AI-Aware Teaching in Political Science with Eric Loepp

The ebook version of my new book, The Norton Guide to AI-Aware Teaching, co-authored with Annette Vee and Marc Watkins, is out! To celebrate its release I’ve interviewed a few of the faculty whose AI-aware teaching we highlight in the book, and I’m excited to share another one of those conversations here on the podcast today.

Eric Loepp is a professor of political science at the University of Wisconsin at Whitewater and co-director of Whitewater’s Center for the Advancement of Teaching, Learning, Scholarship, and Technology. Eric is passionate about teaching political science—something that comes through loud and clear in our interview. He’s also had to figure out a variety of AI-aware teaching practices in the last few years in response to generative AI’s impacts on learning—and on our political discourse. In our conversation, Eric shares some very concrete and AI-aware assignments and learning activities that will, I think inspire you as you adapt to AI in your teaching. 

Episode Resources

Eric Loepp’s website

Transcript

Eric Loepp 0:34

And the goal is not that everyone changes their minds, but that they walk out and appreciate that in a lot of cases it is a matter of trade-offs. And if you want to be able to articulately and effectively communicate your worldview, it really helps to understand the worldviews that would challenge yours because then you can form your responses, clarify your positions, and better make the case for them in the future.

Derek Bruff 0:58

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. The ebook version of my new book, The Norton Guide to AI Aware Teaching, co-authored with Annette Vee and Marc Watkins, is now out. To celebrate its release, I've interviewed a few of the faculty whose AI aware teaching we highlight in the book. And I'm excited to share another one of those conversations here on the podcast today.

Derek Bruff 1:33

Before I do, however, I have a request for you, dear listener. This is episode 95 of Intentional Teaching, which means that the podcast's 100th episode is coming up soon. And I would love your help making it a special one. Here's my ask. What's one idea or example or question or quote from an intentional teaching episode that you keep thinking about? Let me know via email or record a short voice memo on your phone and send that to me. I'm planning to make the 100th episode a kind of a clip show where I share excerpts from past episodes featuring powerful insights and your comments about why those insights have stuck with you. Your deadline for this assignment is September 15th. Okay, now on to today's interview.

Derek Bruff 2:21

Eric Loepp is a professor of political science at the University of Wisconsin at Whitewater. Eric is passionate about teaching political science, something that comes through loud and clear in our interview. He's also had to figure out a variety of AI-aware teaching practices in the last few years in response to generative AI's impacts, both on learning and on our political discourse. In our conversation, Eric shares some very concrete examples of assignments and learning activities that will, I think, inspire you as you adapt to AI in your teaching.

Derek Bruff 2:56

Eric, thanks so much for being on Intentional Teaching. I'm excited to have you on the show and to get to know you and your teaching a little bit and how you're navigating this AI world that we're in. Thanks for being here.

Eric Loepp 3:06

My pleasure. Looking forward to it.

Derek Bruff 3:08

Nice. So I'll start with my usual opening question. Can you tell us about a time when you realized you wanted to be an educator?

Eric Loepp 3:17

It was some time in college. I helped facilitate some programs for high school students and found myself really enjoying the process of leading a group in discussion and doing a reasonable job of connecting with people in that age group. I suppose it helped that I was only a few years older than them at the time, go figure. And although I don't get confused for a college student anymore, I hope I can make up for that with some of the experience I've picked up along the way. And, you know, really, I just love being a part of people's aha moments, you know, especially in political science where everyone's trying to find their place in our political system. And, you know, maybe it's a student finally figuring out why these two different political parties are different. And, you know, now I can be more confident in my vote choice. Sometimes it's students saying, uh, I really want to jump into these family discussions at the holidays, but I don't have all the background knowledge yet. So anything I can do to help people find their place in the political world is wonderful. And I'm lucky to have gotten to do it for so long.

Derek Bruff 4:13

Yeah.

