Category: edtech

Show Notes

• SUNO playlist of 16 songs on thermodynamics
• SUNO
• Tom Lehrer, The Elements Song (video)
• Neil Garg, Garg Lab Music Video Hall of Fame (2010-2015)
• Dirk Mateer, Using Media in Class
• Dirk Mateer, Dirk’s Media Library
• Mateer, Dirk (2024). Economics in the Movies. Cengage
• Kent Kauffman, (2026). Intellectual Property. Tea for Teaching podcast. Episode 457. August 5.

Transcript

John: Providing students with multimodal content can increase student learning, but creating such material can be costly and time consuming. In this episode, we examine how AI tools can be used to generate songs and lyrics to support student learning.

[MUSIC]

John: Thanks for joining us for Tea for Teaching, an informal discussion of innovative and effective practices in teaching and learning.

Rebecca: This podcast series is hosted by John Kane, an economist…

John: …and Rebecca Mushtare, a graphic designer…

Rebecca: …and features guests doing important research and advocacy work to make higher education more inclusive and supportive of all learners.

[MUSIC]

Rebecca: Our guest today is Oliver Steinbock. Oliver is the Cottrell Family Professor in the Department of Chemistry and Biochemistry at Florida State University, who specializes in research in physical chemistry and materials science. Welcome, Oliver.

Oliver: Thanks for the invitation.

John: Today’s teas are:… Oliver, are you drinking tea with us today?

Oliver: Yes, but I must admit that I’m usually a coffee drinker, but for this special occasion, I brought some ginger lemon tea.

Rebecca: Oh, that’s a great choice!

John: It is.

Rebecca: I have blue sapphire today, John, because I am recording from home, so I was able to make a pot.

John: …which you haven’t had in quite a while on the podcast. It was one of your old favorites.

Rebecca: I’m excited to have some today. [LAUGHTER]

John: And I am drinking an Earl Grey tea today.

Oliver: Good choice.

Rebecca: Good classic choice. We’ve invited you here today, Oliver, to discuss your recent experiments with using AI to generate songs that explain thermodynamic concepts. Can you talk about how you decided on this approach?

Oliver: Yeah, I’m very interested in AI and integrating it into my research and teaching. So I try to keep up a little bit with generative AI. So, AI methods that can generate text, music, pictures, and you see these fantastic videos of fake songs, [LAUGHTER] fake music videos. So I thought, okay, I want to play a little bit with the music generation part. And so in the beginning, I kind of just took abstracts of my papers, very technical text, and turned it into Korean pop songs, K-pop, [LAUGHTER] and I was kind of amazed how well it worked. I mean, of course, ridiculous, [LAUGHTER] but you can take very technical text, and it works surprisingly well. So, okay, that was maybe day one, and I shared it with some colleagues who also played a little bit around with it, and I’m teaching a lot of physical chemistry upper-level undergraduate classes, and so I thought I can take just an equation or some concept and see whether I can turn it into some kind of catchy song, and so I started playing around with that, and that’s how the whole thing started.

John: And we will share a link to the playlist that has those songs in it, so they will be in the show notes.

Oliver: Yeah, please listen to them. [LAUGHTER]

John: What was the process of creating these songs? What tools did you use to create the lyrics and then to create the actual music and the singing?

Oliver: Yeah, so there are many different ways how you can do this today. But what I did kind of converge to was that I used ChatGPT or a similar large language model, and I had an idea, let’s say I wanted to do the first law of thermodynamics or some kind of fundamental thing,
and/or an equation. And so I said to ChatGPT, “I have this idea. I would like to turn this into a song. Can you suggest some lyrics?” And I gave a little bit of guidance because part of the problem is you have to find the right balance between scientific rigor and technical rigor and making it kind of a little fun. I mean, whatever fun means to different people. So you walk that line a little bit. And so, I prompted it for lyrics, and it’s just amazing. These large language models-they are so supportive, let’s say. [LAUGHTER] They’re like, “Yes, the first law of thermodynamic is an amazing idea, and I think it would work well as a country song.” [LAUGHTER] And I think it doesn’t make any sense really. So, I generated the text and some mild instructions in there, like chorus lines and stuff, and then you transfer it over to a AI music web page. I used SUNO without advertising for it, and so you essentially just copy paste the text in there. You give some guidance on the music genre, let’s say K-pop. You can pick whether you want a female voice or leave it up to them, or a female-male duo, and then you have essentially only two things that you can select. It’s like how closely it adheres to the style and how weird it is. [LAUGHTER]

Oliver: The weirdness is interesting because I wouldn’t say that I fully understand, but it gets more interesting, but sometimes slightly awkward if you dial this up. And then it generates, really within a few seconds, two songs for you. And if you’re in the free level, these are limited to a minute or so, I don’t really remember. And if you have a fairly inexpensive subscription, you can get longer songs, and then you listen to it and you can play with it again. And so that’s how I converged eventually to my playlist. [LAUGHTER]

Rebecca: I love the weirdness is another line.

