Episode 2 - Happy Little Writing Assignments

This transcript was created by Otter.ai. Although it does a reasonably good job of transcribing conversations, you'll notice a lot of grammatical inconsistencies if you pay attention. We did only very minimal editing to this transcript, feeling it best to showcase the level of quality achieved by the tool.

:::::

Bill De Herder 0:00
If you're listening, that means you've made it to episode two. And by this time listener, you're probably wondering which Bill I am. Am I bill? Or am I the other bill?

Bill Williamson 0:09
Right? Which one of us is which?

Bill De Herder 0:11
Right? And let me tell you that. That's a good question. And I think the only thing that we can do is wait it out and see what the writers room comes up with. As the season unfolds. I'll remind you that my name is Dr. William Emmet De Herder, the Third. I'm used to vying for Bill supremacy. That's what I've known my whole life. So I'm quite comfortable with not knowing

Bill Williamson 0:41
Well, I suppose if you're gonna pull out the whole formal name, technically, I am. Dr. William James Williamson, Jr. Not the second because I did not pass that along to my son.

Bill De Herder 0:54
Oh, you didn't pass?

Bill Williamson 0:56
Which upset my father at the time? Yeah. But hey, what you're gonna do?

Bill De Herder 1:01
I mean, William Emmet De Herder t he fourth does sound ridiculous. Yeah, yeah.

Bill Williamson 1:07
Sounds very ... efite.

Bill De Herder 1:15
Welcome to bills and bots. I'm Bill.

Bill Williamson 1:17
And I'm Bill,

Bill De Herder 1:18
Be like Bot Ross. Make happy little writing assignments.

Bill Williamson 1:21
That that's the theme for today is making happy little writing assignments. Happy because? Well, we'll get into that.

Bill De Herder 1:41
Right, so for today, we're thinking about some questions that we got from faculty here at Saginaw Valley, we sent a survey out, we asked, you know, what questions do you have about AI working with AI and the implications in the classroom? And some people wrote in with things like how do I write completely AI proof assignments? And how do I write writing assignments that limit AI's ability to meddle in my classroom? What makes a meaningful writing assignment I think is sort of at the core of those questions, right?

Bill Williamson 2:18
Yeah. And although the we pulled these questions, specifically, out of those surveys, I'm in a number of forums, on social media and a variety of social media platforms. So these are really common questions, and they've been questions that have been coming up for a little over a year now. And, you know, one of the things that I kind of assumed a year ago is that by now this conversation would have turned but it really hasn't, there's still there's a lot of anxiety, there's a lot of uncertainty, there's a lot of tension. And actually, at times, there's a lot of anger about AI, and what's going on with it, and the impact that it has on our teaching. But, you know, I think the the arc that we are on today is really about reconsidering what it is that we are doing as teachers, what it is specifically as teachers of writing. And if we come back to the core essence of teaching there in lies the solution to how do you do something that makes AI less of an issue for you.

Bill De Herder 3:21
So the first thing that we have here, is recognizing that there's a tension here, as we try to figure out how to build AI out of our classroom. Absolutely, right. Because we have to recognize that a lot of the stuff that has been circulating around about, you know, what can we just easily and actionably do to make sure that we don't have to worry about AI in our classroom. A lot of that is like Rewinding back to like the 1950s. Right, you know, you're gonna handwrite your essay and regurgitate an argument I've already spoon fed you, and you're not going to have any digital or information literacy built around it necessarily.

Bill Williamson 4:01
By the way, when I say that, we need to get back to our pedagogical roots. That's not actually what I mean. Yeah, we'll get to that. Yeah.

Bill De Herder 4:09
So I think that in, in recent years, we've been making a lot of improvements in making our classrooms more accessible. And a lot of that is facilitated through technology, right? Things like screen readers for an easy example. Right? And so if we start thinking about how we can frustrate a robot, we're also making it more difficult for people to just engage with our classroom content.