Eric Loepp 4:14

What got you on that path of political science? You know, it's it's funny. We don't, I don't come up from a family of politicians or academics or even from a state. I'm I'm from Oregon. It's not really a state that has a lot of political history the way maybe Pennsylvania, Virginia do. Um, so I had one of those formative moments in my youth where the third grade class got on the bus, went to the Capitol, took a tour, saw the chambers, met their legislator, um, and I got a real kick out of it. I remember the the teacher uh invited a few of the students who had a kind of seemed to show a lot of interest. They invited us to, I forget if we went back the next day or stuck around after, but I remember being asked if I wanted to spend some more time there because I looked like I was enjoying it. Um, and uh sure enough, it turned out to be my thing. And what I found I really love about political science and politics and government is you know, this is how we figure out how we are going to govern ourselves. You know, how should the world be set up? Is the way that things work now the way they ought to work, or can we make our system better? Right. Especially as we're coming up, I'm I'm getting a little sentimental coming up on the 250th thing, how cool is it that we have you know managed to make it this long? And yet we're also surrounded by some big questions about what the next 250 should look like and what we might do to make our system even better. So I love that conversation and hopefully inspiring students to want to get into that conversation too.

Derek Bruff 5:44

That's great. That's great. I love that. What do you think are some of the ways that your teaching experiences prior to Chat GPT's arrival in late 2022 prepared you to navigate this new thing in our teaching landscape called AI?

Eric Loepp 5:58

So some of the challenges posed by AAI came along in kind of in different forms several years before ChatGPT in 2022. Uh, you know, if you talk to political scientists, you will hear terms like fake news and alternative facts and some of these terms that became pretty prominent both in the popular lexicon and within our discipline in the 2010s. So, in some ways, the politics of a decade ago really helped position us to be ready to teach in a world where evaluating the veracity of claims is paramount and preparing students to come to claims with an open but skeptical mind is paramount. Um, and you know, even before those terms were common, we've long dealt in my field with issues like leading questions in polling or misleading claims by politicians. So we've spent a lot of time in this kind of trust but verify mindset. Now, I don't know if that makes us better agents or critics of AI, but it was certainly helpful for me to have a number of years working with that mindset when the substance shifted from interrogating political claims to interrogating kind of the nascent claims from the first couple of generations of AI.

Derek Bruff 7:06

Okay. I find that a lot of people, and I say people because sometimes it's students, but it's like family members and randos on Facebook, like tend to put a lot of trust in what comes out of their favorite AI chatbot. And I'm wondering, do you have do you have a theory as to like like why why people do that? Um is it do we just trust lots of things we shouldn't trust? Was there something about AI that made it seem more reliable to some people? Do you have any thoughts on on that, especially kind of how it situates with kind of other other ways where we have to verify claims?

Eric Loepp 7:41

Yeah, you know, it's interesting. In political science, we will study things like anchoring bias uh in ballot design, where the first name on the ballot we kind of compare everything else to it. And uh there could very well be some parallels in the AI space. Uh as you know, there's a lot of literature and commentary out there about the tendency for anything that looks authoritative, anything that looks formal, um, or really anything that just comes off as confident. I mean, that's where there's a great parallel with political science. You know, the old joke of if you just say something and you're confident enough, it just starts to sound true, whether it is or not. And in my world, when you couple that with the tendency for people to seek out information that confirms what they already believe, um, I can absolutely understand how people come to maybe it's a specific AI tool that was the first one they learned, or the one they have a subscription to, or just the easiest one to use. And it's comforting, right? There's some really interesting literature that came out about AI pandering and the tendency for AI tools to kind of pat you on the back, no matter what you say, to kind of pat you on the back and say, that's a great question, you know, whether it's a good question or not. So it it really kind of taps into our wanting to be validated and feel a sense of belonging, even if we ask a completely ridiculous or nonsensical question. You know, if you ask the human, they might say, What on earth are you talking about? But AI may say something like, you know, that's something that confuses a lot of people. I really appreciate you giving me the opportunity to provide some clarity. So I can definitely see that uh relationship between a tool that provides some of that validation and someone's uh sense of or need to belong, to feel confident in their own answers and and to not feel like they're asking a silly question.

Derek Bruff 9:31

Yeah, yeah, that makes sense. And part of that is I think how some of these chatbots are designed by their creators to to to engage the audience in certain ways that are maybe less about truth verification and more about keeping them talking, right? Um that's fascinating. Okay, so we could go down that rabbit trail for a while, but let's talk about your teaching here in 2026. What is AI aware teaching look like for you now?