Oliver: The whole thing is very weird. [LAUGHTER]

Rebecca: So, be careful. [LAUGHTER] But songs themselves are not an unusual thing to find in a learning environment. Maybe not as common at the college level. Can you talk a little bit about the history of song in learning?

Oliver: I feel like there’s something to music and especially rhythm that helps people memorize things. And when we learn the alphabet, no matter what country, that’s usually an alphabet song, and there’re jingles. Maybe they’re a little bit less popular nowadays, but they can be very catchy, although they’re actually annoying. [LAUGHTER] So I would say, in the context of science and chemistry, it’s not frequently used, but probably the early classic is Tom Lehrer’s song about the periodic table, where he, at extreme speed, recites all the elements in the periodic table. It’s real fun, but it’s from the 60s, there’re really cool YouTube videos of it. There’s actually even a version where he shortens it to the four classic elements, and the song becomes very, very short: earth, air, water, and so on, fire. And so that was probably one of the memorable and remembered chemistry songs, but then more recently people played with it on social media. Usually, it’s students performing a song or lip-syncing or changing the text, but you keep the music. A person who did a lot of this and quite successfully, I think, is Neil Garg. He used to be at UCLA. He’s now at Baylor, and he used it in teaching organic chemistry, and to produce great videos. You can see the students really love it, and yeah I would say these are maybe it’s a little bit the context of where all of this work kind of sits.

Rebecca: And although jingles can be annoying, you do remember them.

Oliver: You do remember them, and you know the way I see it is many people, unfortunately, have bad memories of their chemistry classes, and they’re just like negative emotions, to be really honest. There’s not even like there’s this one particular thing that I remember but don’t like. So to maybe make it a little bit more approachable and maybe a little bit more fun, at least to some people, could be a valuable tool. So I’m not thinking of doing every lecture in some format, though, right? But to use it as a small tool and sprinkle a little bit of it in can be fun, maybe.

John: Many years ago, I had listened to Dirk Mateer, an economist, talk about using music and video clips in classes, and one of the things he had done, along with some of his students and colleagues, was creating short little videos that had the songs with the lyrics on them and some connections to economics in it, which I used to play at the start of each of my classes. He also worked on creating this economics in the movies platform, which provided information on short clips from movies that illustrated economic concepts, and I used to use those fairly regularly in my classes. With the songs, I’d start my class with them… I teach a class that back then generally had 350 to 420 students in it, and they’d come in and they’d sit down and they’d be quiet and they’d be listening intently. And I’d start the class by saying, “What economic concepts did you see here, and how might this relate to what we’re doing for that day’s class?” And that worked really well. The only problem is I got really concerned about copyright issues, so I stopped doing that. But what I really liked about this approach is that because the material is generated by AI, it’s not subject to copyright, so there is no concern about copyright infringement with using it. In fact, we had a podcast on that that was released on August 5 that discussed that very issue. So that’s a really nice feature of this that you can create these songs and share them freely with your students without worrying about infringing on any musician’s or author’s or producers’ rights.

Oliver: I would say, first of all, this is really awesome and inspiring to some extent because I will also use it a little bit more in the next semester. But I would say that there are, of course, very serious concerns about copyright infringement that occurred during the training of this generative AI, and many people very rightly point this out, and some AI companies were hit with pretty hard financial punishments. I’m not a lawyer, but for the user point of view, if you’re okay with this or the way it was settled, I would say that my understanding is that this material cannot really truly be copyrighted, and you’re the owner of the song for whatever that is worth, and that’s pretty much how much I know. But one thing I would like to add is that this copyright thing is, of course, a very important issue. But there’s also, of course, the technical musical ability to do this, right? So in a certain way, this generative AI approach democratizes the approach to it. Most people don’t have the musical skills, knowledge, the time, the money, the recording equipment to really do this, and while it’s a major shortcut, and the product is of course different from human-generated music, it is kind of opening up this to more people and more crazy ideas.