Bill Williamson 4:39
Well, when we think about things that we have taken for granted that are that lie, we think of them as lying outside of the teaching moment. So I'll give you a very specific example of job seeking. Job Seeking has been mediated or moderated by robots for a very long time, like our data gets fed into what well into a variety of algorithms, it might be a software package, it might be some sort of, you know, an archive or whatever. And all of those details that are stored within those documents that are uploaded, they've been processed by bots for more than a decade. The fact that it we're just getting around to thinking about these kinds of things. Now, in a more general way, you know, again, we said this before, this is not new, there's there's very little that's that's new here. What is new is that there's an explosion of tools, apps, software, packages, websites, whatever the case may be, that are now claiming to be aI driven. You know, Grammarly, for example, is one of the ones that we talked about before. But if you jump onto any forum online, there are dozens, if not hundreds, of different apps that are available for almost any writing purpose that you can imagine. And all of those things are AI driven. It's not just ChatGPT, that's, that's one, it's not even maybe the most effective one, there's a countless array of things that people are using. And so the point is not to be bot proof. The point is to be pedagogically sound. And if you do those kinds of things, you're gonna have a better experience as a professor and your students have a better experience as students,

Bill De Herder 6:23
right? I feel like a lot of the sustained anxiety in the classroom around AI, comes from the instructor, they aren't aware of how to engage with these systems in practical ways, right? Because we are all still figuring out how this stuff can even be used. Right, let alone like what is what is an efficient, like method of achieving A, B, or C? Right?

Bill Williamson 6:49
Well, and I think that if, if people took the time to play with a handful of tools, and didn't try to use those tools to do the assignments that they're assigning in their class, even the assignments that are in no way shape, or form bot proof, you're gonna see that the results are not that amazing, right? Will they get better, perhaps.

Bill De Herder 7:06
So, we do have some ideas for things that you can do. And it's not, it's not what you might instantly discover, if you start Googling this and reading a lot of the stories that came out last year about right what you can do in your classroom, these things are a little more subtle. And they're also informed by what these systems really kind of drag at, like they don't perform very well in doing these particular kinds of things.

Bill Williamson 7:35
Well, and I'll be clear upfront that when I talk about designing an assignment, that is good meaningful for my class, and that is also along the way going to be something that AI cannot do, at least not to the level of achievement that I want my students to perform at. It doesn't mean doing things like you might if you if you do a little bit of research, find people saying bury text that is colorless. That is that the students will not see the buried in the middle of your paragraph so that when AI takes the text that gets copied and pasted into the system, and tries to reproduce it, suddenly your students are going to be writing essays about Frankenstein, and or chili peppers, when you're not really talking about those things, and that that's your key for catching the cheaters. If you put the energy that you would otherwise waste to trying to catch cheaters into redesigning your pedagogical approaches, you'll have a better class, your students will have a more rewarding experience. And AI is not going to be an issue.

Bill De Herder 8:35
Like Isn't it just a better life to live? Like not being so stressed? About You know, whether or not this person did like if the paper stinks, just mark it down? Right. Right. Right, and make sure that your rubric supports those decisions. Yeah, I think that's, that's important. I also want to point out that, yes, that example you just gave went viral, right?

Bill Williamson 8:56
Oh, I, know.

Bill De Herder 8:57
Well, we also know it just doesn't work. No, because if you feed an entire assignment prompt into one of these machines, I don't know how many data points it actually can latch on to right, but not the whole thing.

Bill Williamson 9:09
No, let's let's take a more positive turn that instead of talking about what not to do, let's talk about the kinds of things that we do want to do. And so one of the first things that you can do in terms of any any assessment that is meaningful in the context of your class, is to make it human focused and student centered to revolve around activities that are live that are in person or that doesn't even have to be in person but just that we're human beings are interacting with human beings and that generates something.