Eric Loepp 9:58

So I'm pretty skeptical of our ability to effectively police AI through submission scanning technology, browser lockdown tools and the like. Uh now, there may be some contexts in which these tools work really well, so I don't want to reject their place entirely. But in my space, I've focused much more on revamping how we challenge students to tackle our course material and how we assess their learning. So, one example for me is I've selectively deployed structured AI critique exercises where students are tasked with critically analyzing AI output. Now, a lot of people do this, and we need to be really careful because this is something that can easily be done poorly. Uh, you know, students need to have the right existing base of knowledge in order to evaluate output well, right? We don't want to have our students spending most of their time reacting. We want them to be proactive as well. We want them to be writers, not copy editors. Uh, we need to make sure we're aligning our assessment goals with our exercises and that we don't even accidentally oversell the capacity for AI to deliver quality and accurate information. So we do need to be careful. All that said, though, I think these kinds of activities, if they're structured, can be really helpful, especially in a formative context. So, for example, when I teach about political groups and affiliations, I will have students peruse data from the US Census, Pew, Gallup, some of the highly reputable resources of information to familiarize themselves with party coalitions and who tends to belong to what group and just kind of get a sense of the portrait of America. Now I've found that it's a relatively few number of students who love pouring over data tables. Right. Not necessarily all that much fun. But I've also started class by putting an AI-generated image on the board following a prompt I gave AI saying something like, draw for me an image of a group of Democrats or a group of Republicans. And I just throw it up on the screen, and we start class by having students diagnose whether and to what degree that picture just feels wrong. I mean, I call it a kind of a just feels wrong test. We're not doing formal work yet, we're not using the data separately. We just like gut reaction. Does it feel like this is accurate? Um, and most of the time the students will say, I like some of it, but something kind of feels off. So it's a great way to kick off our class. Students take that image, they say it doesn't quite feel right. I don't like this, I don't like that.

Derek Bruff 12:25

And then at that point, they're kind of comparing the image to whatever preconceived notions they have about the party in question. Okay. Exactly. Which could be accurate or could be not, but it's it's it's whatever's in their head.

Eric Loepp 12:38

Right. I mean, I remember the first few generations of AI images where people would have a right-hand thumb on their left hand and there were just a lot of technical inaccuracies. A lot of that is resolved now. But if we're looking at, again, say the demographic profile of a group or the type of instruments in people's hands, right? Like a like a farmer's pitchfork, right, or something like that. Um, it's a fun way to get students to just kind of think about what is the state of reality and to what extent does this image reflect that reality. Um, but again, the trick is not to stop there and just say, is this right or wrong? The trick is to say, okay, now let's bring in that rigorous data that we studied before or that you prepared before class. I want to provide a structure where they're evaluating those images for accuracy, reasoning through the strengths and the weaknesses of the output. And then the critical part, owning the process of correcting it using that evidence. Are there factual inaccuracies? There may be. There may not be. What I find is more likely to be the case, though, is where does the image need nuance? What is AI kind of mostly getting right or kind of leaning in the right direction, but could use some additional detail? That's where I think there's a lot of value in coupling some of the traditional rigorous database work that we've done with AI output that can maybe make life more fun or bring data to life, or just give students an interesting place to start. And that doesn't even include the potential lessons in AI literacy and hallucinations or all that kind of stuff.

Derek Bruff 14:10

I gotcha. Okay. So you're you're kind of having the students triangulate their their internal sense of what this should look like with the AI created image and the data. Um, how do your students respond to this?

Eric Loepp 14:23

Most of them really enjoy the process. You know, in in the olden days, right, every everything pre-AI, right? It just feels like a completely different world. In the before times, you know, we would work primarily with the data, and it would be a lot of graphs and Excel, and you know, that's all fine. But the AI piece kind of brings a three-dimensionality to it. Uh, it it lets us start with a visualization rather than with data. And I find students really appreciate starting with that. They also, I think there's a confidence that comes from starting off with that image or whatever parallel it might be in other disciplines, where students go through that kind of what feels off process. And there's a confidence that comes from saying, you know what, I I'm pretty sure I'm onto something here. I have a I have a hypothesis where something feels right, I'm already in this place where something feels a little bit off. And if I can provide them the data, the resources to be able to pursue that gut feeling, they go into it already having something of a starting point that they're looking to validate through evidence and not just here's a stack of data, tell me what you see when it comes to party coalitions and their demographics. So I think there's a number of things that benefit students, ranging from it's just fun all the way to building confidence. Yeah. And making sure that they feel comfortable using the data because they're not just starting with a stack of numbers.

