AI & Debate Interview Part II

Part 2

Eric Byron  0:09  

Welcome to the education innovators Podcast. I’m Eric Byron. And it’s an honour to host this show where we get to hear from talented educators who are willing to share their stories of the incredible things they are doing in learning environments all over the world.

Stefan Bauschard  0:28  

The reality is AI is here, it’s not going anywhere. It’s getting a lot better, it’s going to change the world. It’s going to change your classroom. If it doesn’t change your classroom, then I don’t know what relevance your classroom is going to have to the world.

Eric Byron  0:42  

That’s the voice of Stefan Bauschard, a debate coach and advocate for debate-centred instruction. As you can tell, he gets a bit wound up when he talks about AI’s impact on education. To be expected, as he published a book on the topic. In this episode, we get into a passionate discussion of the impacts of AI, and how debate centred instruction can help develop the durable skills students need even more in a world where AI is getting better and better. This is part two of my conversation with Stefan, fasten your seat belts, this is going to be fun.

Stefan Bauschard  1:18  

There’s some different types of debate. And my background is in more policy debate, which is really information, actually, processing. And that’s where a lot of the full time coaching came out of because you have a lot of arguments, you really go in-depth on a topic. Like they say even high school kids with like, the amount of research like a master’s student would do. They’d be reading all these articles, journal articles here, you’re reading all this stuff organizing dates, these ideas, right, like, so this took a lot of time. It’s very cerebral focused. And even the evaluation of those debates is kind of more on the arguments, a little bit less on the performance. You have other debate events that are focused more on the performance of the speech events, like they have their little bit of niches right? I always felt very comfortable in this information space, right? So it’s like, yeah, I can take an idea and break it down relatively easily, you know, some of this stuff, the concepts are more complicated. Like it took me a while to really understand backpropagation the development of a neural network. And that was mostly just doing that. So you know, I could understand it. Or sometimes people ask me about it. I’m not going to be I’m not on my way to becoming an AI scientist.

Eric Byron  2:22  

But it brings up an interesting point, though, Stefan, and this is something I’ve seen, almost generationaly, on different. I always wanted to know, too, how do things work. Don’t just tell me, “Oh, you push this button. And this happens, right?” I want to know what happens when you push the button, right? What’s going on behind the scenes. But now and I taught computer science courses, I wanted to impart on my students, right, that this isn’t just you got to know what menus to go to to configure a system. I want you to understand what all those settings do and why they exist and, and stuff. And, and I had students push back to saying, “I don’t care. I don’t want to know all that stuff.” And then when I started my own education startup in Hong Kong, we were doing a coding bootcamp. And I wanted to add some kind of computer science to the curriculum. These kids were learning to code— kids, I say, they were mostly in their 20’s— but they didn’t want to know. And the even the other instructors, the faculty didn’t want this stuff introduced because it’s so much harder to teach, right? It’s so much easier to just say, no, here’s a library, and this is the syntax. And this is how you do this, you know, just follow this formula, cut and paste this code, it works. You don’t need to know why it works or how it works. Just know that this works. And I kind of hated that. I want to know how stuff works. 