Bill De Herder 9:43
Yeah, and and this is pedagogically rock solid. If you look into the wealth of research around cognitive psychology and how learning happens student to student dialogue is really important for cognitive modeling. So that's, you know, modeling the sort of thought processes that are going inside, going on inside one students head to another student. And that helps both people in that conversation, navigate their way through an emergent communication situation, right? One of my my easy examples to wrap your head around what we mean by a human focused activity would be assigning a podcast episode. But but we all know that there's some stuff out there, that will, it will simulate someone's voice I sent you wanted a text message the other day. Okay, so it's 11 Labs. And they're advertising that you can clone your voice, and create artificial intelligence voiceovers for all of your your videos and stuff, right?

Bill Williamson 10:49
However, consider for a moment that if you want to produce something that way, the fastest way to do it is to write a script. If the students are struggling to write in the first place, why would they take the time to write a script to have an AI approximation of their own voice expressed for them? Why would they just sit down and do it?

Bill De Herder 11:07
Yeah. And while it's true that an LLM could, you know, generate some script Text , you probably won't be up to the level of sophistication than that you would? Yeah,

Bill Williamson 11:19
it would take you less time to do the work than it would be to cheat.

Bill De Herder 11:22
Yeah. So I also want to throw in an additional complication to this. And that's if you're having them do a podcast episode in teams in pairs, or even small groups. You are capturing that student student dialogue, you're capturing that social dimension that preserves the learning process?

Bill Williamson 11:41
Yeah, absolutely. In the exchange, especially you can you can give them questions, you can give them prompts, you can give them things that they can explore together, or have it be part of the assignment that, hey, they've got some direction, they know what, you know what topic they're supposed to be discussing. So the first thing that they do is generate a set of questions for themselves, then they have to consider you know, what would be meaningful here, what would be useful here. And then they have to follow through and do the actual dialogue and record the dialogue. By the time you do all of that, guess what, there's a whole lot of mastery, well, mastery may not be the right word, there's a whole lot of engagement that's going on there. And the ideas are driving the conversation, not some, well, not something that would be less useful.

Bill De Herder 12:29
So another version of this would be incorporating a sort of prewriting activity. So this is before you write your first draft in text form. You try to get students engaging with each other, maybe it's over to microphone, as you know, a sort of sandbox for working through their ideas. It's a pre write. I've done that in podcast form. I've also done it in presentation form. So an example. Last semester, I had my students do pecha kucha presentations using artificial intelligence generated images. So all the images were like wacky and weird, wrong fingers, fingers wrapping around, like hot dogs, all kinds of stuff. And they presented what they wanted to write about to the rest of the class. And we had question and answer time. Yeah, as I've sort of built an in between thing. So it was a pretty fun way for everybody to benefit from the intellectual labor that everybody else was doing. And it was also a way to sort of build in some student a student exchange.

Bill Williamson 13:40
Did you have them record those presentations, even the audio on their own phones that they could transcript to later just out of curiosity?

Bill De Herder 13:47
I did not because I needed to find as many ways as possible this to de escalate the presentation. Got it element, right. tried to keep it chill.

Bill Williamson 13:58
Yeah you were looking to reduce anxiety, not add elements? Yeah. Well, and you know, tied to those same kinds of things I have now for my goodness, I think it actually goes all the way back to the 90s. I have done workshops in class in a workshop is just, it's just code for an assignment that you can do alone or in collaboration with others within the space of a single class period. And it contributes to the greater knowledge for the course. So it might be something from, we're experimenting with an analytical scheme, and they do an analysis together of a website, an app of a piece of writing, whatever the case may be. I often do have them record those dialogues, like on their smartphones, because then they can go back and they don't have to take notes on what they said because they can go back and listen to what they said. But the idea behind that is, is it's all very dialogic, and yes, although there's a piece of writing that comes out of it too. To clear they've created the document or the it's, it's more along the lines of share the highlights of your conversation and then reflect on it briefly, right, where there'd be no point in having ai do any of that, because you'd have to feed so much information into it, that it would be absolutely pointless, you'd waste so much time telling the machine what you talked about for it to spit out a summary write again, you sit down, you write the summary, it takes you 10 minutes. And if you allow them to do it collaboratively, you know, it's all about the exchange, and it values yes, there's an outcome, but the outcome so that I can have some notion of what they talked about. So I can bring that up in class next time. I'm not looking for a super high quality piece of writing. I'm looking for highlights, I'm looking for concrete details. And the moment is the experience, the experience is the value that they're taking from it. And that's cumulative. So do I grade those? Yeah, but they're super low stakes. But it's a way of it's an alternate pathway to getting a participation grade, if you want to put it that way. That's initially how it came about. But the more that I did it, the more I realized, no, this is knowledge building, and it's giving them credit for investing. And that I found really, really valuable. And again, as it turns out, then meaningful assignment makes it less difficult for someone or like there's no incentive to incorporate AI into the process unless it's transcribing your talk as you're doing it, which is absolutely fine.