Derek Bruff 15:47

Right, right. You're giving them a more directed task with the data, which is helpful and and probably kind of reinforcing the notion that they should have a bit of a radar when they encounter political messaging, right? That sometimes it's going to be something will smell fishy, and and also we we have tools actually, maybe, to figure out is that smell accurate or not? Yeah. Yeah, absolutely.

Eric Loepp 16:09

And there's any number of ways you could take this, right? You could say, uh, draw me uh the portrait of a typical Fox News viewer or MSNBC viewer or something like that. Um, and students almost always have uh an image in their head, right? Um, but one of the things that is really important to do in that information literacy vein that we were discussing is for students to be able to validate that gut feeling because in many cases their hunches correct, but it can also lead, especially in political science, to the overuse or excessive use of stereotypes. And so helping draw that line, like, okay, this image may roughly capture the reality, um, but in what ways does it fall short? You know, maybe it turns out that you know a larger percentage of Fox viewers than you expected were women, or a larger percentage of MSNBC viewers than you would have expected are not Christian, right? Or or whatever it may be.

Derek Bruff 17:04

Yeah, yeah, I love that. Do you find that um because one of the challenges with these critique the chatbot assignments is that the chatbots keep getting better at stuff. Do you find that you have to kind of calibrate a little bit or maybe like prompt poorly so that you get an image that has enough flaws that there's something to talk about?

Eric Loepp 17:24

It definitely has improved. Uh, I don't recall having to change the prompt too much. Um, granted, of course, it's only been a couple of years, and some of these classes I don't teach every term. So for some of them, we're I'm a year out from the last time I did it. Uh, but it's definitely something on my mind. And indeed, the the degree to which the critique the chatbot concept will be valid. Uh, you know, I could easily see this one having having its time, you know, kind of like VCRs did. But then we just kind of collectively move on to a new way of doing things. So I tend to think of it as more of a bridge activity or a bridge philosophy. Um, but I would say in the last year or so, we're we're kind of in the sweet spot where AI is good enough to give us something that you know is like valid enough. It's a valid enough starting place. It's not just completely random, right? Um, but it's not so good as to take all the census data and just spit out something that is, you know, to the decimal point perfectly proportional, um, which you got to be careful because if we ever get to that spot, then the students are gonna say, What's the point of doing this? Like, why don't I just like I'll just give AI the Pew data that's 40 pages I don't have to read, and in one image I can get the critical bumper sticker. That was the goal of this whole exercise, anyway. So, I mean, one thing I found with AI, I'm sure you found it too, is uh it's a little unnerving whether you're writing a book or doing a podcast like this or doing an exercise in class. That exercise, you know, maybe from 2000 to 2015, I only had to nominally update it. Yeah. But between 2022 and 2026, or let's say 2028. So between the 2024 election and the 2028 election, I can imagine that my parties in elections class is going to dramatically change in terms of the role AI is playing. Whereas, you know, 2004 to 2008, eh, you know, a lot of the things are basically the same.

Derek Bruff 19:15

Sure. Yeah, yeah. So this critique the chatbot. I I I I've heard you talk about kind of another angle that you use with this, and that's asking the AI to do some type of forecasting or predicting where like there's no way it can be right because no one knows actually. And so um, could you say a little bit about kind of what that looks like in your class? Because I feel like that that might offer a way. Um, I mean, presumably the AI tools won't be magic predictors of the future at any time in the uh soon. So this this could be a way others might use AI and a critique the chatbot assignment.

