Intentional Teaching

AI Writing with Robert Cummings

November 15, 2022 Derek Bruff Episode 2
AI Writing with Robert Cummings
Intentional Teaching
More Info
Intentional Teaching
AI Writing with Robert Cummings
Nov 15, 2022 Episode 2
Derek Bruff

Questions or comments about this episode? Send us a text massage.

A few years ago you could assume that if a student submitted an essay in your class, some human wrote that essay, hopefully the student in question. That’s no longer true, however, as AI-powered writing generators get better and better at producing intelligible text. What are we to do, whether we’re teaching writing or having students use writing to represent their learning? 

On today’s episode of Intentional Teaching, I talk with Robert Cummings, associate professor of writing and rhetoric and executive director of academic innovation at the University of Mississippi. Bob has spent his career exploring what’s coming in terms of teaching and technology, particularly in the field of writing instruction. These days, Bob is collaborating with computer scientists to figure out what role AI technologies might have in writing instruction. 

I reached out to Bob to talk with me about the state of affairs in AI and writing, and we had a wide-ranging conversation that I’m excited to share here on the podcast.

Episode Resources:

Robert Cummings’ faculty page, https://english.olemiss.edu/robert-cummings/ 

OpenAI examples, https://beta.openai.com/examples

OpenAI Playground (account required), https://beta.openai.com/playground 

Fermat Generative AI, https://fermat.ws/ 

Michael Wooldridge, professor of computer science, University of Oxford, https://www.cs.ox.ac.uk/people/michael.wooldridge/

Peter Elbow, emeritus professor of English, University of Massachusetts Amherst, http://peterelbow.com/

Is There a Text in This Class? by Stanley Fish, https://www.hup.harvard.edu/catalog.php?isbn=9780674467262

Emad Mostaque on the Hard Fork podcast, https://www.nytimes.com/2022/10/21/podcasts/generative-ai-is-here-who-should-control-it.html

“Moore’s Law for Everything” by Sam Altman, https://moores.samaltman.com/

Broad Band: The Untold Story of the Women Who Made the Internet by Claire L. Evans, https://amzn.to/3UNC1Ed


Music:

"The Weekend" by chillmore, via Pix

Podcast Links:

Intentional Teaching is sponsored by UPCEA, the online and professional education association.

Subscribe to the Intentional Teaching newsletter: https://derekbruff.ck.page/subscribe

Support Intentional Teaching on Patreon: https://www.patreon.com/intentionalteaching

Find me on LinkedIn, Bluesky, and Mastodon, among other places.

See my website for my "Agile Learning" blog and information about having me speak at your campus or conference.

Show Notes Transcript

Questions or comments about this episode? Send us a text massage.

A few years ago you could assume that if a student submitted an essay in your class, some human wrote that essay, hopefully the student in question. That’s no longer true, however, as AI-powered writing generators get better and better at producing intelligible text. What are we to do, whether we’re teaching writing or having students use writing to represent their learning? 

On today’s episode of Intentional Teaching, I talk with Robert Cummings, associate professor of writing and rhetoric and executive director of academic innovation at the University of Mississippi. Bob has spent his career exploring what’s coming in terms of teaching and technology, particularly in the field of writing instruction. These days, Bob is collaborating with computer scientists to figure out what role AI technologies might have in writing instruction. 

I reached out to Bob to talk with me about the state of affairs in AI and writing, and we had a wide-ranging conversation that I’m excited to share here on the podcast.

Episode Resources:

Robert Cummings’ faculty page, https://english.olemiss.edu/robert-cummings/ 

OpenAI examples, https://beta.openai.com/examples

OpenAI Playground (account required), https://beta.openai.com/playground 

Fermat Generative AI, https://fermat.ws/ 

Michael Wooldridge, professor of computer science, University of Oxford, https://www.cs.ox.ac.uk/people/michael.wooldridge/

Peter Elbow, emeritus professor of English, University of Massachusetts Amherst, http://peterelbow.com/

Is There a Text in This Class? by Stanley Fish, https://www.hup.harvard.edu/catalog.php?isbn=9780674467262

Emad Mostaque on the Hard Fork podcast, https://www.nytimes.com/2022/10/21/podcasts/generative-ai-is-here-who-should-control-it.html

“Moore’s Law for Everything” by Sam Altman, https://moores.samaltman.com/

Broad Band: The Untold Story of the Women Who Made the Internet by Claire L. Evans, https://amzn.to/3UNC1Ed


Music:

"The Weekend" by chillmore, via Pix

Podcast Links:

Intentional Teaching is sponsored by UPCEA, the online and professional education association.