Bill De Herder 16:27
Yeah. That could be that could be another episode thinking about how to automate some team oriented business oriented tasks.

Bill Williamson 16:35
And in fact, even let's, let's take that a step and say, let's say that they did record it on their phones, and they were using ChatGPT are one of those services that will do a transcript of something and the transcript than the it spits out a summary because I can do that. Okay, cool. That doesn't mean that that's the only thing I have them turn in now I'm going to have them reflect on it. Is that an accurate summary? Is that what really happened? You know, still pull out some of the key moments, what are some of the key things that were said? And then what does this mean to you? So even, even if they could use AI along the way to accomplish some goal, if that's not the only goal, which, you know, there's the clue? No, that should not be the only goal. As soon as there's a higher stake and higher stake is reflection, in this case, more often than not, there's your value that's coming out of it. So they use the tool to give a shortcut? Sure. But they're not shortcutting. Anything that's a learning experience, that's a meaningful learning experience that because the experience is the doing, and then the reflecting on it.

Bill De Herder 16:36
Yeah. So in your example of a project team, I see two things that build AI out of the classroom for those assignments. The first is yes, they're working collaboratively socially in groups. And the second is that they're working on real world problems. Right. Right, probably exist in media formats, like physical format, right, you know, that AI can't grapple with or doesn't have a sort of database on right.

Bill Williamson 18:00
So Oh, absolutely. Yeah. And like one of the things that I have them do in some of my upper level UX courses would be something like a spatial analysis. Well, guess what AI is not going to there's no time soon that it's coming where you can like scan a room, right? And have a doing an analysis of it. Right? What would it do? What would come out of that that might be interesting to see what it would do. But like assuming that there was a an app that was capable of performing such a task, yeah.

Bill De Herder 18:26
So it can't walk into our writing center, and analyze all the chalk that tutors have been scribbling on to the brickwork now,

Bill Williamson 18:36
well and pedagogically speaking, there's, there's a secondary purpose for it is that I often encourage them to shuffle the membership of their in class working groups over the early part of the semester, until they find a crew that they're really comfortable with, and that they trust. So then what have I done, I've constructed on the fly a base for later collaborations. And I have created ways for them to connect with one another and for them to assess one another. You know, if I'm working with Bill, and I think, hey, this dude, he's, he's smart. He's reflective, I want to be on his team. That's fantastic. If instead, I come along, and I'm thinking, Oh, this bill, you never had anything to say when it's time to reflect that he's kind of a slacker. You know, you don't he missed the last meeting. You didn't want him on my team. That's valuable knowledge for them. And they can pick it up without having to really think too hard about judging someone else. It's just part of the fabric of the class.

Bill De Herder 19:30
Yeah, and, and moments of interprofessional skill development. Yeah, are critically important. However, however they come. So learning to work through tensions and challenges. That's important. Yeah.

Bill Williamson 19:44
What are the soft skills that get celebrated year after year, decade after decade, things like talent assessment, things like emotional intelligence, things like figuring out how to work together with others. I mean, those are all just kind of built into those processes.