Eric Loepp 19:51

Yeah, if they are, then you and I can just get one, invest our life savings and and retire, right? Yeah. Uh yeah. So I mean, one thing that really turned me on to kind of being future-oriented in my coursework and my assessment work, it's it's really twofold. One is as a discipline, you know, students who are working in the political science field, it's it's very future-oriented. You know, if I have a political science major that wants to go work on a campaign, you know, you can only talk so long about how great the 2022 campaign was that you worked on because 2022 is a very different context. So we're always future oriented uh within the discipline. But the other thing was when AI first came along, and it's still the case today, maybe it's getting better as it always is, but one of the areas where AI really struggled was in that forecasting, that future predicting. Uh, there are so many variables that are volatile, information that can't be known yet. Um, and so if you're forecasting particularly further out in the future, uh even a student who wanted to use AI and perhaps do so without authorization would would only find it of limited help. Uh so in my classes related to elections, let's say, you know, one of the big things that students do now is as kind of a capstone experience, I usually teach it in the fall of presidential election years. Now, historically, the capstone experience was more of a an autopsy, right? The students had to evaluate the election we just went through and you know make points about why so-and-so won, and et cetera. Um, and that is the kind of thing that AI could pull very easily from all sorts of commentary out there and give you a pretty defensible paper if you asked for it. So in 2024, the second time I taught the course, right, since arriving here, I really reoriented it to be much more about okay, what recommendations do you have for the two political parties for the 2028 election? You know, I want you to diagnose exit poll data. You know, we look at some economic forecasting, we look at projected demographic shifts. These are all projections, right? The US Census will do kind of between census sensi, right? Between them, there will, you know, community surveys and some projections of where we'll be annually, right? Uh having students work with those to focus a little bit more on what we can expect and what do you recommend in the future, I think has the twofold advantage of one creating a more compelling artifact for people that are going to work in the world of politics and government. But two, also creating something that is much more meaningful. In an era where AI can't really do that work so well. The autopsy stuff, yes, it needs a human touch. I'm all for doing some version of that, but I've really shifted the emphasis of that capstone from looking back on our course to looking forward to the next one. And I was joking with the students in 2024, like you should come back into 2028. You know, once you've graduated, let's do a panel discussion and compare your experience going through this uh as an alum. Yeah. And you know, talking to the students who are doing it currently about what their process was.

Derek Bruff 22:57

Yeah. Wow. Um I want to circle back to AI. And, you know, part of our AI aware teaching framework is having a better sense of how your students are thinking about AI and how they're using AI. And so you've talked a little bit about how you use AI as an instructor to kind of set up some activities for students to help them dive deeper. Um, how do you go about figuring out kind of how your students use AI, how they think about it? How do you have those conversations with students about the role that AI might play in their learning?

Eric Loepp 23:28

I think this is as an important conversation to have at the beginning of the term as conversations about inclusivity and civil dialogue and kind of the broader culture of the classroom. One thing that amazes me, I mean, here we are two to three years, even beyond that, from the initial iterations of AI. And I mean, I'm kind of amazed that even though we've been wrestling with the perils and the promises of this tool or this suite of tools for several years now, there's still a wide variety of experiences that my students report. And we say that anecdotally, just in office hours and such, and also in formal research, you know, these national reports coming out. Um and I'm a big fan of starting all my courses with some open and direct conversations about a few key things. Again, inclusivity, civil dialogue is a big one, but AI is in the same category now. Um, how do I think about it? How do I plan to use it? What are my expectations? Uh, but it's not just a matter of communicating what the policy is as if it were an attendance policy. It's really about I want to hear from you about kind of what your hopes and dreams are, what you're concerned about. Um, I'm sure you've seen the same research I have that one of the things that, you know, really, really worries students, I mean, almost terrifies some students, is the concern about being falsely accused of using AI and being very, very unsure of when and how they can use it. I've even dabbled in the idea of having each of my classes kind of design their own AI policy based on the SLOs and based on, you know, the size and some of the different course assignments because students may feel differently if there's going to be a lot of group work than if there's gonna be a lot of in-class exams.

Derek Bruff 25:07

Sure. Yeah.

Eric Loepp 25:09

Uh now I've not done a live version of that yet. I'm a little bit nervous about you know, five, four classes and they all come up with something pretty different that would make my life pretty difficult and you know, applying standards and all that.

Derek Bruff 25:21

Um, but far and away, the the biggest although it occurs to be your students are probably taking four or five courses that may have four or five wildly different AI policies. So maybe maybe we should be taking that burden.