Stefan Bauschard  3:47  

Yeah, I think in this instance, I mean, look, I never know, I have an iPhone, I’ve generally got the the improvements as they come out. I’ve never really cared to know exactly how it works. I’m on this desktop, I think in this instance, and it’s helped me explain other things to people. It is important to understand and I think because people people bring their— and this happens in instances, I did it too when I first started using it— you bring your prior knowledge and association. So the first time I used ChatGPT, I thought it was a search engine, right. I thought people would call it that, right? I spoke with someone who actually had used it a lot in different things. He told me he thought it was using like it would just draw on every search engine and put everything in there. But when you explain to people, and you don’t have to get, I used to get too complicated, right? Even though I wasn’t an expert, I learned enough that I could really explain the neural networks and the deep learning and all that kind. I found out that, well, that’s not important. But what is really important is that, it is two things. One, it’s predictive text, right, it’s predictive text that follows the next word, but two, it’s not just randomly. It’s not just random predictive text. If it was just random predictive text, we wouldn’t be talking about this right? That its output is, you know, improved, right, you have all the parameters, you have your fine tuning, you have your reinforcement learning with human feedback, and now even AI feedback, there’s a new paper out on it. But that, that makes the output, you know, then there’s these contacts and sentiments, you don’t even have to go into that. But I just say that there’s another layer on top of that, that makes that output meaningful and relatable. Because then once they understand those two things, right, they can understand one, that the output isn’t going to be the same. And this is one thing I’ve heard is, I’ve had people start doing, when I teach them about it, I make them enter the same prompt twice, back to back. And they’ll see that the text output is at least slightly different, especially if you have a broader prompt, right? If you’re trying to get something a little factual, right. But anyhow, they see that too, they can understand why you don’t want to rely on it for facts. But also, given what I’ve explained about the fine tuning and the reinforcement, it’s not terrible at facts, right? It does well, in some of these tests, it gets a lot of facts right. Like, you just can’t rely on it. And then third, they understand like why the plagiarism detectors don’t really work that well, because what they’re obviously trying to do is look for patterns within the output. They’re not, they’re not pulling something from a database that you can go back and verify, right, like Turnitin. You can’t see well, oh yeah, student, you took this out of Wikipedia. Here it is. There’s nothing, there’s nothing to back verify. So if it does, if it misanalyzes— I don’t even think misanalyze is the right term, but obviously there’s false negatives and false positives, even without students perturbing the text, right. But even if the student perturbed, if the student perturbs the text in you know, reasonable ways, right— you increase the burstiness, or the complexity— now that detector is basically gone. And they understand that, once they understand how it works in a predictive text, and they’re saying, I’m getting two different pieces of output, and how it works, you know, you can’t go back and say, “Oh, did you write this?” Somebody asked me that, someone I knew was a professor and said, “Oh, well you know, I caught a kid.” I said, “How did you catch him?” You know, he said, “Well, what I did is I asked ChatGPT if it wrote the paper. And it didn’t really say, like, but there were some out pieces of output that I found in ChatGPT that were in his paper, and I just, well, this is like, by random chance. Like if a kid wrote a 10,000 word paper, there’s probably something, about a particular topic, there’s probably some similar words that appeared to one another, next to the paper, but I had to explain all that, right. So I find that, once people understand that, they can understand a lot of other things: A. what it’s useful for and where you know, what it’s more reliable for, I guess I’ll say it that way. But it’s less reliable for but it’s more useful forward, especially useful for and then kind of some of the quandaries of, of using the detector. So ordinarily, I wouldn’t say that it’s essential, or even, you know, there’s a big issue about, you know, representation and output, right. And, you know, I’ll say, like, look, if I say, you know, on an image generator, I’ll say, give me an image of a CEO, I said, I don’t almost I haven’t counted, but almost every time, they’ll give me a white male, because that’s what most of the pictures of CEO, but I said, if you put in there, give me an image of a black female CEO, it will do that, right? Or what I do classroom makeup, because sometimes I read articles about debate or classes, I’ll say, “Please,” I say please to get my sentiment analysis, right, you know, “Please produce a picture of a classroom full of students. There should be some white students, black students, Asian students, Hispanic students,” right? And then it does that, right? So you can direct it. But if you just say, give me or to, you know, give me a teacher kind of generally gives you this white female female teacher, inyou know, her late 20’s. Right, so the generic feature, but once you understand, I do think in order to use these tools, at least right now, that may change, right? Maybe, you know, if you’re somebody who just have an interface, you know, teacher, right? Those kinds of things, right? That’s easy to do. Right? But which is basically what a lot of these companies are doing, oh, we made a thing where a teacher can generate a lesson plan. It’s like, okay, well, you just, you’re just through you just through a graphical interface on top of it, right. And you had to click a couple of buttons and you hada lesson plan. But anyhow, I do think in this instance, it’s important to understand a little bit how it works, because I think if you don’t understand it, I don’t I don’t really think you can use it properly. 

Eric Byron  9:15  

Yeah, AI literacy is, is huge. And teachers and students, right, everybody’s got to understand how it works. And when to use it. And all that.