Bill De Herder 19:58
I'll generalize I'll expand a little bit About your example of project teams. And I'll also say that you know, in a classroom setting, whatever your classroom looks like there are opportunities to sort of break down elements of your assignment, and have students pair up and talk about what's going on, what are the challenges, you know, what progress have you made, and then come back to the group and sort of share some points. And your local friendly neighborhood Writing Center, yeah, is also an excellent resource. Absolutely, you can require visits, you can assign students to have a real conversation with a real actual person, about what's going on with the project. And that sort of required visit slipper, email, whatever you get, can verify that, you know, there is some conversation and therefore some knowledge building

Bill Williamson 20:47
happening. So these are the assignments, some of them that we're talking about here. They all they all have some sort of an outcome, they'll have some sort of a tangible product that you can assign to it or draw from it so that you can, you know, give them a score, if that's part of the point of what you're trying to do here. But it's also like, it's it's building a strategic mindset, as a teacher and building a strategic mindset as a student that values the time in class and that values, colleagues, values, peers, values, what is said? And how you process it, not just showing up? How many times do we see people show up to a class? They're not taking notes? They're not engaged yet? Technically, they were there, but where are they? Where are they? What are they taking away from it, and the more engaged work that we accomplished in the context of a class it which doesn't involve listening to the professor speak, you know, those are things that you're building a very different kind of atmosphere in your classroom, and you're building an atmosphere that, you know, it's you're pulling away from the tools in a variety of ways, making the tools, work and service to the students, not the other way around.

Bill De Herder 22:04
So our next big point is thinking about how you can design an assignment that complicates the modes or mixes modes, that an AI system would have to engage with, so that the human actor is really the one left doing the intellectual labor. Right. We talked a little bit about how podcasting sort of does that. And we talked a little bit about project teams working with real world problems, also kind of does that. Right. I would also point out that if you're dealing with a document that has tables, figures, and has to incorporate that into text, right now, these machines can't just manifest that.

Bill Williamson 22:44
No, no, the tools are not good at processing. It can't can't process an image unless there's all text written for the image. Well, I've talked about in my classes, many times how you could teach an entire course on how to write alt text for an image, because most often it would be Man sitting at a computer with a microphone in front of him, right? Technically, that's correct. But what's the context? What's the meaning there's no

Bill De Herder 23:10
music on, there's not even any function behind it. So that that was the whole thing about how it doesn't understand hands, it can't understand the function and utility have a hand? Well,

Bill Williamson 23:20
and the alt text for a table is the contents of the table, the alt text for a chart, or the contents of the chart. So it's the data behind it. And you can't feed that in in the same way. Like, it's not the way that AI works. If you feed it an Excel spreadsheet, yeah, you're giving it data. That's not the same as giving it an image or giving it a table. Right.

Bill De Herder 23:46
And I'll also point out that finding ways to build digital literacy in your students pays off. Oh, well, yes, yeah. And it rolls forward, we can think of it as knowledge transfer. So if you're learning on one platform, one system, one set of technical problems, it probably rolls forward into the next context, right?

Bill Williamson 24:04
You know, I've talked to a lot of professors over the years who count on their students knowing more than them about certain tools that they're implementing in class. And it might be something like Microsoft Word it might be. It might be something like Excel. Or it might be something that's much more esoteric and specific to the discipline. But what's always amazed me and sometimes amused me is that those same professors sometimes are afraid to admit when their students knowledge surpasses theirs. I thrive on it. And I tell my students, if your knowledge does not surpass mine, eventually on some of these things, you aren't doing it right. And I want to know what you've learned that goes beyond what I've taught you. Because then you're probably going to teach me something about it. So it's an opportunity to engage it's an opportunity to put students in a position of expertise and authority. And that doesn't mean that I lose mine. It's not the Zero some kind of thing, everybody can have some expertise. And the more accomplished they feel, imagine if you were the student, and you realize that your knowledge had surpassed your professors in a specific way about some tools that you're using in class, or you're not going to feel just really damn intelligent. Yeah. And really accomplished and be high fiving yourself, well, that's an amazing thing to give them.