Subscribe to the Intentional Teaching newsletter: https://derekbruff.ck.page/subscribe

Support Intentional Teaching on Patreon: https://www.patreon.com/intentionalteaching

Find me on LinkedIn, Bluesky, and Mastodon, among other places.

See my website for my "Agile Learning" blog and information about having me speak at your campus or conference.

 [00:00:00] Derek Bruff: Welcome to Intentional Teaching, a podcast aimed at educators to help them develop foundational teaching skills and explore new ideas in teaching. I'm your host, Derek Bruff. I hope this podcast helps you be more intentional in how you teach and in how you develop as a teacher over time.

Back in high school, I had a graphing calculator that I used in my calculus course. My teacher, Ms. Steelman, did a great job of teaching with calculators using the technology to help us explore the concepts and techniques in the mathematics. For several decades now, math teachers have had to figure out what role these devices would play in their classrooms and in student learning.

Today, teachers in a variety of disciplines are facing similar questions about the use of a different kind of technology: writing generators powered by artificial intelligence. A few years ago, you could assume that if a student submitted an essay in your class, some human wrote that essay, hopefully the student in question.

That's no longer true, however, as AI powered writing generators get better and better at producing intelligible text. On today's episode of Intentional Teaching, I talk with Robert Cummings, associate Professor of Writing and Rhetoric and Executive Director of Academic Innovation at the University of Mississippi.

Bob has spent his career exploring what's coming in terms of teaching and technology, particularly in the field of writing instruction. For example, he wrote a book on teaching with Wikipedia back in 2009, back when most college instructors were forbidding their students from using Wikipedia. These days, Bob is collaborating with computer scientists to figure out what role AI technologies might have in writing instruction.

I reached out to Bob to talk with me about the state of affairs in AI in writing, and we had a wide ranging conversation that I'm excited to share here on the podcast. Thank you Bob for being on the Intentional Teaching Podcast. I'm really glad to have you as our guest on this episode to talk about artificial intelligence using perhaps some human intelligence to do so.

First Bob, tell us what you do at the University of Mississippi, What's your, what's your title and what's your, what's your job there? 

[00:02:16] Robert Cummings: Sure. Thanks for having me on, Derek. My title at the University of of Mississippi is Executive Director of Academic Innovation. And so I one way to think about it is I try to

help us look around corners and think about what's coming next in terms of teaching and learning. That involves both importing external practices, but also just as much trying to lift up internal practices and help our faculty develop new teaching learning practices. Depending on what their areas of interest are and where they're going next and what they're trying to improve.

[00:02:56] Derek Bruff: Yeah. That's a space I like 

[00:02:58] Robert Cummings: to be in. 

To be in yeah. And it's about helping people figure what's around, I mean, what's around their corner. So if, you know, if there's an aspect of teaching practice that has is necessarily novel, but it might be novel to you That's We're game to help people figure that out too.

[00:03:18] Derek Bruff: So before we continue to look ahead, let's look back just a little bit. Sure. Bob, can you this is a question I like to ask some of my guests. Can you tell us about a time when you realized you wanted to be an educator? 

[00:03:32] Robert Cummings: I think it has to do a little bit with the, I feel like I've never really fit in as a learner.

So I have been someone that's always been disruptive in the classroom. Not intentionally so, but I learned through dialogue. So I learned by listening to somebody or talking with someone else or receiving a new idea, thinking about the new idea that questioning. The new idea, just sort of endlessly iterating over it.

And that's kind of a really difficult, can be a really difficult situation if you're a teacher, and I didn't realize that at first. And so sometimes that would be interpreted in ways that I was being. Intentionally disruptive or trying to hijack the classroom. And so I had to sort of learn to moderate my learning style a little bit.

And so I think in some ways that gave me an appreciation for, for teaching and what challenge, why that would present a challenge because I was really completely unaware of that. And then also I knew that Particularly the way that I had engaged technology was unusual. So I remember. And I think we all have these sort of like moments where a comment comes at you from a teacher and you just, you're, I don't know, it has an outsize impact on you.