Stefan Bauschard  9:25  

Yeah, and then you can get into those secondary questions when you’re using it for, what are you using it for? You can start thinking about it. You know, obviously, my background is in debate. And of course, a lot of debate coaches, and teachers are all freaked out, “Well the kids are just gonna get all their arguments” and I said, “You know what,” I said, “I think it’s great.” I said, “I’ll let this thing turnout 10 arguments, then I’ll talk with the students. What are the better arguments? Why? if you only have a two minute speech, which arguments do you choose?” Sometimes it generates a list of arguments and they’re inconsistent with each other, right. So which ones are you going to pick? Why are they inconsistent? Which ones are just kind of reasons that you could state in the debate? Which ones do you need to go research to? Like back up? Right? So there’s so much more you can do. That’s why I really liked the AI, the AI augmented, right? Because if you just try to compete with the AI, especially over time, it’s gonna, you’re shaking your head. You’re gonna lose, right? So I don’t know how to say his last name, but Mustafa Suleyman, who’s the co-founder of DeepMind, worked at Google, now has Inflection AI Pi, I was watching a podcast at the end of the day. He said towards the end, he said, “So far, at least,” I don’t know if he said so far. But he’s basically saying, “So far, the human and the AI together can defeat the AI,” right. So the AI first defeated the chess champion, then the Go champion, which, sounds like there’s like four quadrillion moves or something like that. Right, then Diplomacy, which involves all these interactions with people overseas, it’s so far when you put the human with the AI, that generally tends to defeat the defeat the AI. So that’s why I think yes, I think conceptually at least right, there’s a lot of things that affect whether somebody wins the debate, but a student using AI well, to enhance their debating is obviously having both an advantage over a student who’s not using AI, right, and I think the AI itself, especially AI, you know, these things have a lot of limitations, right? Where they’ll go, you know, I talked to people who I knew from debate, you know, who graduated from debate, they went on to get CS degrees, you know, PhDs and you know, they have really high positions at Amazon, and all these places, and they say, they say, “Yeah, eventually it’s going to be able to do the whole debate.” And they said, you know, “If Amazon didn’t have me doing something that was more important than trying to win a college debate, I could basically train it. If I had, like, 10,000 videos, like we have with everything else with college debates, and then I heard, you know, heard, so to speak, how it was decided, right, it’s only going to break down in so many ways, which is true, I could train it, right, to debate like that. So it’s eventually I mean, but look, eventually, I think, eventually it’s gonna be able to do what we can do, right…

Eric Byron  12:07  

I’m working on my own AI that I’m going to train, so I’m going to feed it all of the transcripts from all of my episodes. And all of the papers like yours that I used, as, you know, in my research to prepare for the, the podcasts, I’m putting that all into a database for a chat bot. So I can later go in, right is that over time is that starts to really build up a knowledge base there. And I can go to the chat bot and say, “Oh, give me a good reference from, you know, an AI conversation with either Stefan or, you know, somebody else, you know, I need a quote.” And instead of me trying to remember, dig through, go, you know, scanning, the AI can do it for me.

Stefan Bauschard  12:59  

Yeah, just take some, you know, material on a debate topic, right? Have it trained just on that material, right? Because that reduces likelihood of hallucinations, right. And then add in my material just about debate formats, and some different skills and then kids wherever they’re starting, they can always Q&A, that limited, that limited material, just like what you’re doing right, like, and it works well. And then…

Eric Byron  13:23  

Well, for practice, Stefan, right, I’m gonna practice my debate against the AI. Right, before I go into the debate competition. So you create a tool…

Stefan Bauschard  13:32  

Well that’s in people, you know, because people are still using it primarily as a search engine. They haven’t realized all these things yet, but you know, I have some prompts that, you know, that I share with people that say, imagine, right, that you are a high school debater, and you are doing a cross examination with another high school debater. Ask the debaters some questions, right. And then you ask a follow up question, I’ll ask a follow up question. Right. So the students, you know, people because you’re still seeing it as a search engine, right, you don’t realize it’s like interactive, and it’s just gonna get better at that. And especially once it gets, right, right, multimodal and like immersive and right, there’s just gonna be like, you’re they’re debating, right? But it can already do that at a pretty basic level, as long as you do the prompt properly. It’s a great way for kids to practice because you can’t, you know, some of your coaches say, well, am I gonna lose my job? It’s like, great, now you’re not,  you can hear, maybe you can hear like one practice speech from every kid or a couple practice today. Here, a kid could sit at home and just just keep practicing their arguments with a bot until they don’t want to do it anymore. And every kid on your team can do that. So I just think, it’s incredibly empowering, yes.