Bill De Herder 25:22
And that. So that brings us around to one of our final big points here. Engage with meaning, and making new knowledge absolutely in crafting your assignments, right. And I'll seize on that making your knowledge point to connect to what you just said. So I think about a lot of the assignments that I design, and I purposefully construct them so that students have to theorize and apply theorizing, right theorization. What is the word? To an emergent context? Yeah, you know, I talk about them as artifacts, you find some sort of cultural artifact, you apply some theory we've been talking about, in class, you make that abstract theory connect to that concrete, emergent situation, connect those two dots. For me, that is building new knowledge.

Bill Williamson 26:19
Yeah. So I'm about to launch a class is a class that I teach relatively often for the university. For those of you who are in the know, and disciplines, we think of the generic technical writing class, or technical communication class, we often call it in fact, the service course. But it is it's the tech writing course, for non majors. And the way that I teach the class, following up on what we're talking about here is that they work in small groups early in the semester, to learn analytical methods. So they're going to pick an object of study, so something like an app or a website that's meaningful to them professionally. And they'll conduct three different kinds of analyses using these different methods of analysis. And then from each of those moments, well, each time that's engaged with a new method, they're learning the method. What does it look like? What do they get from it? What data does it generate? How do I interpret that data? And what does that data mean, in terms of my understanding of this thing that I'm examining, they do three different variations on that they're learning different things about their object of study. And then they have to figure out how to write that up. Well guess what I've just described, if there is something that's AI proof, I just did it. Because it's all built on engagement. It's all built on interpersonal exchanges. It's all built on data that they are generating in the process from original research. Yeah, they have to figure out what to do with that. All of those things, AI can't do it can't do it. Well, at all it it generates a big report, it's probably the most robust report that many of them have written to the point where they get into that class. But it's all driven by their own expertise, their own knowledge. And they are acquiring knowledge, they are creating knowledge. And it is meaningful in a variety of professional contexts. And so again, if I want to break down the you know, what happens in a writing assignment and create something that a bot cannot perform that suit? That's exactly the kind of thing that we're talking about, is that the only way to do it, no, but it's a way that does work.

Bill De Herder 28:20
So I just finished an English to 12 class where I got the entire class to perform a cultural study. So they, they picked a cultural artifact, and they picked an ideology. And they tried to demonstrate how that ideology is playing out in the tensions and contradictions at work around that artifact that they picked some people, they picked muscle car culture, other people, they picked all sorts of things. I know, an LLM can't do that. Because these machines don't have any sense of meaning, these artificial neural networks, they can categorize things in terms of like what a noun is, what a verb is, they can categorize things pretty simplistically in terms of like positive or negative tone, they don't have access to the kinds of maps of meaning that we as humans, who have lived in a culture, with competing histories of struggle, have have created and we hold in our mind and our cultures, right. So I have used the example before of the apple. So this is what I often do in my class, I write the word apple on the board, and I say, Okay, let's look at the map of meaning just around this, right? And I start like drawing a web. And you know, in one direction, we have a crappy image of an apple that I draw on the board. And then we have branching off of that the idea of a piece of fruit coming from an apple tree, right? But we also have other things buried deeper in our culture. The Apple also is a reference to elementary education, right? There's the stereotype of leaving a shiny red apple on your elementary school teacher's desk. For some reason, right? And then I trace even deeper down into that to the story from the Bible about the tree of knowledge. So just in that one example, we can see that, okay, this is our moment in our culture, where this is one map, there are other maps of Apple, from other cultures that can also be drawn here. And they are coexisting in, in potentially the same communicative act when you use the word apple, right? I mean, Sheen can't mess around with any of that, because it's constantly shifting in relation to all this other stuff is the result of struggle, histories of struggle, and I'm saying plural histories because, you know, different cultures have different struggles. Right?

Bill Williamson 30:43
And I've heard people say in response to that, well, you can you can ask for, like, what does apple mean? Or what is the meaning of the word Apple beyond? Its denotation? That is beyond a dictionary definition. But here's the problem with that, yes, there's some data out there, that is connotation that is the cultural state or value or, or, you know, usage of the word apple. But the computer cannot distinguish between connotation and denotation. It's all the same data as far as that's concerned. So those things overlap and intertwine. It's not like the machine cannot consciously separate it. So as soon as he started talking about the meaning that we make about things, the computer is at a loss, right? It does. It does formula, it does linear, it does. denotation, but it does not do cultural value.