And it might not have been something that was intended. So I had a high school teacher and it was in a, a computing class and we had been given this common assignment and it was a coding assignment. And so I knew a little, something about it was BASIC. We were coding in BASIC. And so I knew something about that.

And so they'd given us the, he'd given us this assignment. The class said, Well, they, one person just said, Well, okay, if we need to see another student for help, who should we talk to? And so there was a guy in the class who was really good at math and they said, Well, you know, if, if, if you guys wanna talk to each other about your approaches, Talk to Patrick.

And then he said, But I don't know if you're trying to do something weird or unusual, talk to Bob . It was because I wasn't a great math student, but I really took to programming languages. And so my approach to this problem, he already knew I was coming at it from a completely different angle. Yeah. And so I don't know, you take those comments in an outsize way sometimes when you're a student and the teacher has authority.

Two things. One is I, I, I started to learn that I was sort of disruptive in the way that I learned that made me curious about teaching. And then two, I was really interested in the fact that. As in as, as teachers, we can have more influence and power than we ever realize in ways that offhandedly might shape people's future.

[00:06:50] Derek Bruff: Well, speaking of computer science, let's talk about the term artificial intelligence. . So I, I hear ai that term used in a lot of different contexts. And I think it can mean a lot of different things, but in the context of writing, what does AI mean? And does it include spell check? Cuz I kind of like my spell check.

[00:07:13] Robert Cummings: You know, I got a lot to say all that. So first of all, I'm not a computer scientist, so I'm not really qualified to answer the bigger questions about what a AI is or is not. I have been using a definition that came from Michael Wooldridge, who's a computer scientist at the other Oxford. And he.

Has a definition. It's something along the lines of helping or trying to use Computer technology to perform tasks that human brains or human bodies would otherwise do, which I think is a, this is not a bad one. It kind of puts us in the center of things rather than the, you know, the, the, the Hollywood version of AI and the, you know, Derek, I think the thing that I'm interested in is probably more generally digital writing.

So the, the. It's interesting you said spell check, right? So I think a lot about spell check. I was very interested in spell check as a technology and my current research project is on the use of Google Smart Compose and I'm interested in how Google Smart Compose or other predictive text sentence completion technologies affect the writing process.

So I'm very interested in understanding what a composing brain space looks like. . And how, when, what you're using to be a sort of a tool for expressing yourself starts writing back to you. To me, I see that as an inherently disruptive process and, but I'm very aware that other writers don't feel that way.

and I'm aware that it varies according to the rhetorical situation. So it very much is seen as an asset if you're trying to do something that might be very well defined and repetitive and not involve a lot of original thought or original expression and to a known audience. But if there's other types of more open-ended tasks, I think there's more room for seeing it as interference.

So it's, I don't think that is really settled as a question. So I'm going to be doing research an experiment here in starting in two weeks, a partner with someone in computer science and a great guy assistant professor named Thai Le. And we are gonna get 50 different writing samples.

I'm sorry, a hundred writing samples, 50 using. Google Smart Compose and 50 without and have ask these people to write on a the same prompt and we're gonna be able to compare the form and the process. So we'll compare a writing product and writing process. And due to some of Thai's brilliance, we're gonna be able to create, I think for the first time, as far as I can see some measures that say, right,

derek's writing this sentence, he's gotten an auto complete suggestion. He's gonna have three possible outcomes. He's gonna accept. He's gonna edit it or he's gonna delete it. Okay. And how long does he take on those? Then we'll compare those measures to what Derek's non smart compose enabled peers do.

Compare the structure of the writing and compare the the lexical range that's there to try to look at some ways we think about curiosity and creativity and expression. That's the idea going in that is very staid compared to what's going on right now. This is ancient three year old technology that we're looking at.

Yeah. This is not using the. The a you know, writing models large language models that we have that are really getting everyone's attention. And I'm hopeful that what we're also doing is Ty is doing a lot of coding to enable us to make these measurements. And I'm, I'm hopeful that these tools that we're building Yeah.

Will enable us to take, to experimenting forward into the large language models where there's so much in. I don't know if I answered your question, . 