Eric Byron  14:37  

I want to talk a little bit about assessing durable skills through debate. So imagine, right we’re in a classroom. Let’s say it’s middle school. Okay, I got a middle school classroom. I’m introducing a debate-centered instruction element, and I want to come out with an assessment of stuff your critical thinking skills or their persistence, their curiosity, their creativity, right, you know, some of these skills that we’ve all said employers want, right? LinkedIn says it, Economic Forum’s Future Jobs says it, right. How do we assess these things? I mean, quantifiably, consistently, fairly. How do you do that? Because that’s, to me the real problem with targeting those skills.

Stefan Bauschard  15:27  

I mean, there’s some of those skills, they’re in the, I don’t know if I categorize, but some of them are a little easier to assess than others, right? Like, so you can have a like public speaking skills, right, you can say, here’s a rubric for things I want you to have, you can just think of the basic ones, you know, eye contact, like, you know, looking at your audience, you know, reasonable amount of like hand gestures, but not too many, right? And you give them scores on the rubric. And I think that with, and then you know, the same thing you could do with your arguments, right? Do you have like, at least so many arguments? You know, do you have so much evidence to back that up? Is your evidence like from reliable sources? And you can build those in over over time, right, like the first time kids debate, you don’t worry about how reliable the sources are or those kinds of things, right. So you kind of build that in over time. And of course, with the computer, it’s easy to keep track of it. Because in the old days, you just have one random paper, and I wouldn’t remember what Eric got on his last paper. So you know, even if I gave him a different score, I wouldn’t know if he got any better or worse. So I think with computer systems, right, like, that can help you track that. And it can help you track his skills. And I think the second thing that that enables, which is gets harder, right, but there should be more cross-disciplinary learning, like kids should be, you know, debating in like science class about whatever it is, they need to learn history. When you know, and this is true in a debate, like with different judges judge across the tournament, it’s not the exact same point of reference, right? We all like slightly different expectations, right. But over time, those things, those things can score. And that’s why when you can do it, and of course, like I said, I you know, I brought up the knowledge thing, even though you know, I’m more on the skills side. I don’t think it’s so bifurcated. I think kids need both, but it’s important to teachers, and you know, there are still standardised tests. And that’s the way kids are evaluated. So you’d also see how they did on the test, right to, you know, at the end of the semester, whatever the the unit, to see how they’re doing skills, like resilience and stuff that you’re talking about, like empathy. I mean, those are much harder to develop. Evaluate, excuse me, I should say, but I think that I wish I knew the name of this, but there’s someone named John Kelly, who, Colegio Ikigai, he started this school in Mexico, it’s all VR-based and it’s all skill based. And they do, you know, they take a theme, there’s a name this too that’s escaping me, they say, “Okay, well, I’m gonna learn about theme like climate change, I’m gonna learn about it, and history and biology and English, and we’re gonna kind of have this theme that crosses.” And, you know, he talks about the skills like resilience, and I went to one of his things he was just talking about it, and I said, “Well, how do you, do you measure these things? Right, like, how did you measure them?” And he said, the first year of school, he didn’t, they didn’t measure, right. But he said, “I don’t care,” he said, “they’re getting better.” Right. So he’s, but you know, obviously, probably, you know, to make it more, now he’s working with some company, and I wished maybe I’ll look it up and I’ll send it to you. Maybe there’s some way you can put it in. But there’s some company that has some ways that you can at least measure these things. But they’re obviously harder to measure, right, than just content. It’s a little bit harder to measure. In debate and competitive debate, kids do it so long, it is kind of like, you know, they get better at it. Right? It’s like, you know, if they just debated a few tournaments. Sure they learn on them on that much. Kids debate over one year, two years, three years, four years, you, you can kind of see them, you can see them improving a lot. But it’s definitely something that’s a little bit harder to measure. But we do have, we do have public speaking classes. Right. So that’s, that’s not impossible to measure. We have argumentation classes, you can see our kids break down an argument. You can, it’s not terrible, to also give them a test and say, “Okay, well, we’ve been working on, you know, I taught you what an argument is, a basic component of an argument. You’ve got a couple of worksheets where you have samples to work through to break down an argument, you had an actual debate where you had to build an argument and refute an argument, right?” And then like, “Okay, here’s another example, like, what are the weaknesses?” At that point, that kid should be able to identify the weaknesses of the argument, right? So you know, you’re assessing it in the debate to a degree, it’s a little more kind of touchy feely, right. But you can also assess it with a test. Like, I don’t think we should get rid of tests. I mean, I’m, you know, we should have quizzes and things, maybe fewer high stakes tests that are like, well, gee, the kid knows this, or they don’t. You know, tests, quizzes, they force you to kind of learn some stuff and pay attention, right. So I think, I think all those things are assessable. I mean, when I taught at Boston College for five years, I taught argumentation and debate, right. We had some debates in class, and then the kids had, like, you know, some sample like tests, or they’d have to go through and break down an argument. It’s obviously, like I say, it’s harder to assess than content, but it’s not impossible. 

