– [Olivia] So I’m Olivia Fleming, I’m one of the co-founders of OneHE, so I’d like to welcome you all in today and thank you very much for attending. I think today is going to be a really interesting session. So we have Todd and Lew here. So Todd is the associate professor in the University of North Carolina Chapel Hill, written numerous books on teaching and learning and the science of learning. And we have Lew here, he is a maths professor from Denison University, so which is in Ohio. For those who might be attending outside the US, I’m not so sure. So I’m gonna hand it over to the two guys because they’ll be very interesting and entertaining. And I’d like to say thank you so much guys for your time and your work here.

– [Todd] Sure, thank you. I sure sounds good. So we’re gonna kick it right off and just get rolling. We don’t have a whole lot of time here so we’re gonna kind of move quick paced here. Wanted to start real quickly with an issue that’s happened for a long time in higher education and teaching in general. I guess it’s even outside of that is that concept of an illusion of consensus. So we often think that other people think the same types of things we do or think similarly to us. So when we say specificity words or numbers, like you know, get about what you say you want three to five references and how many times I’ve heard faculty members will say, “I want five references.” And then I’m helping them and I say, “You know, when you’re grading it’s like, this one didn’t seem like it’s got a lower grade.” And they said, “Yeah, they only did five.” I said, “But you asked for five.” And they said, “Well yeah, they should have done more.” So that concept of you should know kind of things. What is respectful? I was having a conversation with a student one time and it was really interesting that my phone rang and I reached in my pocket and turned it off. And he looked at me, he says, “Aren’t you gonna get that?” In my day, if you’re talking to someone and your phone rings, you don’t stop talking to them and pick up the phone. And so I said, “No, I’m talking to you.” And he said, “You should get that.” And I said, “Well, I’ll get it when I’m done talking to you.” And he said, “What if it’s an emergency?” And I said, “Then the person should call 911.” And I realized real quickly that he thought it was rude that I didn’t look at the phone. I thought it would’ve been rude to look at the phone. So this concept of what is respectful and what is appropriate. Expectations of assistance. How much assistance should you get on something? I’ve had students who are charged with plagiarism or cheating on exams when they’re working with groups. All semester long and they’re working in groups all the time and then they have a take home exam. Teacher says, “Hey, did you work with somebody else ’cause two of you got the same answer.” And students says, “Well yeah, I worked with Joe on this.” And the teacher says, “Well, that’s an F because you’re supposed to do it alone.” Student says, “You never told us to work alone.” Teacher said, “You should know better.” Those kinds of concepts where we think what’s going on and what other people are doing is the same. I think we have to be very, very careful with that. Even when we say things like, “I don’t accept late papers,” I can’t tell you the number of times that I’ve heard faculty members say, “I don’t take late papers,” and then the student comes in with a really, really good excuse and the faculty member say, “Okay, I’ll take that.” And then a teacher will come along to say, “I never take late papers.” Student tries to hand in lateness, faculty member says, “I told you I never take them.” And the student says, “Yeah, other people have said that too, but I have a really good reason.” This idea of what we’re doing. If we don’t talk about it, then we have to be careful because the students don’t just know. So the next thing I want to try is this real quickly, this concept of the specificity of sometimes words and how this works. If we consider this real quickly, I wanna just try something fast in the chat over here is think for a second if, if we have somebody who’s like Chris who’s interrupting someone. And we say, Chris always interrupts, that means 100% of the time Chris is interrupting. Every conversation that Chris is in, he interrupts. At least once, he or she. If we say that Chris never interrupts, it means 0% of the time. So the first thing I’d like to try real quickly and think for a second before you actually type this thing in. So think about a second and then you can type it in. Let’s go with what percent of the time would you say often? If I say a person often interrupts, think about a number for just a second. And what percent of the time would you consider to be often? Okay, just a second for you. All right, go ahead and type in your responses and we’ll see what kind of numbers you get here. Oh, this is perfect, thank you. So this is what I’m really after and this is why we have to be