Bill De Herder 31:34
Yeah. So unless I'm holding up my phone, this, this is an audio format show. Unless there is already in existence, an essay written about the ideology embedded within a touch phone, right smartphone, right, the machine won't be able to just automatically generate an essay, where it convincingly argues that this phone is designed to interpolate us as subjects of pleasure. Right, right. Right. very abstract concept. very concrete example I'm holding in my hands. Right. Right. That's, that's what we're talking about when we say, you know, engage with meaning in the assignment, how do you design a rubric that is testing them on, on their ability to create new knowledge and grapple with meaning, right? Well,

Bill Williamson 32:25
and I'll take it down a predictable pathway question that I just feel that from someone a little while ago, and they were saying, Well, you know, if I have my students go to the library, and research a topic that is new to them, are they not making knowledge? Sure, they're expanding their knowledge by exploring what other people have written or said about something. But it is not the same as engaging in research directly themselves. It is a form of research, it is reading research, and it is at knowledge acquisition. Absolutely. But they're not forming any sense of expertise. They don't own anything. They're, they're borrowing it from others in the moment. Now, fast forward, if I start as an undergrad in a class, and I explore something that's new to me, and wow, is that exciting stuff. And I turned it into a career arc. And I continue to learn about it. I read more, I do more I talk to people more over time, I could become an actual expert in that field of study. Absolutely. But only then, am I able to really contribute new meaningful knowledge. Otherwise, I'm just riffing on published stuff and maybe forming some sort of semi uninformed opinion. It's not the same as what we're talking about in terms of creating new meaning or, or, or gathering original data.

Bill De Herder 33:53
I'm going to talk about things to avoid here. Yeah, I've got a shortlist, I would avoid something that only summarizes a pretty accessible source. Like if it is like, really, really challenging reading what the summary comes out as with an L is probably going to make no sense. But if it's already accessible reading, and you just asked to summarize a case study or something. Yeah, it's probably going to be relatively easily automated with the push of a button. If you're just asking students to deliver very, very broad information, like the history of country music, right, right. Something that is just so easily doable, and therefore summarized level by one of these machines, don't even bother analyzing materials that have already been written upon. And all of those essays are just floating out there on the internet and its database, I would avoid it. So I don't know if that means we have to really get creative about teaching Shakespeare now. But probably even assigning something where they're just producing a sort of reflection that is composed of like a bunch of unverifiable details about themselves, right, you know, um, That's easily just sort of fabricated. And then if you're asking them to critique something, but their critiques building on critiques that already exist plenty of fully out there on the internet, yeah, of course, the machine will be able to refund those as well. Right. And that's sort of the illusion that the machine can like engage with meaning comes in. So all

Bill Williamson 35:18
of those things that you just described, are the kinds of assignments that I could consider using in my class, but not as the end game. And I think that's the key is when we talk about something having pedagogical meaning we talk about what's the what's the end game, what's the what's the value at the at the outcome of it. And summary is a valuable strategy. But if it's the end, it's really easy to reproduce with a computer. But if you do a summary as a stepping stone to something else, okay, that's when as soon as you move beyond the summary of the summary, you have acquired it, you have you have had ai do it for you. There's nothing wrong with that. Now, what do you do with that, right? As soon as you have another step, or maybe seven other steps, you're depending on the complexity of the activity that you're doing, you know, that's when you are devaluing or decreasing the value that AI contributes to that moment, it becomes a tool that is part of the process, rather than the end in and of itself. And if you can accumulate moments that are meaning building that are built from those stepping stones, that's a valuable thing to do. But it means reconsidering what you're actually getting in terms of pedagogical outcome. Remember, we said that at the very beginning, it's kind of bringing it full circle? What is it that you're trying to accomplish with the assignment? What do you want your students to know? What do you want your students to be able to do? What do you want your students to be able to take from that moment and apply in another context? As soon as you can answer those questions, you can look to see how to the stepping stones, contribute to strategic processes or methods that then contribute to the accumulation of knowledge, meaning strategy, when you're giving those bigger outcomes. That's what we're talking about, where now you've given your students something that they can apply in another context, not something that they can just do in the moment. You know,