Eric Byron  19:51  

So, one of the things that also concerns me a little bit, is that debate, generally, right, it’s a competition. There’s winners and losers. But if you’re going to bring this into the classroom, is that the way to do this? Do you want to have winners and losers? Because you’re going to have kids who are going to lose every time, right now they’re completely discouraged. Because they’re not inclined, this is not their skill. This is not how they’re wired. They don’t do this well, they’re not critical thinkers, right. This is unnatural to them. And yes, they can progress given enough time, but they’re gonna keep losing over and over again, because there’s other kids in their class who are just better at this. So, how do you avoid that stigma, of I’m never going to pick you as my partner for the debate because you lose every time, right, and now I’m the kid who sits on the bench and never gets picked?

Stefan Bauschard  20:46  

Well, I mean, to your last point, right, I’ll start there and kind of build backwards, right. Like, yes, generally, if I’m teaching regular classes, you know, first couple of debates, the kids, you know, pick who they’re comfortable with. They’re probably going to pick their friends, right? Even more than whether they think they’re gonna be a good speaker or not. They’re gonna pick their friends, and then on a competitive team, kids will kind of move to, you know, the better kids will want to be with each other, right? So you’ll see that phenomenon. But you can also intervene as a teacher, you can also mix up the kids, right? Like, you don’t want to always say like, okay, well, you don’t want to always make Eric debate with Stefan, because, you know, no one wants to debate with Stefan, right? So you always pick the strongest students, so but you can mix that up. It’s also, because it’s graded, it’s not, sure, there can be a winner and a loser in a particular debate, right, but because it’s graded, I mean, you could lose and kind of still get an A, right, just as long as you met, like, all the criteria of the rubric that the teacher shouldn’t grade the debate on who won or lost, right? Like, that would be inherently like, could just depend on who you debated. Like, “Oh, yeah, I’m gonna, I’m gonna get an F on this assignment because I have to debate Eric,” right? So the grading, right, it has to be a process. You can also reward different things, right, you can reward, you know, some kids are actually better at the arguments than they are at the speaking. I was actually kind of in one of those categories— better with the arguments than I was with the speaking. So, you know even in a debate tournament, you get awards for winning, you get awards for speaking, obviously, there’s, there’s some overlap between between the two of those, right? But it’s not, it’s not completely the same, right? It’s not always a student wins the most debate, like becomes the top speaker, you can give, you can give awards for working together, right? You give awards for improvement, you give awards for like, facilitating a group. Right? So it’s a little bit of leadership. So if you can go back to those durable skills, right, what are the durable skills you want to award like and it goes a little bit of it, you can think of it as a rubric. You can think of it awards, but I wouldn’t make it, you know maybe once, I wouldn’t make debating in the classroom so competitive. It’s going to be inherently competitive. So I’d kind of do everything I can to kind of reduce that, right. And there’s always, there is going to be that time, like, yeah, where you probably want to have the four shyest kids in the class, you’ll probably have them just debat each other, right? But then there are other times where you got to mix it up and, right, let the shyest kid debate with Eric and maybe they’ll come a little bit out of their shell, right? Because there’s like a little bit of ten-, like, the kids who are the shyest who are like the least capable, they’re the ones who like need it the most. Right? They need it the most, because like we all recognise these are really essential, durable skills. Right? And these kids, they need to develop them, right, we need to push them it’s just like anything, right? You’re always trying to push somebody to be a little bit better, right. And they say that’s one thing like a learning bot, you can call it a learning bot, a tutor bot, you know, they bring up different ideas, but that’s what they’re going to be really good at right. Why is one-on-one tutoring effective? A one-on-one tutor, I mean, who knows what he or she is doing, and can just push that student a little farther each time, that’s how that student grows. If you give them the impossible task, if you’re the shyest kid in the room, and the teacher is like, you have to go up front and stand in front of the class and give a five minute speech, it’s going to be a disaster, right? But if the student just gives like a short speech in a small group and just gets a little bit better, right? Maybe it builds up to the end, you know, they’d probably make a movie out of that. “Johnny, now he’s, now he’s speaking in front of the auditorium and he’s the President United States.” Probably not gonna happen. Right, Obama probably wasn’t the shyest kid in his class. You got to think through, right, like how you want to achieve the goals. And as soon as you think, well, there’s a problem. Yes, this could be a little too competitive. Okay, so how am I going to ameliorate that, right? How is my situation different than like, a competitive debate team are okay, maybe I want to have my best two students debating together because I’m trying to win the debate tournament.