careful. So we have Kim in here at 80, we’ve got a 75, a 60, 75, here’s a 50, a 33%, a 40, 40, 70, 60, 50. This is why we have to be careful. If we say to the students, we should often do something or you should look at, make sure you have a sufficient number of citations. Anytime you were use words like sufficient and often or rarely, we have to be careful because in this example here we have numbers that range. Think about this for a second. The numbers range from 33% to 80%. So somebody who says often and thinking 80% and somebody else is saying often and thinks 33, a third of the time is often, then that’s huge. Same type of thing I mentioned before. If I say make sure you have an adequate number of references, are we talking in the range of 80 here or the range of 33? And so I remember one old assignment teachers used to say all the time is you’d ask, students would ask, “How long should it be?” Teachers, “As long as it needs to be.” The teacher has an idea in mind. So the concept here is we just have to be careful whenever we’re doing that because we can, it runs into it. So if we say something like, “Some of the time I’m not gonna do it again,” it’s the same kind of thing. We run into the same thing. So the whole thing I’m after here, the point of this whole exercise real quickly is that we wanna be careful with language because the specificity of it is something. I wanna pinch something that’s kind of funny. Coming to North Carolina, it’s interesting because it may sound funny but people hold doors here inordinately long periods of time. And so I went to ChatGPT, by the way, here’s a citation that a student could use if you say, if you use ChatGPT or Dolly, make sure you cite it. These are the kinds of things that Lew will talk about in just a second. But I wanted to point out, it’s kind of funny, is this concept of how long before you should like, how long do you have to hold the door open? First image I got of course is the person’s walking through it in front of the individual who’s holding it. So then I thought, well, I’ll try a different prompt here. So I went with this one was, please draw another image with a second person further away and the individuals are entering the building. And I did this one, which I then said, hey, entrance is spelt incorrectly. The individuals that you’re holding the door for are in front of you and could you draw another one? And it came up with this, which was nice because now it has entrance spelled correctly except the individuals are still in front of them. The reason I wanted to point this out real quickly and I can tell you the difference, by the way, is if you’re in Michigan and upper East Coast, it’s about three feet behind you, you’d hold the door, otherwise you’re not. In North Carolina, if you make eye contact, you can be across the parking lot and they expect you to hold the door. But what Lew is gonna kick off a little bit too is the concept here is that you’ll notice on the prompts that I’m doing. I get longer and longer prompts, more specific prompts. But as we explain things and try to do good prompts, keep in mind that this has always been around. When we talk to students, when we talk to colleagues, the kinds of things that we explain sometimes we are not as clear as we think we are. One of the things I like about AI is I’ll explain I want something and it gives me something totally different and that’s like, hey, that’s not what we are really after. And so it helps with kind of figuring out where you’re at. So overall, it’s imperative that we consider the implications of vague language. When it comes to exams, be very careful ’cause students may or may not understand what’s appropriate, not appropriate. Acceptance of late papers. Will you take a late paper? If so, under what conditions? And I will tell all of you, you can have any, any policy you want for late papers. But I’m begging you as a colleague, whatever it is, stick with it. If you say no late papers, it doesn’t matter. No matter what, you don’t take ’em. If you say, “under extenuating circumstances,” then you gotta be careful ’cause that’s a bit vague. So be careful with those types of things. And outside sources, when you say “It’s okay to have somebody work with you on the paper,” now I’m going pre-AI, it’s okay to have like a family member help you. It’s okay to work with a friend. It’s okay to work with other person. You have to do it completely alone. You have to make sure you cite all your references. All of these criteria that we put on papers that often we never said, it’s just do it. And then, the next thing you know, students charged with plagiarism, a very egregious offense because they didn’t understand what you meant with “You can get some help.” And so when it comes to AI, this stuff all goes on like high octane. And so the question would become: How would you get started with students and talking about AI and specific language too? But what about a new topic that’s coming along? Lew is gonna take it away here.