Rebecca: I think the interesting piece of this is if it’s used for learning, and that’s the point… the point is not making a good piece of music. It’s helping students [LAUGHTER] learn an idea and remember it, and maybe enjoy that learning experience. So the way music would be measured in those different scenarios would be different.

Oliver: Exactly. I don’t think that if you listen to the songs, I like them, but they’re not really trying to compete with human-generated music.

John: They’re not going to make it to the top 100 list in any genre, most likely.

Oliver: Unfortunately. [LAUGHTER

John: As you noted earlier, there is a long history of using music for teaching, these songs tend to stick with people for extended periods. You had mentioned the alphabet song, which is how many people learned the alphabet, and I know many people have said that when they’re trying to alphabetize lists, they often hear that song in their head as they’re doing so.

Oliver: I think that’s true for me too. Not always, but depending on my mental state. [LAUGHTER]

John: And I remember when my kids were very young, they learned other songs in school, including ones about the days of the week and the months in the year that helped them make some early connections that stuck with them. Songs have often been useful in creating long-term memories for many individuals.

Rebecca: Yeah, we hear a lot of those songs in my house, John. [LAUGHTER]

Oliver: I must admit that I sometimes wake up in the morning… I listen to these songs a lot, unfortunately, maybe… [LAUGHTER] and I have a song stuck in my head, which is good and bad at the same time. [LAUGHTER]

John: Two years ago, Rebecca’s daughter was learning how to play guitar, and I found some music to share with her. What were the songs that I shared the chords with you to give to her?

Rebecca: Oh, my favorite was “It’s a Small World.”

Oliver: Nice.

Rebecca: It’s painful.

John: One of those little earworms [LAUGHTER] that just sticks in your head if you’ve ever heard it.

Rebecca: Yeah. Oh, yeah. And then she thought that was really funny, very funny. [LAUGHTER] How have your students responded to the music?

Oliver: So I think the first response is surprise, people don’t see it coming, [LAUGHTER] and so it’s a combination of an onslaught of: I have to adjust to the idea “this is AI-generated,” and then “oh my God, on top of it, it’s physical chemistry.” [LAUGHTER] And then I would say that people started to respond very differently. Some started to speculate on what artists could be copied here. I honestly think it’s not like that. This is not how AI works. But sometimes maybe. But there’s also negative response, to be really honest. People, for example, I think a valid criticism is that people can kind of say this undermines the value of human music or makes fun of it. This is clearly not the intention, but people respond very differently. Overall, I would say positive, but there’s also some people who really don’t like it. Just to keep it real. [LAUGHTER]

Rebecca: Do you think it’s helped with their learning?

Oliver: It’s very hard to measure, right? Of course, I want to believe it. [LAUGHTER] I think of it mainly as some kind of tool to break up the monotony of a typical lecture with something that is a bit surprising and different. One could also turn it, and I will try to do this next semester, into like an exercise or homework, where you don’t use a large language model to generate the lyrics, but well, as long as the students play long, you try to ask them to write the lyrics themselves. And so you can also one-shot them. But I think that’s the least interesting thing.

John: And that is another thing I have done in my classes. I’ve allowed students to replace their lowest grade on an exam with a song or a video project or something else. And I’ve got a nice collection of songs that students had written over the years and produced as videos and shared with the class, and that again was another way of getting around the issue of copyright because the students owned the work that they were sharing. And that was a fun project. And students really appreciated hearing their colleagues do that. But I never required that. I always gave them multiple options. It was basically an unessay-type project.

Oliver: Yeah, I think one has to be careful. Both some people might not want to do it. Some students might be willing to do it but don’t want to share. So one has to be kind of considerate of this.

Rebecca: And having those options, having the song as an option as a way to learn or remember something, is a good way to implement UDL, universal design for learning, or different formats for learning.

Oliver: I agree, and there’s another dimension to it. I really think it’s very important to integrate artificial intelligence education into all kinds of college and maybe even high school education, and you can do this in a very technical, mathematical way, or you can do it in a more playful way with generative AI, maybe.

Rebecca: Yeah, I like that approach for sure. Have you had some other ideas about how you might use AI-generated music to assist learning?

