, these are skills that may be, may be replacing some of the skills we've traditionally taught in schools. Over the long term, but I didn't think that I could graduate a class of seniors last year without teaching them much about chat GPT.
I thought that would be malpractice. I needed to send them off into the world, knowing what this tool was and how they could use it and how they couldn't use it and how it could help them and what could go wrong with that and how to navigate it effectively. And I think that's our job to prepare students for their lives and what's going to happen after they leave us.
And so I, when I teach. academic skills with ChatGPT I'm teaching for transfer, not just teaching for, my particular class and my particular subject and my particular assignment.
Well, thank you for joining us today Jen on the thoughtful teacher podcast
Jen Roberts: Now, this has been a lot of fun, Scott.
Scott Lee: The book is power up making the shift to one on one teaching and learning,
With Diana Neebe. , and the blog is lit and tech, which is lit and tech. com.
The Thoughtful Teacher Podcast is brought to you as a service of Oncourse Education Solutions-if you would like to learn more about how we help schools and youth organizations strengthen learning cultures and implement restorative interventions, please visit our website w w w dot oncoursesolutions.net.
This has been episode 1 of the 2024 season. If you enjoy this podcast, please tell your friends and colleagues about us, in person and on social media. We also greatly appreciate positive reviews on your favorite podcast app. The Thoughtful Teacher Podcast is a production of Oncourse Education Solutions LLC, Scott Lee executive producer, in partnership with Chattanooga Podcast Studios. We encourage diverse opinions, however, opinions expressed do not necessarily reflect the views of producer, partners or underwriters. Guest was not compensated for appearance, nor did guest pay to appear. Episode notes, links and transcripts are available at our website w w w dot thoughtfulteacherpodcast dot com. Theme music is composed and performed by Audio Coffee. Please follow me on social media, my handle on Instagram and Twitter is @drrscottlee and on Mastodon @drrscottlee@universedon.com