– Thanks, Todd, so I think Todd did a really nice job of kind of setting the stage that historically, we’ve always had trouble with our students and kind of, you know, making sure they understand what we expect and we know what they should expect just because, you know, that lack of conversation that we often have. But what I wanna talk about today is how can we have a productive conversation with AI, with students about AI. ‘Cause clearly, AI, if you think about it, it’s kind of like the 800-pound gorilla in the room. So this was an arrival technology. We did not ask for it. It was just kind of unleashed, right? Everybody had access to it. There was no vetting, there was no testing of this thing. And everybody all of a sudden had access to this, you know, very powerful technology. So as faculty members, we were really kind of taken aback ’cause we’re used to being the experts in the classroom and all of a sudden we’re on the same, I dunno if I’m gonna call it level-playing field, but the same playing field as our students are. And the problem with that is that we need to kind of have those open conversations with our students about what we expect from AI. The problem is we’re so new to AI, we’re not really sure what we should go about asking. But I do wanna point out again that your students are turning to you. They want to get some feedback from you, some insight like Todd was saying about what you expect. So Marc Watkins pointed out, he said, “Many students are still trying to figure out how to best use or not to use AI and turn into faculty for advice. Many faculty may not realize it, but having discussions with students about generative AI can deeply shape how students view their agency and voice with this new technology.” So what we wanna do today is actually talk about how you can have those conversations, but specifically we’re gonna do it uniquely for your own classrooms. ‘Cause I can’t tell Todd how to do this. He can’t tell me how to do this. I can’t tell you how to do it. We’re gonna give you some examples and some frameworks that you can actually develop this for your own specific courses. And what we’re gonna do today is create something I like to call an AI engagement scale. And this is from work from Ryan Watkins. This came back in July of ’23, which in AI world seems like forever ago. But what Ryan did is he kind of created some surveys like a 10-point scale of given a specific assignment or exercise that students had to do, how much could they engage on AI on an increasing scale? So let’s look at an example that he had. So here’s an example, an essay example that Ryan had, and here’s the directions that we’re gonna have. At which point on the progressive list below, could a student use AI to assist in completing an essay assignment without crossing the line that violates academic integrity in this course? Okay, so I wanna give you a scale, one through 10 and here’s what he has. So idea generation, title suggestion, all the way down to number 10 which is the complete essay writing. So what Ryan had in mind here, if you landed on a number, say four, then you were okay with one, two, three, and four, okay? So you have this kind of progressive scale of how students could use AI. And then once you landed on a number, you agreed with that number and everything prior to it because it was an increasing scale. Now again, this thing is not perfect, I understand, but let’s give this a try. So we’re gonna do a similar thing that Todd did in one of these zoom cascade things. So here’s what I want you to do. I’m gonna show you the list here again in a second. I want you to land on a number. You’re gonna type it in the chat, but don’t hit Enter yet. We’ll wait until the count of three, and then we’ll all hit Enter and we’ll kind of see that scrolling by a waterfall or a cascade feature, if you will, to see how we responded there. So again, here’s our list. So land on a number. If again, if you pick five, that means you’re happy with one through five for an essay assignment on how students would use AI. Type that in the chat and then in a second, we will all hit Enter. So please choose your number. Oh, and by the way, no judgment here, right? No judgment, we’re among friends. So, all right. And again, not a perfect scale, I get it. Some people are like, I want one, two, and seven. No, no, just pick the number. On the count of three, here we go. One, two, three. Ooh, we got some fours, we got some nines, fours, a seven, a four, a five. Nice, I didn’t see any zeros. I didn’t mention that. I guess you could go with a zero, right? Like I’m not buying into any of this. So there you go. Okay, so what Ryan did in this article, he hit not only had the essay example, but he had five other ones that he shared. He’s very generous. You can follow the link. We’ll provide these resources later. But he actually provided the Google form that would actually generate this for you and you can adapt and modify. So he’s very generous with that. He did one for testing, he did one for a lit review, a collaborative project, a coding project, and even a data analyst or data analysis research project. And you’re like, great, that’s a great idea but I don’t teach any of those. How’s that helping me? So what if these examples don’t work for your class? So what we’re going to do is we are actually gonna think about how you could specifically create one for your very own class. And here’s how we’re gonna do that. We’re gonna use AI to actually create that engagement scale. I’m gonna be using ChatGPT today. You could use whatever AI, generative AI works for you. But I will be using that and I’ll show you that in a second. And here’s kinda the steps that we’re gonna do. First of all, we’re gonna go to Ryan’s webpage and basically just kind of cut and paste this different survey examples that he had. Again, I’ve spoken with Ryan with this. He’s very happy to share this work. So we’re gonna share that with the AI. Then, we’re gonna get very specific with what we instructions we wanna give AI to do, as Todd was saying, to make sure that we, you know, really kind of lead this thing along with our prompting. We’re going to give it a roll. We’re gonna tell it about your class. Then we’re gonna tell it what format, what you want to produce and what format. And then we will actually share the assignment with AI. So let me go ahead and see if I can