Eric Byron  24:35  

Yeah, well, I think it’s really important, you bring up a great point about emphasising other things. Because I’ll admit, you know, in my experience, you know, honestly that the guys like me who liked to speak and get up there, I’d be, you know, all over it all. Let me give this argument, right. Quite often, they’re not the ones who are going to win the debate, right. They like to hear themselves talk, but they probably don’t have the best argument. It’s the more quiet one who’s thoughtful…

Stefan Bauschard  25:06  

Yeah, it made me think about when you sit there like to hear themselves talk, you also have to listen. Right? You have to listen. And then you have to, they call this, there’s a term for this now, I heard it in a Lex Fridman podcast, it might exist elsewhere, called steel manning, where you imagine what the best possible version of your opponent’s argument is and you respond to that.  Right? Because that will give you the best possible response. And now, but if you’re the person who just wants to hear yourself talk and you’re not going to listen, you’re certainly not going to steel man. So yeah, there’s a lot of, there’s a lot of other skills you need. And you know, that’s why I think it’s important to bring it back to what are your objectives. Your objectives, like I say, yeah, you’re kind of stuck a little bit, the kids have to learn this content. And you know, some teachers are really into it, they became history teachers because they love history, so they want to teach the content, right? You also want the students to develop these skills, that there’s a lot of skills you learn in debate. I mean, people debate stands out is like being this competition, you know, this clash, but there’s so many other skills that students learn and develop over the course of debate. Yeah, and you think, okay, well, man, you know, you may even observe in your classroom, well, you know, I thought the students, you know, they really kind of got the research part of this down, but they’re not really working well, they didn’t work all that well together. So maybe the next time I use debate, or even just some other instructional method in the classroom, I really need to kind of find a ways to kind of reward like more cooperation, like, you have a lot of control. But if you always keep what your learning objectives are in mind, you’re gonna have more than one, then you can draw that out. You can structure the debates, however you want. If I take students to a debate tournament, okay, “This is the structure of the debate, there’s a tournament, people are trying to win.” But in the classroom, I can do anything I want. I’m the teacher. And no one’s telling me, you know, the principal is not going to come in and say you have to organize the debate this way, right? So it’s like, organize the debate to reward what you want to reward.

Eric Byron  27:03  

All right, let’s talk about resources. Because all right, I’m in right. I agree with you. This would be a great element in many classes, particularly if we start them young, and got to train them, they get comfortable, they get used to this all the way through school. Now they’re really building durable skills. So, but I’m a teacher out there who also likes this idea. Where do I go to get started? Where do I find the resources, the tools to help me implement DCI in my classroom?