Eric Loepp 25:33

You know, that you know, that is that is fair enough. You know, it's kind of the it's kind of the flip side of like I have 40 names to learn in each of times four classes, and and you have only a handful of instructors to know. Like it's important to appreciate both sides of that. Yeah. Um, but far and away, uh, what I found both anecdotally and in my my work at our teaching and learning center and in reviewing the national reports and such is that the worst AI policy far and away is no AI policy, right? If there's a lack of clarity, um, that is way worse than an absolute prohibition. Um so that's why I always start is like, hey, here's where my head's at. And and there are times, you know, students may say, can we do things slightly differently? And I'll say, no, like this is a pretty important line for the type of work we're doing in here. Like we can't just eliminate writing, you know, things like that. Um, but the other thing that I've found is it's not just a matter of talking to students. I mean, it's very important. They should be a part of the process. I think they should inform how we design our syllabi to a degree. Um, but I I think it's also important that we be mindful of the homework we can do and the steps that we can take before the class begins to help set that stage. Um, so you know, one easy example of that is, you know, taking a good hard look at our assessment, right? If I'm very concerned about what AI might mean for a particular course I'm teaching, then there's also a burden on me to make sure I'm designing learning experiences that are rich, that are meaningful. Um, I can control that to a degree. What I don't have as much control over is how students use AI when I'm not around, right? You know, my general policy is that AI should be a tool and not a crutch. And with some exceptions, I encourage students to use it. If it helps prepare them to deliver things that will be graded to the course, uh I always tell them that I, again, with some rare exceptions, depending on the course, I don't want them using it to produce anything they're submitting for a grade. But if it prepares them, if an AI tutor helps them create practice quiz questions, or if they want to run some logic through AI and say, Am I missing something here? Um, that is the kind of thing that I would do. So it's really not fair to suggest that that no one else can do it. Um, I have my fair share of academic integrity issues. Um, you know, we've we've all dealt with them, but I I at least have found them to be relatively uncommon. And I think part of the reason for that is starting off doing that that work at the front end to set those expectations, to invite feedback, but also to try to design a course in a way where it's just not that tempting to use AI. I mean, this is one thing that I don't think it's talked about enough. You know, we'll we'll hear things like, uh, well, you know, students don't know that AI is often wrong or hallucinates, and so they'll send in, you know, this this material that's completely wrong. I haven't found that to be the case. I found students to be quite sensitive to the fact that AI can be wrong. Okay. Um, but I've also found that when when we're concerned and frustrated about students using AI when we don't want them to, you know, again, this is just Eric's experience, but I found that it's often the case that they're tempted to use it because they're not sure how to proceed, or they're really running short on time. Um, and they get to a point where it's just worth taking the chance to, you know, either submit information and hope it's correct, uh, or take the chance on submitting something that ends up being wrong, or taking the chance on getting caught. Um, I want to be clear, this is not to say that students do not own a big part of the academic integrity process. They they really do. Um, but I do think that we can help create a culture both in our policy making and in how we try to create norms in our classrooms and in our digital spaces where students are less tempted to rely on AI. Because I find that oftentimes the reason they do, it's not that they're just lazy or don't like the course or want to stick it to me or something. It's often I'm not sure how to proceed, or oh crap, it's due at midnight. So I love that universities are really investing in you know new student orientation programs and and you know, English 101 classes that really focus on things like time management, writing skills, and the like. Uh easier said than done to do that, I know, but I think there's there's a lot we can do in addition to creating policy and communicating that policy to try to create a culture of ethical AI use in our classes.

Derek Bruff 29:54

Are there you mentioned your um the the presidential election course that you changed the post-mortem um activity near the end to something different, uh, partially in a response to AI. Are there other kind of assessment moves that you've made or assessment changes you made so that it it kind of reduces that temptation for students?