share this with you folks. Let’s see. Todd, can you give me a thumbs up is this looking okay this screen? You good? Awesome. Okay, so here we go. So this is my ChatGPT screen. And here’s what I did. I said look at the AI engagement scale Ryan Watkins created by various college level assignments. That is for each assignment, how much AI can a student use to assist them? After you look at and understand his example, I would want you to create a similar scale for one of my students. And then I literally copy and pasted the survey questions that Ryan had and pasted those in there. And as you can see, here’s the first one and there’s the scale. AI does not care about formatting. It’s very good at scraping the web. So things can look cluttered to the, you know, human eye. But AI can read this very handily. So there it is. The six different prompts that he had. And then it says, “Okay, hey I’m all set, what do you wanna do next?” So now is where I’m gonna lean into this a little bit more and then specifically ask AI what I want it to do. I want you to act, I’m a mathematician. I want you to act like an expert in mathematical pedagogy at the college level. You will help me develop a scale for the various ways that students can use generative AI on a related rates question. That’s the type of calculus question that I’m gonna ask it you. Your work will look similar to Ryan Watkins. That is you will create a scale of how students might use AI to assist them in solving its problem, but not actually solve it. ‘Cause if you think about Ryan’s essay question, he actually had ’em right at the end. I don’t want to quite my number 10. I don’t want it to go quite this far. And then I say again, hey, don’t solve a thing. And then I actually go through and here’s my very simple assignment that I give to it and it’s a laterally and against the wall kind of problem. And then off to the races, ChatGPT very quickly creates these scale of 10 for me. So problem understanding, diagram assistance, identifying unknowns, problem-solving strategies, and then we’re kind of off to the races so to speak. So again, what that does for us is it gives us kind of a very concrete way, sorry, get my right slide here. It gives us a very concrete way that we can now talk to our students about something like here’s an assignment and here’s the scale. Let’s vote on where we’re at on this. And they can, you know, maybe do that unanimously or anonymously and then you can have that kind of concrete or that practical illustration that you can talk with them and then work your students through how you might go about doing this, okay? So this is very good. I think it’s a really good start to kind of get students in that conversation. But just like Marc Watkins says, our students are still looking from guidance from us. They are still very new to AI as well. And if they just cut and paste, right? AI is gonna get them in trouble ’cause AI’s just gonna basically be very kind and dump the answer out to them, and that’s not what we want. So what we need to do is we need to help our students with their prompting as well. How do they ask a level two question or a level three question of AI? And you might be saying yourself, well, I’m not so good at that myself, Lew. Guess what we’re gonna do? We’re gonna use AI to create those prompts. So let’s give that a shot. So let’s take a look at what that might look like. So what I have here is I say, hey, below is my AI engagement scale that gives increasing levels of how students can use AI for the next assignment. Here’s what I love. Unfortunately, my students are not very good at writing prompts. For each level of the AI engagement scale below, can you, yes, it’s a typo, create two different cut and paste prompts they could use in order to have AI help them with that level of engagement. And then I give it the scale and then again off to the races. Here it goes. Problem understanding diagram assistance and it gives a prompt for each one. Let’s look at this one here real quickly. Diagram assistance. This is the prompt that the students can then cut and paste. Guide me in drawing a detailed diagram for related rates problem where a shadow is lengthening as a person walks away from a streetlight. Again, a very classic mathematical example, but it’s not the example that they were doing for their assignment. Their assignment was a ladder thing. It’s similar but not exactly the same. And that’s what I love about this. So what AI is doing, it’s almost acting like a tutor. So if a student went to a tutor and said, hey, can you gimme some help on this? The tutor wouldn’t do the problem for them, they might do a similar problem. That’s exactly what AI’s helping the student and kind of in doing that prompting. So again, this is a kind of recap here where we’re at on this. So what have we done? So the AI engagement scale. So how do we do this? First of all, we use AI to help create that engagement scale for us and that’s gonna be specific to our individual assignment. Then we tow through and take that scale and we rate it ourselves where we, you know, land on a four or a five or however comfortable we are with that. Then you can give that scale to your students and let them anonymously assign that a scale for that as well and kind of vote on where they land on those numbers. And then you have that productive discussion in your classroom about your assignment, specifically how students might use AI for that particular assignment. And then you can go through and develop guidelines and you know, kind of share some promptings, some guided AI prompts that students can use as well. So not only are you having that open discussion of what you expect from them, but you’re also giving them some training on the prompting and then they can cut and paste and they can modify and they can adapt these prompts to make them their own. And of course you can revisit this more than once. Now I don’t know if I would do it for every assignment but maybe every other or maybe you know, every two weeks or something like that. Because as this technology continues to change, we’re gonna have to change and adapt with it and to keep repeatedly coming back to it and having those conversations is gonna be very, very important for us as we move forward in this new AI world that we are living in. Okay Todd, I think that’s it. I’m gonna hand it back over to you, my friend.