Bill De Herder 37:13
if, if you just have a screwdriver in your toolbox? Yeah, that's pretty limiting. Right? So if your assignment is just summary, and that that was the entire aside, right? What is it? What is, what is the real utility of it? Right? If you put it in combination with other things, greater possibilities emerge,

Bill Williamson 37:32
right, and you could turn a screwdriver around and use it as an ineffective hammer, you can try to cut something with the side of it. And it's a really blunt knife. But that adaptation is not necessarily creative, you know, and it's not necessarily all that useful.

Bill De Herder 37:52
Okay, so those were those were like, the things to avoid, at least exclusively. Yeah, right. Here's some things that you could lean into, we already gave you, you know, some human focused stuff, some some stuff about genre, and so on, and so forth. But, you know, you could think about being very specific about the sources that you want students to use and how you want write them to use it, so that you can watch to see how it is or is not actually being incorporated into a paragraph. Right, right. Because we know that that's where these machines fail miserably. They, they can grapple with bibliographic information and really, really limited ways. They don't necessarily know what quotes mean. And they also don't know when to stop, when to stop referring. So it'll just start re mixing stuff that was in an actual article and sort of make something that wasn't in that article. And then finally, my, my other idea is to you could lean into like more discipline specific genre conventions, right? Particularly around new knowledge creation. So you can really teach them the proper format of creating a lab report in your specific disciplines, right. And then arrange the assignment so that they have to take a theory from class and apply it in a lab or something.

Bill Williamson 39:25
Well, so what have we talked about here today, making happy little writing assignments that are about cultivating experiences cultivating engagement, and building meaning building meaningful engagements in the context of classroom with ideas with people, and so on. And the more immersive the experience is, the more valuable the end product, whether it's a piece of writing or a podcast or something like that, you know, the the more human connected the assignment, the more the more oriented around Knowledge Building, the more likely it is that AI is not going to have a significant role or at least not a meaningful role, other than as a crutch or a tool. And, again, that's going to make the classes more engaging for students and ultimately more valuable, they can have something that can transport instead of something that they just regurgitated.

Bill De Herder 40:18
So I guess to sort of like contextualize that philosophy a little bit more, I'll go ahead and read this sample syllabus language that we hashed out Yeah, months ago. And I've been using in my syllabus, but I think this sort of describes my my sort of perspective here. So in my syllabus, currently, it says, You will not be policed around your use of smart writing tools in this class. But be aware they will be of limited use. Assignments will ask you to craft innovative interventions, apply theoretical concepts to emergent situations, work with others, create discipline specific documents, and consider ethics and strategy. Smart writing tools are ill suited to meet the demands of these communication challenges. While I encourage you to find ways to incorporate smart tools toward more efficient writing processes, wielding such tools effectively requires writing skills this class aims to build, you are ultimately responsible for the quality and validity of your work, and best practices towards students success. So I don't have to get anxious about whether or not I'm accusing somebody of cheating. When they're using AI, I can just mark them down. Because I designed something that was really challenging for an ally, you've

Bill Williamson 41:36
produced an opportunity that is not about a simple goal or a simple gain. In the moment, you're emphasizing really building knowledge. And when we talk about what is the value of a college education, it's, it's not to get a degree. It is not necessarily specifically to get a job, those are things to the degree might be a pathway to the job. You know, it's you're accumulating experiences that are meaningful, and professionals contexts. And so the more that you keep that kind of stuff in mind as you're designing a course, even a general education course, if you're not contributing knowledge, what are you doing?

Bill De Herder 42:21
Alright, so that's all that we have today. Thank you for joining us on this wonderful sequel episode of Bills'n'Bots. Stay tuned to find out if there's going to be another one.

Transcribed by https://otter.ai