Stefan Bauschard  27:34  

You know, there’s a couple of things that are online. So there’s the Chicago Debate League and the Boston Debate League, they really tried to help schools like put this into the classroom. And New York City Urban Debate League started a program this year to try that in a couple of different classrooms. There’s some resources online where you can learn how to structure a debate, there’s some books about debate in the classroom. I think the second thing is there’s also this thing called ChatGPT, which is really, really good. Okay. I did you know, and maybe I’ll send you the link, if you want to include it in the podcast. Before I started working on the paper, I just asked it a couple you know, and you know, obviously, you know, the better your prompts, right, the better you get, but said I am a you know, so situation, right? I said, I am a teacher in a classroom where my students are not that engaged.” That’s my problem. Right? “I want to have a debate. I don’t know anything about debate. So first, can you give me some sample formats for debate in the classroom? Second, I don’t remember the exact orders of these, but I’m teaching history and I have to teach this unit on the Constitutional Convention, right. And the original learning material, the thing I’m supposed to be following says, these students are supposed to learn how enlightenment thinking impacted the Constitutional Convention. And I pulled this from the internet because I wanted to take an actual lesson and then make it into debate in the classroom. So when I saw this, and it had, and now there was an explanation of how it did it, there was some multiple choice quiz. I was like, Oh, my God, I can imagine here being a kid being like, “I don’t really care.”  Right? But then, you know, if you take it, you say, Okay, well, what are the ideas? What are the different people? At constitutional convention, some of them are influenced differently by some of the different ideas. Who was? And why did why did some people of course clash with the Constitutional Convention, because they brought these different ideas in there, set up a debate, right? I’m one of the players, you’re one of the players. Right, so that’s how that could work in that context, so dictate your lesson and contextualize it. And really when I loaded, honestly, when I loaded up the code interpreter and ChatGPT-4, it really broke it down because I offloaded that PDF. And once I got it to where I kind of generally started to say okay, how am I going to break this down, right? How am I going to break this down? Then I uploaded what was already there and said like, “Please retain the learning objectives.” Right, because like I said, for whatever we think that teachers have to teach these things, right? So, “Please retain the learning objectives. My students are new to debate so they need to give short speeches, so my format needs to include short speeches. Please explain all this to me, because I don’t know anything about debate.” Okay, and it kind of provided that thing. “Please provide a sample rubric that I could use to evaluate individual students.” And then I said, “As in a group, please give me some suggestions for how I could break the class down, you know, how I could organise the class.” And sure, I mean, some of the ideas are better than others. I know that because obviously, I know a lot about debate, but I’m saying a teacher could obviously follow up with any of those. And of course, like anything you’re learning, okay, well, you’re gonna try this, that part didn’t go, well, it’s gonna take them a little bit of time to learn, they’re not going to become a magician, right? Like, I even think, “Okay, well, I mean, I have a lot of experience in debate, but I don’t really have much classroom teaching experience.” So if I went into a history class and tried to use debate, as much as I know about debate, and even though, you know, I could skim the book and learn the history, right? Like, it would just jog my memory? So, I’m sure it would go okay, the first time, but it wouldn’t be like, amazing, because I haven’t used it in the classroom, right? That much. It’s like, in the traditional classroom, so I would, I would learn it has to be iterative. And, you know, a teacher needs to embrace that with open mind. But I think you know, and of course, we’ve seen different perspectives on this, right, but like, the reality is here, it’s not going anywhere. It’s getting a lot better, it’s going to change the world, it’s going to change your classroom. If it doesn’t change your classroom, then I don’t know what relevance your classroom is going to have to the world. So that’s all going to happen, right? So just, it’s a way to, you know, if you don’t use debate, you don’t have to use debate, you’re gonna have to learn something else, right? So you could try that but ChatGPT can really help you because it can help you with that dialogue, and the individual follow up conversation. As far as stuff like I say, just on the internet, if those debate leaks and things and you can, you know, Google debate in the classroom, of course, you know, people are welcome to contact me, I’m happy to help them. So, you know, there’s a lot there to use that. And, you know, like I say, have reasonable expectations. I thought, you know, the professors said, well, the first time they debated they, you know, they just read stuff. It didn’t go that well. Yeah. The first time they played football it was pretty bad. They dropped the ball.

Eric Byron  32:09  

Yeah. Anytime I’ve introduced something new in a classroom, I’ve always assumed the third, by the third time I do this, I may have it pretty good, right. First one is, you know, I’m just, you know, this is an experiment, right? It may go terribly wrong. And I have that expectation that I may not have enough content, I may have too much content, you know, the timing may be off, the student may not get it. They may hate it, you know, whatever, right. But you try. And by the third time, it’s starting to get better. Right? 

Stefan Bauschard  32:43  

Yeah, and if kids debated more, it would also help the teacher. Like, you shouldn’t just be debating once or twice in one class, right? Like, think of a debater. I was like, you know, I remember in ninth grade, our teacher, like had a debate. She didn’t know anything about debate but once a year, she set up this thing to have a debate. I’m like, “Okay, well, I got this.” I can debate against people who have never debated before. Like, I got this, right. So like, you know, any debater, they’re excited when once a year now that maybe the teacher has them going to debate in the classroom. But if kids debated more, and they were regularly more engaged in class, then this wouldn’t be so hard. It’s just like any kind of shift. It takes, you know, experimentation, effort, it takes time, it takes iteration. It’s all good, there’s a chance we can improve. You know, obviously, there’s a lot of good things going on in education, but at least in the United States, it’s, it’s also the reality that 50% of the kids basically don’t make the National Assessment of Educational Progress and just basic, like math and English skills, right? So there’s things we can do better, you know, we could probably take some chances and let them debate and have some things come out. So I think there’s a lot of opportunity to do that. And, you know, it’s not just time, I mean, it’s money, but it’s not, it’s not really that, like, teachers might need some training, but schools have professional development budgets, they buy books, like, chatGPT, you know, even even three, five, which is free, it does pretty well, you know, it is interesting for you know, of course, I use the code the code interpreter so it would help more, but you know, four is better at like organizing things and breaking them down. But three is, three will get in pretty far.