– [Todd] There we go. I keep getting unmute there. Thank you, that was fantastic. I love the specific example and I have to admit I learned a lot from Lew as we go along too. And this whole concept of asking AI, what we should ask AI so that it can give us better stuff and then we can kind of formulate from that I think is fantastic. I only got a couple things to kind of wrap up with here. What I would say is, let me see if I can, there we go. Ah, let me get that in a second here. Is number one is one of the things that Lew did was this prompt engineering. He’s basically showing us also what we can do. So when you’re working with students, one of the things we have to do is we have to become more and more sophisticated with AI, which is a little scary. I like to point out that, you know, technology has changed for like a long time all the way back to good old Socrates saying you shouldn’t write anything down because it weakens the mind and people come along who say you shouldn’t, you know, with books. You know, we don’t have to have strong minds anymore. We can just read it. All this stuff has come along. Internet is the same thing. Each time, what it boils down to is as faculty members, we kind of help to guide how this moves forward, but we have to learn it as well. And so I think this is really important to have that. So it’ll help us in terms of coming up with better prompts. Love that about what Lew said. I do wanna circle back to the very beginning, I was talking about specificity and I think it’s interesting that when Lew put the scale up there and said what level is okay, the numbers came out to four, between four and nine, which are huge, huge differences. Which means if I said it’s okay to use AI a little bit but don’t go overboard and then a student turns a paper in, I’m thinking four, they’re thinking nine. I call ’em into the office and say, “This isn’t appropriate.” Now here’s the reason that’s important. The whole session’s about student agency, you can’t get agency if the person who basically has control over the situation is telling you that you’re wrong. I mean, how do we get somebody to have agency when they do something like, what I just mentioned, they turn in to use AI to the scale of an eight. I’m thinking a four. I say, “You shouldn’t do that.” And the next teacher comes along and says, “You can use AI.” And the student saying, “I don’t even know what you want.” And there’s something called outgroup bias, which is cool. Because in their minds, we are all the same. I don’t know what you teachers want. And so keeping that in mind I think is really important. And that specificity and that comes down to the agency component. There is one other small component to it. There’s really good research out there that individuals that have more structure in the class do better than when you don’t have structure if you give some amount of structure to it. The really interesting thing in the research shows is that students from underrepresented groups and first generation students disproportionately do worse when there’s no structure. So these concepts of students should know better. That’s different for a person who has had parents, uncles, grandparents all have gone to college because they know aspects of college because they’ve lived it, they’ve heard it, their parents will talk to ’em about it. I was a first generation college student. I didn’t know when I started school that it was permissible to drop a class. You couldn’t drop a class in high school. And so I was flunking a class and then I got in trouble with another class and I thought I’m just gonna flunk out. Advisor said, “Well, just drop the class.” I said, “I didn’t know you could do that.” And his basic response was, “How would you not know that you can’t drop a class?” That’s at least a double negative in there. The point here is with AI, if we’re not careful in working with the students and setting this up, we can’t create agency for them and disproportionately for students in underrepresented groups if we don’t talk to them a bit about what we mean by this things. And Lew gave an excellent example for this. And I think as time moves forward, we’ll get more and more of those types of things. I have been using a slide for a while that was kind of a long and winding road. Like today’s session is just getting started and there’s so many things that are gonna be happening, and this is what it’s always looked like to me in the past. The dilemma is that things are slightly different now. So instead of being this long and winding, it’s gonna be okay kind of road. I mean just to be absolutely fair and upfront is things change so fast. And Lew’s got the number better than I do here. I think it was when it was, how long it took to get to a million users. I think it was Netflix or was something around two years and ChatGPT was five days. I think 10 million users was a couple months. It’s just an incredible speed. And the pacing now, which things are changing. I think we just have to keep in mind it’s moving fast, and as Lew at the very beginning, we can’t control that. It’s here, but what we can do is help the students as we work through the system. And to wrap up, I guess I should say we probably should have done at the beginning is we’re fully aware of the resources and everything to AI. I mean there’s lots of controversy and discussions about what this does. I think it’s an important conversations, but I also don’t think you can ignore something that’s gonna change all this stuff by saying I don’t know that I agree with some of the policies. I think we have to figure out ways of talking about the policies, what should be done along with helping to oversee the system, so to speak it together in a way. I think we’re just about at time. So I’m gonna kick it back to Dasha and let’s get rid of our screen here so we can take that right out. And Dasha, I’m just gonna take it away, or Olivia, sorry, I wasn’t sure, I should have paid more attention.