[00:11:59] Derek Bruff: I think you did, and I think you're, I mean, one of the reasons I, I, I mentioned spell check is because I feel like if, even if those of us who haven't thought much about artificial intelligence and writing, we probably use tools, writing aids at various kinds all the time and may not think of.

Like we think of that robot that tried to take over the world in the Will Smith movie AI, right? Like it's there. There, there's a big gap between those two. But this is future Derek jumping in with a quick correction. That Will Smith movie that I was referencing is actually called I Robot and not AI.

There is a movie called AI but Will Smith is not in it. Spell check, grammar, check autocomplete, right when I am texting my wife to say I'm heading home. Auto complete does a really good job of knowing who I'm talking to and what I'm likely to say. And when I type, heading home is one of the suggestions.

And I have just saved myself three key presses. , right? Yep. And so it's a small example, but it also I think, illustrates. What you said about having a known audience and a targeted genre, right? There are are places where those types of tools can be very effective. And it sounds like your question is, well, if we, if, if we move out from that, you know, texting your wife, you're heading home to other types of writing, are these tools helpful?

Are they disruptive? Do they slow us down? Do they speed us up? Do they help us be more creative or less creative? Because as you note, autocomplete may be an old technology, but there are a lot of other technologies that are coming. And you could ask those same questions about those technologies. 

[00:13:36] Robert Cummings: I agree, and I'm excited to, to, to do this work and to, to move forward with it.

I think that the I agree with exactly everything you're saying. Like I don't want to lose my auto complete functionality on my phone. It's helpful, you know, it's a, using your phone to send a text message is a compromised situation for expression, and yet it's not one that we would want to give up because it gives us immediacy, it gives us this context.

It's just, it's a beautiful communication tool for certain needs and certain messages and certain contexts. So, right, understanding what the tool does and understanding how you want to use it is. Is key. One of the issues there. Two issues I can think of immediately is the technology technologies are evolving so quickly.

It's very difficult for us to understand the full scope of the tool. So it's not like someone's coming to us with a proposition and saying, This is a pipe wrench. It is really good for helping you tighten up a nut around a pipe. You may wanna be careful on plastic pipes though, because it may break something, you know, like you don't, You're not getting that as a proposition.

These are things that just appear in the technology that you're using, and I think it has a lot to do. With the way the people that develop the technologies look at writing. Mm-hmm. , I often hear them talk about writing as a problem to solve. . They also, they, they like to talk about solutions a lot.

And so I don't, you know, I have a bit of a, I, I, That might be unfairly reductive. They're obviously very intelligent people. I just come from a different background that has sort of a respect, a more, maybe a, a respect for the range of writing genres, the range of approaches and the range of goals that that writers might have.

And so, I think that the technology is developing very quickly and then it's never presented in such a way that it's a defined tool and it's also a black box technologically, often for proprietary reasons. 

[00:15:59] Derek Bruff: So there's 

a lot of consternation, I think, in some parts of higher ed about these AI tools that seem to be capable of generating, you know, student essays, right?

We assign our students to write an essay and you know, here's three paragraphs and now we don't know, did the student write this? Did, did they, did they prompt an AI tool to write it? . . , Did they write most of it, but a couple of sentences were enhanced by Google Smart Compose, right? .

Would we have consternation over spell check or auto complete? Maybe not. Right? And so, yes. Is there value in a student who can come up with a really good prompt for an AI generator? Right. That's as I'm learning, that's not trivial. And so maybe that's something that's, that's worth exploring. 

[00:16:42] Robert Cummings: That's an employable skill, right?

[00:16:43] Derek Bruff: Right. Exactly. So, so when it comes. And, and the other thing I wanna acknowledge is that, you know, you are in, in writing and rhetoric, right? The, the teaching of writing is, is your, is your jam . And so as you think about the teaching of writing, what are some things that you're excited to explore, tools that might help students learn to write better, learn to be better writers?

[00:17:08] Robert Cummings: Yeah, I think this is such a great question. I'm glad you asked it. You know, the first thing I think about is I think a lot about Peter Elbow and Peter Elbow's struggle with writer's block. And I think of all the people out there that struggle with writer's block, how much these tools are gonna help them.