Eric Loepp 30:15

One thing that I've tried to do uh across my courses is to shift some of the deliverables to be more uh personal and more about evaluating normative arguments. We get up to do a lot of that in political science. Um, so for example, in one of my courses, students we we study uh representation and voting and the different challenges, barriers, access to voting in different states, right? The rules vary quite a bit, you know, how early do you have to be registered and you know how late are the polls open and all that kind of thing. So uh in this class, I'll have students as part of a broader assignment where they they analyze voting access and you know which states are harder to vote in. Do those states tend to be home to more voters of color or whatever? In addition to some of that analytical activity, students have to design their own ideal voting access rule system. So, you know, do you have to be registered before or can you register on election day? Are you automatically registered when you turn 18, or you do you have to fill something out? If you have to fill something out, is there an online option? You know, all these different things. Do you have early voting, vote by mail, all the different parameters that we study? And students have to design what they think is the ideal system that balances election security with election access, right? Because it's all trade-offs, right? If you want a super secure, like it's virtually impossible to cheat kind of election, which we would all love, right? We don't want any cheaters. That would mean adopting a lot of rules that require a lot of voters in order to cast their ballot. If on the other hand, we want to uh elevate the idea of let's make it as simple as possible for everybody to participate, then we might have to let up on some of the rules that promote security. So it's all about establishing that trade-off. And so I'd like to put students in a position where they have to understand the arguments from different sides, you know, whether it's voting ID laws or closing a border or whatever. And they have to design an ideal system, they have to defend that system. But then one of my favorite things to do is say, okay, now imagine that there's somebody who is thoughtful and principled, just like you, but their principles and their values are very different. They don't see things the way you do. Make the case for their interpretation of the ideal voting system. What critiques are they going to bring to you that are valid, right? And which critiques maybe aren't so valid? Uh basically, you have to be in the position to make the case against what you think to be the right way for our democracy to work. And the goal is not that everyone changes their minds, but that they walk out and appreciate that in a lot of cases it is a matter of trade-offs, in this case, access versus security. And if you want to be able to articulately and effectively communicate your worldview, it really helps to understand the worldviews that would challenge yours, because then you can form your responses, clarify your positions, and better make the case for them in the future. You know, if somebody wanted to outsource an assignment like this to AI, you know, they could certainly try. And what AI can do reasonably well is, you know, hey, AI, you know, give me a list of bullet points that make the case for why we should have very restrictive voting access laws. Why is that good? What can that do well? And then the flip side of that, what's the benefit of opening them up? And AI can give you a bland and I mean accurate, but you know, kind of bland, like, well, here's what some people think. But what it doesn't do so well is, you know, provide the student. What do they think? Right. I mean, my goal is even if AI helps you get these thoughts organized and helps you understand the technical basics, I want them to be able to advocate, no, here's why I think that a votering ID law is so important. And here's the evidence that backs up my position. It's not just give me the pro and the con of requiring voter IDs. It's about what do you think? Why is this the right ratio of access to security for you? What about your life experience? Maybe, you know, I I've had students who are voting for the first time and they have no idea where do I even go? I had to register beforehand. Do I put my dorm address down? Like they actually go through the process of becoming a voter in many cases, their first couple of years in college. And I like them to think about what is it about the current system that you think is reasonable and maybe what isn't so reasonable. So AI is not so great at helping you figure out what the right trade-off ratio is for you. Can it help you inform what the trade-offs are? Sure, great, go for it. But it's not so good at that last mile, if you will, of but what is it that you believe? What's the part that only a human can do?

Derek Bruff 34:52

Yeah. So I know there's a lot more that we could talk about, um, but I do want to ask about um, I feel like in your discipline, uh there's a potential for some change in learning objectives because of AI and the way that AI is used in our political discourse. Have your course goals and objectives changed any since generative AI entered the scene?