– I can just pick it up. There is a question in the chat here. Let me just pull it up and it’s from a member: I attend another similar session and they proposed utilizing this framework that has been adopted to AI usage and Bloom’s taxonomy. The lower the level of the pyramid, the more AI use is permissible a given assessment. Okay, so you’re putting that in. Amber, thank you, as a resource really. Any comment guys on this?

– I agree, Lew?

– Okay.

– No, I just. No, I think this is the way we’re looking at it.

– So Leon Furze, F-U-R-Z-E, has done a lot of work on this and he talks about the acceptance scales and things like that. So yeah, this is definitely right in line with what Leon would be saying about these things. But yeah, it is definitely a brave new world as we step into this and what’s acceptable. One of the things that Leon really focuses on is, you know, the idea of assessment and how we’re going to have to think about the validity of our assessment, and some things are gonna have to change as we do this. So something to keep in mind. But Blooms’ definitely helps out with that.

– Todd, did you wanna comment?

– No, I think that’s a really good point. In fact, that’s one of the big criticisms of higher education in some respects is in some classes, students are working at understanding or even remembering phase. And faculty are spending a lot of time with that and less time on the critical thinking, higher level things. And so we’ve long talked about the fact that we should get through that, you know, get into the higher critical thinking. Now very quickly, careful with this stuff, is Bloom’s taxonomy was never, ever designed as a race to the top. And the concept is the first couple levels are your foundational knowledge and what really, really limits what you can think and talk about is the processing speed. It’s basically automaticity and learning stuff. If you don’t learn those first couple levels, you can’t think critically later. If the words that Lew and I are using, Olivia everybody else, if the words didn’t come to mind like this because they’ve been used so often, they’re automatic, you can’t look each word up, as you’re listening, it doesn’t work. So we have to be very careful, If students do skim past those first couple levels quickly, even using AI, they’re not gonna be able to critically think about things later. So I think it’s fine to use AI in terms of those Bloom taxonomy’s levels, but be very careful about students who take the easy way. I’ve used the example for a while of if we all take cars, automobiles everywhere and nobody runs, walks, or jogs, we get to our destination faster. But once we get there, we don’t like what we are when we get there. With AI, if we use it all the time, we’re gonna get there faster but we’re not gonna like what we are when we get there.