Oh yeah. Because it's going to put them seemingly, and I think it's really important to bracket this off, but it's gonna seemingly put them in dialogue. Which might be a mental state, which is, makes it easier to produce text. So for folks that feel bottled up sometimes, well, everybody does from time to time, but for folks who really struggle with getting fluency and getting, getting things moving on the page, I think these, these, a lot of the AI writing generation tools are gonna be really, really good.

I mean, they're gonna change some of their struggles, 

I hope. 

[00:18:07] Derek Bruff: Let me jump in for, so for listeners who haven't used one of these tools or maybe have a vague notion, what might that look like for a student? Can you, can you give a, . A concrete example. , 

[00:18:17] Robert Cummings: so it's important to distinguish, I think when we talk about Google Smart Compose, we call those writing assistive technologies.

But the tools that I'm shifting to talk about now, the ones that are capturing a lot of tension, like Open AI's Playground for instance we call those writing generative technologies or writing generators. And, and that's just because you can put a phrase in and you will get a much larger textual response than you first put in.

And now seemingly it's in response to what you put in and with, from a technical standpoint, it is in response. To what you entered, into, whatever the system is that you're using. But it's not the response that you might assign it. And so that's sort of like the, the card trick. The parlor trick that's happening right now.

I think to users, they're putting words in. It might be something along the lines, along the lines of, is smoking bad for you? Knowing the answer, but just testing the system to see what kind of response it gives you, and it gives you a pretty impressive response. That's not because there was cognition behind that, 

[00:19:25] Derek Bruff: and in that case you might get, you know, four or five paragraphs from the tool.

Explaining why smoking is bad for you. 

[00:19:32] Robert Cummings: Four or five would be a bit of a stretch right now. Okay. You might get a good paragraph on that. Now. You can coach it, right? If you, if you start to use the tool, you can coach it up to give yourself longer responses, and it might be a period of three months, it might be a period of nine months before.

Yes. We'll get an entire essay out of it and I might be misspeaking like the folks that really stay on top of this and know that developments on every 24 to 48 hour basis might be, might tell me, No, you're wrong, Bob. We can get a full paragraph or a full essay out of that question. Is smoking bad for you?

But I think the real thing to think, well, one of the issues to think about is it's not giving you that response cuz it understands what you've entered. It's only giving you that response because it has incredibly fast processing on incredibly deep databases to know what the probability is on what words will follow the words that have been entered.

It's just, it's probability and associations. . There's no cognition, ? And the real problem that we're going to experience as teachers and the real problem we're gonna experience as students, and the real problem we're gonna experience as a public is we've never been in a situation as, at least as I can think of, Where writing has been divorced from thought. So previously, all, every time you encountered writing somewhere behind that writing was thinking. It might not have been great thinking.

Sure. It might have been a stranger, it might have been a group of people, it might have been plagiarism, but at the core, somewhere there was thinking and there no longer is. So we're not, I don't feel like, I don't feel that. As a public and as educators, certainly and as students, we are really prepared to look at to, to separate that out.

And that's something that we're gonna have to be able to do now going forward. We're have to be able to look at a piece of writing and say, Okay, this is a piece of writing. I don't know if there was thought behind it. So it gets into. Really, it's a Stanley Fish territory. I don't know if you've read Stanley Fish.

"Is there a text in this class?" But it was sort of like reader response theory and some of the other literary theory, which I'll need to go back and brush up on now, I suppose. But it, it gets into the idea of where does meaning get assigned? Is it upon the reader reading the text? Is it upon the author writing the text?

And so different people said different things, but it really hasn't felt like it had as much importance as it does now. 

Well, and 

[00:22:19] Derek Bruff: I'll say I've, I've, I've probably without meaning to follow the conversations around AI generated art more than I have, yes, AI generated text, mainly because there's usually a visual that gets shared on Twitter, and so it's, it's kind of easier to, to, to share about on social media.

And you know, my wife is a former high school art teacher, and going to a museum, an art museum with her is a great experience because she has ways of looking at a piece of art that aren't in my toolbox, right? I'm learning how to do this. And I know, you know, when we encounter written text, it's the same way.

There's. There's this reader response and there's lots of different ways to be a reader and to have those responses. And some of them don't depend on the author's intent. And so whether or not there was intent there, maybe this, this was a piece of art generated by AI without human thinking directly related.