Eric Byron  34:23  

I don’t want to take your whole evening here. It’s…

Stefan Bauschard  34:27  

I could talk about debate, I could talk about debate all night, or evening. I can talk about AI a little bit, right. No, I can talk about these two topics forever. But this is great, because other podcasts I’ve done have mostly been about AI and you know, what it means for education and, you know, what is it those kinds of things which you know…

Eric Byron  34:47

And I did listen to a couple of podcasts you did. 

Stefan Bauschard  34:50

Yeah, but it was great to really just talk about because, you know, it’s interesting, you know, obviously, when this came out, I got very involved with this AI stuff, But once, you know it was really about a month ago, six weeks ago, I thought, I got more excited about debate again. I mean part of it, I’ve done this a long time, right. So you get a little bit burnt out doing something. But I’m like, debate is one of the answers. And that’s the title, you know, part of the title of the paper. Because people, and I think it’s fine, people are looking at AI, and saying, “Okay, well, the world is going to become like AI. So what AI’s are we going to use?” Well, I think it’s important to think about that, because AI is irrelevant in the world of AI. AI’s can help teachers in the classroom, but there’s a lot of things we’ve been doing in education. And, you know, portfolios is another example, right, that transcend whether we have AI. In the 1800s people went to the town square, right before the industrialization, right, in the early 1800s. Before the industrialization, before the information economy. They went and they debated and they argued, and they talked and they thought about things, and they learned some content. This is work, you know, back to the Romans, right? Like it worked then too. Right, we’ve done this. Yeah, Cicero, right. I mean, this has been going on for thousands of years. So we should keep doing, we don’t need to change everything about education because of AI. We need to think well, what what is what are the most important things to do that we already know how to do? So, will classroom debate be a little new for a teacher? Of course, but it’s not as new as an extra prediction. Right? Okay. And then all the changes that are going to come like in the next five years, that they’re going to try to keep up with, that’s just not going to happen. Right. So, well…

Eric Byron  36:34  

Well and we all realize, right, we’ve all been debating, whether we’re formally debating on a debate team, right? I mean, this is a life skill. Too many people these days, just want you to tell them what they want to know. Right. So you know, tell me what I want to hear. And I’ll accept it as fact. Right? Yeah. I don’t want to challenge the source. If you’re telling me what I want to hear, then it’s all good. Right? And it’s a real problem because the bots and anybody else out there who wants to be popular, right, will tell you what you want to hear.

Stefan Bauschard  37:07  

Yeah. Yeah. Yeah. I’ll figure out what they want you to know in a way that you want to hear it. That’s what’s gonna get a little bit, yeah…

Eric Byron  37:15  

All right. Well, um, one last question. You’re, you’re in New York, right? Are you a Yankees fan?

Stefan Bauschard  37:23  

Not really, because I’m originally from Boston. I spent a lot of time in Boston.

Eric Byron  37:28  

Oh. Yes, yes, I am a huge Red Sox fan. All right. Let’s stay in touch. 

Stefan Bauschard  37:35

Sounds good. 

Eric Byron  37:46  

Well raise the roof, drop the mic. I don’t know what the current expression is to say how much fun that was. I think if I hadn’t cut that conversation off, we could have talked for hours. Stefan Bauschard is a very interesting guy and I love the concept of debate-centred instruction as becoming a common approach to active learning that kids practice regularly starting from a young age. If you’re enjoying this podcast, please subscribe and share. Up next is Dr. Sean McMinn, the director of the Centre for Education innovation at the Hong Kong University of Science and Technology, a must listen-to conversation with tons of practical advice and examples from his experiences within the university setting. I’m Eric Byron. Thanks for listening. And thanks to all those education innovators out there, you are making a difference.