– Thanks Todd, I’m just gonna say to anybody, you can put a quick question into the chat if you wish. I’m gonna ask a quick question. Well, we’re actually up against time, but my question is really about you talked a lot about given guidance and consistency. Just very quickly, guys, if there’s a lack of consistency across educators within an institution, what needs to be done in order to not confuse students going into different classrooms?

– Oh, students are so confused right now. There was a “Inside Higher Ed” article back in December about the student’s burden of how, you know, what to do with AI. They have four professors. Two say don’t use it within an inch of your life. One says use it any way you want. And then, the fourth one doesn’t even pretend the AI, it exists, right? So what’s a student to do? And oh by the way, the internship this summer expects them to use AI. So our students are really struggling out there trying to figure out, trying to navigate this path. You know, Todd talked about being a first gen student. Well, you look up to your faculty members. Well, what if your faculty member you look up to says never use AI in your entire life. Is that going to, you know, set you up for a productive job as you get out into the workforce? So there’s a lot of concerns we have right now with students in the mixed messaging that we need to get. And higher ed, we’re gonna have to try to get to on some kind of same discussion to have these things to move forward. And to Todd’s point, I think there are gonna be different parts where we’ll engage in AI in different ways. There are gonna be instances where we do not want AI in the room. We might want a little bit of AI in the room. And then maybe if it’s something’s more elaborate, maybe more AI in the room. But it’s never going to be AI 24/7 like his car analogy. So our students, yeah, definitely struggling out there. That’s why we need to have those conversations with them so that at least in our own classroom, right? They have some guidance of what we expect and where they’re headed.

– Yeah, and this is also why we started the way we did. It’s just this is not a new phenomenon. Students have struggled and complained about faculty being inconsistent forever. Participation in class, coming late, some faculty will lock the door at the time of the class starts. Other faculty say, “Come when you come, it’s okay.” Some faculty say, “If you come in late, make sure you sit.” My policy was, “It’s okay to come in late. But if you come in late, you sit in the seat closest to the door. Please don’t muscle your way to the middle of the room. If you know you’re gonna have to leave class early, please sit by the door so that you can leave but also let me know so that I don’t worry about you when you step up and walk out.” That’s the way I would do it for basically helping the students to understand this is what I want. Because in another class, it’s gonna be a totally different thing. And I’ll finish with this. When a student says, “I couldn’t make it to class Monday, did I miss anything?” Faculty respond to that in lots of different ways. My way of responding is, “What I’m hearing is you wanna know when you didn’t make it to class Monday, did I teach anything of value? Is that what you’re asking?” And every time I’ve done that, the student said, “No, no, no, no, no.” And I say, “Pause, I’m not being mean. Just take a second, ask me the question you wanna ask me.” And then they’ll say, “Okay, I couldn’t make it to class Monday. Is there anything specific I should know to be prepared for class Wednesday?” And I will say, “That’s a better question.” All of this stuff is we have to help the students understand what they’re saying, what they’re hearing, and for us to do the same type of thing.

– Okay, that’s great. And I think what I take out of listening to you guys today is very much take control. Once you take control, and then you’re guiding the AI instead of the AI guiding you. So guys, thanks so much. Really interesting as usual and interactive. Chelsea, we are recording this so it will go up on the platform. And these Lew and Todd are collaborating on a book. So watch this space, maybe towards the end of the year, we’ll see some interesting work come out of the two guys. But again, thanks so much for your time and everybody else for joining and enjoy the rest of your day.