I mean, there was human thinking that created the ai, but it's not actually directly informing the created output. And that's maybe happening with text too. And so how do we make meaning of this, even if we have no idea what the intent behind it was? Well, like this is something we as humans have long done, actually.

Right? 

[00:23:36] Robert Cummings: Well, and we also have algorithms that read. So there has been a conversation among, okay. I mean, quote unquote read, right? So there has been a conversation among folks that are computer savvy, that are also writing teachers about teaching algorithms. So like the way to think of audience now is to think about, think about it algorithmically, because anything that's gonna be posted on social media is gonna be picked up on by algorithms.

And so think of, you know, some people would argue, think of your true audience as being the algorithms themselves. Oh. 

[00:24:07] Derek Bruff: Which is, I guess, a, a more complex version of what I've learned on Twitter, which is if you include a link to something off of Twitter, the Twitter algorithm is less likely to show your tweet to a lot of people

So I will, I will describe what I'm doing in one tweet and I will follow up in a second tweet with the. Right. And so I'm, I'm composing based in part on the algorithm Sure. As, as my intended audience, Right. So, 

[00:24:31] Robert Cummings: Well, I can say, I can say from the, well I'll say like straight, straight off about the visual generative business.

It makes me really glad to be on the writing side of things because the the, I listened to Hard Fork this weekend and they had a great episode on you know, generative ai and they had Emad Mostaque, who's Stability AI ceo, and this is not an exact quote that he said that the world has been creatively constipated and we're about to let it poop.

Rainbows. 

Well, I will have to put a link to that interview. Yeah. In the show notes. I feel like there's a lot to unpack there. 

There is. It's a really, it's a really good show. And I just, it's like a show I'm gonna have to listen to a couple times. Yeah. One of the other sources that I've leaned on a bit that I've enjoyed is an essay by Sam Altman, who is a principal at Open ai.

His essay is entitled Moore's Law for Everything. And so it's very interesting read in the sense. He sees the arrival of AI as the way to drop costs for everyone on everything. It was really expansive point of view. But he says, you know, human labor is a, is a factor in costs for everything.

And so long term, what we're talking about is dropping the cost of the input of human labor. I think there's a lot of ground to cover between here and that particular vision . Yeah. Yeah. But it is, it is a surprisingly egalitarian spin on the arrival of ai, which perhaps someone who's a principle at Open AI is benefits from floating as an argument.

But nonetheless, it's a really interesting. So, 

[00:26:45] Derek Bruff: yeah, let's go back to the the writing classroom. Yeah. Right. And, and I'll say also, one of the things, one of the reasons I like to talk to people who know writing pedagogy is because a lot of us who are not in departments of writing and rhetoric, End up teaching writing in one fashion or another.

Sure. Right. We teach writing in our disciplines. We teach first year writing seminars. We, we have students learn through writing. We have students express their learning through writing. Right. So, So there's a lot here. So I've heard of a couple of other things that we might be able to have students at least play with right now.

Right. So, you know what, if I ask my student to use an AI writing generator to. Do a first draft of an essay, and then I ask my student to revise that essay. , right? What, what do you make of that as a, as a teacher of writing? 

[00:27:34] Robert Cummings: Yeah, it's, I kind of move. Yeah, it is very interesting and you kind of jumped around your, your original question there.

So one thing I would say is that the folks here in writing and rhetoric we've got a team of writing teachers who are doing pretty much exactly that. So we've decided that the way to work on this is to invite our students in and to do it together because these tools are gonna be there and they're gonna be using them, whether we invite them in to do it or not.

So that's what the, that's the first thing. And so, yes. . I had, I attended a workshop on the use of AI writing generators back in the summer. And there was oh my gosh, I've gotta go back and look for her name. There was a speaker there who was a also a punk rocker. She was the author of a book called Broad Band is about women in the early internet era.

So it was great. Great title. 

[00:28:35] Derek Bruff: That's a great title.

[00:28:35] Robert Cummings: It is, it is a really good title. 