Eric Loepp 35:15

I'd say they've changed a bit. I mean, I I try to be careful not to change too much every year, right? Even if if uh it looks like I may uh need to move more quickly in some courses than others. I think the biggest impact for me or the biggest change is is really shifting more to a process approach over a production approach to how I design my courses and my assessment. Um so, you know, for example, when I teach about parties in elections, one thing we do is look at all the different ways that you could divvy up the electoral college votes that we have in our system. You know, right now in almost every state, whichever political candidate gets the most votes in that state gets the entire bushel of electoral college votes. And that's why on election night every state is red or blue. Right. But there are any number of ways that a state could decide how to divvy up those electoral college votes, right? They could do a proportional system based on the statewide vote. They could do a proportional system where they do it by each congressional district. Now, in the past, I might ask students, I might give them a hypothetical outcome of an election, say in Wisconsin, where I am, and say, okay, let's imagine that Wisconsin divvied up their electoral college votes using this other system. What would the election outcome be then? And how would the candidates campaign differently if that was the target they were aiming for? Now, today, AI could do that pretty well. Okay. AI could do the raw calculations pretty well. I've tested it. I mean, it's it's not just what is, you know, a 20% tip on an $80 bill. It can do some pretty sophisticated stuff. So what I'm moving more now is instead of focusing just on creating a product, like what are the technical outcomes, focusing more on having the students maybe, maybe they still do that product, maybe they still do those calculations, but the the real thrust of the activity is um, okay, now let's look at four different ways Wisconsin could have divvied up its electoral college votes. Now I'm gonna challenge you students to think about which political groups in Wisconsin? It could be bipartisan groups, it could be different industries in Wisconsin, you know, the tourism, industrial, whatever. What groups and coalitions in Wisconsin would favor each of these different systems, right? If all the farmers in Wisconsin could pick one of these four options, which one would farmers by and large want to do and why? Um, which of these systems do we think would the public most enjoy? Right, based on our knowledge of what people have to say about election preferences. Which one of these do you think would most align with our legal precedent, right? Which of these systems best reflects the vision and ideals of the framers of our constitution that we studied two months ago? Right. So it's really moving more towards interrogating the information and again focusing on that normative, that procedural side of things. How ought the world work, right? How should things work? Um, as opposed to kind of a, well, if Donald Trump had collected, you know, how many more votes would Donald Trump have needed in the state of whatever to have won that state in 2024? AI can do that beautifully. And it can handle that. 10 years ago, yeah, 10 years ago, that's a good multiple choice question for students to demonstrate that they understand how electoral college votes are divvied up. But now I try to focus much more on what's good about this system, what are the political motives of all the people involved, and what are the different interests of the different groups out there in terms of the potential alternatives? It all comes back to where we started our conversation. What should our political world look like? To what extent have we achieved our, you know, the top of our pyramid, right? That self-determination to both Maslow, right? Uh I think it's an ongoing process, right? The system will always be one of hopefully progress. Um, but students need to decide what that looks like because for some progress may mean turning this way, others may see progress in a completely different way. So, how can we design our assessments and our courses to facilitate students really challenging themselves to think about how that world ought to work?

Derek Bruff 39:16

I love it. I love it. Well, thank you, Eric. Um, that that feels a little optimistic on some days, but I love having that as our as our North Star. Um and helping students appreciate the complexity in our systems and and different points of view and and and how you how you balance all the trade-offs that are that are necessary. I I just um it sounds like again, a fun course. Uh a certain kind of fun, I think, to be a part of that course. All right. Well, thank you, Eric. We will leave it there. Um, thanks so much for pulling back the curtain a little bit and showing us inside your classroom. This has been great. Thanks for being here.

Eric Loepp 39:49

My pleasure, Derek. You be well.

Derek Bruff 39:56

That was Eric Loepp, Professor of Political Science at the University of Wisconsin in Whitewater. Eric also co-directs Whitewater's Center for Teaching and Learning. Thanks to Eric for coming on the show to talk about his AI aware teaching practices. I really love how excited Eric is to teach political science and to change his teaching approaches over time, not always in big ways, but in significant ways, to continue to help his students deal with the complexity in our political systems.

Derek Bruff 40:24

You can read more about Eric's teaching in the Norton Guide to AI Aware Teaching, which features concrete examples of AI aware teaching practices from Eric and from many more thoughtful, effective instructors. The ebook is available now, and the print edition will be available in September. See the show notes for more information.

Derek Bruff 40:44

Intentional Teaching is sponsored by UPCEA, the Online and Professional Education Association. In the show notes, you'll find a link to the UPCEA website where you can find out about their research, networking opportunities, and professional development offerings.

Derek Bruff 40:59

This episode of Intentional Teaching was produced and edited by me, Derek Bruff. See the show notes for links to my website and socials, and to the Intentional Teaching newsletter, which goes out most weeks on Thursday or Friday. If you found this or any episode of Intentional Teaching useful, would you consider sharing it with a colleague? That would mean a lot. As always, thanks for listening.