Yeah. And I, I need to go look her, look up her name. I'm sorry it's escaping right now, but 

[00:28:43] Derek Bruff: This is Future Derek jumping in again. The author that Bob is trying to remember is named Claire L. Evans. She's the author of

Broad Band: The untold story of the women who made the internet 

[00:28:57] Robert Cummings: as a artist, as a, as a, as a musician. She and her collaborator, as her partner. They went and they knew something about coding. So they had, they decided to use AI and the creation of one of their albums, and so they took all of the music and put it into a data base.

And then they took all of their lyrics and they put it into a database, and then they started their composing process. So they got random, seemingly random outputs in their lyrics. And it wasn't like they just took those random outputs and started putting them into a song. They just looked at it. And every once in a while something would come up and they would say, Huh, that's really interesting.

And that would be a kernel of something they would explore. Same thing musically. They would get a lot of different notes that would come out, but then every once in a while they would find something that they would want to explore and that was their launching off part. So I would go back and say, Hey, there.

What works about that for me as a writing teacher is that they had a very deliberative and defined process and they defined how they wanted to use AI in that process. So to me that was like, it was a light bulb moment. I was like, that's how we can use it in the writing classroom. So I think the way forward for writing teachers with, with AI generative tools is to look at a tool figure.

out figure out what the strengths and weaknesses of the tool are and design specific interactions with specific composing purposes. So if it's an invention stage, Sure. Go into open AI or go. You know, Fermat is a tool that writing and rhetoric is using because it has a sort of spatial canvas aspect to it.

So people who like to do, you know, visual layout as they, as part of their writing process, they could really interact with that allows you to see the structure of arguments and generate pro cons. You know, using at the invention, using it at the drafting stage, using it at the revision stage.

There are AI tools that have strengths at finding sources, so you could get to a certain point where you've got your four page essay on smoking is bad for you and you and the AI will help find, help you find sources for this. And then of course you could use any, a number of the AI tools to give you a reader response.

So as long as we're able to sort of define, I think it can work if we define the purpose for the usage of the tool, the technology, and then be very specific about how we want to apply it at that particular stage in a writing process. 

[00:31:41] Derek Bruff: For folks 

out there who are pretty new to this field, is there a tool or two that you might recommend that they use to kind of play around and kind of get their feet wet?

AI writing generation.

[00:31:51] Robert Cummings: I've used 

open AI a good bit. Fermat is out there and I would lift it up as a tool that's, it's interesting and kind of very useful in the, in the writing process. 

[00:32:05] Derek Bruff: Well, thank you Bob. We've gone in a lot of different directions here today. I appreciate the conversation.

You've given our listeners a lot to think about and some tools to play with, and I, I do, I want to encourage our listeners to play with some of these tools. I think this is an area where it would be helpful to get your feet wet a little bit so that you can start maybe to have some conversations with your students about them as well.

[00:32:25] Robert Cummings: Absolutely. Thank you so much. It's been a pleasure to talk to you. 

[00:32:30] Derek Bruff: That was Robert Cummings, Associate Professor of Writing and Rhetoric, and Executive Director of Academic Innovation at the University of Mississippi. I really appreciated how Bob raised some big questions about writing and AI, while also providing some practical advice for instructors wanting to understand the space better.

And while he acknowledges that writing generators and other AI technologies pose some significant challenges to teaching, he's also optimistic about how this tech might actually enhance teaching and learning in the coming years. I'm looking forward to tinkering with Open AI and Fermat, the two tools that Bob recommended to get my head around writing generators and what they can do. I firmly believe that calculators and similar technologies can make our math instruction more effective and more useful for our students.

And I suspect that I'll feel the same about AI writing technologies once I get more used to them. During our interview, Bob dropped in half a dozen or more references to scholars and their work as a good humanist might. If you didn't catch all those footnotes along the way, please check out the show notes for this episode.

I've included lots of Bob's references as well as links to more information about Bob and his work, and please let me know if you have thoughts about the topics Bob and I discussed in our convers. Or if you've experimented with teaching with writing generators, I would love to hear your thoughts on the future of writing instruction.

This episode of Intentional Teaching was produced and edited by me, Derek Bruff. See the show notes for links to my website, the signup form for the Intentional Teaching Newsletter, which goes out most Thursdays, and my Patreon, which helps support the show for just a few bucks a month, you get access to the occasional bonus episode.

Patreon only teaching resource. The archive of past newsletters and a community of intentional educators to chat with. As always, thanks for listening.

Podcasts we love