Oliver: So a little bit before last summer, I developed… it’s not music, but I developed… what I call Thermo Tutor. So it’s a program where I generated 400 questions and answers and tips, and then I use generative AI… ChatGPT at the end of the day… through an API to grade and give feedback on the students’ answers. So you bring up a question that kind of matches roughly the outline of the course, so you can follow along, and it gives you a question. You write in the answer. The AI grades it. If you are doubtful of the AI, you can look at my certified answer if you want so. If you have a question about this, you can ask the AI the question. And so I think that’s a kind of interesting way of doing it because the problem is if I go to ChatGPT and I say, “Hey, I’m taking physical chemistry in college. Give me a question,” it will generate all kinds of questions, but they will probably not line up with the order of the class and the level and maybe the expectations of the instructor and so this is maybe another way to introduce it. Where I see this whole thing going is a little bit; these ideas will grow with what AI gives us. When you look at social media, you see, admittedly many very bad, but sometimes very funny, [LAUGHTER] AI-generated videos, and I think in the moment it’s a little tricky and requires some knowledge and a little bit of money… not fortunes… but maybe hundreds of dollars to generate this. But soon it will probably become much cheaper and less expensive. And so, what is the song today could be a music video in the near future, and maybe even an instructor can make small video games to kind of complement and add to the more traditional teaching approach.

Rebecca: So, do you have video games and videos in the works? [LAUGHTER]

Oliver: I try very hard to make good videos. I fail very miserably, [LAUGHTER] so I will try again in half a year. [LAUGHTER]

Rebecca: Yeah, you wait about six months, then you have a whole new set of tools, right?

Oliver: It’s true.

John: You mentioned having students create songs. Were you thinking of having them do it on their own, or were you thinking of them doing a similar process where they would use an AI tool to generate lyrics and then create a song from there?

Oliver: It’s an interesting question because it goes to this: How far can an instructor or professor really control whether the student is using AI or not? So my instinct on it is to leave it to the students. You can say, “I really encourage you to do it yourself, write the lyrics yourself, and play around with it, and learn some physical chemistry in the process. Spot some errors and mistakes, but if a student is less interested in doing this, it will be very hard to spot the LLM use anyway, right? So my personal opinion is that it’s kind of a very difficult path to fully enforce this, and maybe also a path that will look kind of outdated in a few years.

John: I was very impressed with the quality of the music that was provided. We’re recording this in early August. I just saw last night that there’s a new release of SUNO, so I imagine the quality is only getting better.

Oliver: I haven’t tried it myself, the latest release, but if there’s one, I really didn’t go to the web page back again. I was busy with other things, but yeah, this stuff will only get better. It will never get worse again. It will never disappear again, I think, unless it’s completely outdated by some or overwritten by some other technologies. Yeah, it’s kind of astonishing. There is a weird dimension to this, that things that we think of very creative at some lower level, seem to be automatable, where you can automate the process, and will not reach good levels or top levels. We don’t know where this will evolve. I don’t know whether you’ve ever tried to write little poems with ChatGPT. They fake it pretty well, I think. [LAUGHTER] I’m not saying that an English professor or an expert or really someone who’s deep into this wouldn’t be able to spot the difference. But it’s kind of shocking that so many people, including myself, can be tricked by it.

John: Musicians and poets learn from what they observed in the past, the techniques and the practices of others, and in a sense, this is automating that same sort of process, and it’s learning from its mistakes along the way, and as people do. I suspect it’s only going to get harder and harder to detect what’s created by AI and what’s created by humans.

Oliver: It’s fascinating how this absorption of information and styles is kind of similar in the net effect, but the processes are a little similar, but already quite different, and then you get very different outcomes out of it. And it boils down to the question: What is really human creativity, and how creative can these systems really be? But creative is not a zero-one situation. I think it’s some kind of gradual thing. You can be a little bit creative, and you can be amazingly creative. [LAUGHTER] You can be creative by combining things that already exist. This is very common in science, and it’s usually a very, very nice accomplishment. If you take a result from one field and you apply it to solve a problem in another field, that was a very routine way of being creative in science, and that is something large language models can and will be able to do. But then there’s another type of creativity. We really create something original, something qualitatively new. I believe, and that is something we will have to see whether AI can ever do this.

John: The best musicians, the best poets. It’s hard to imagine AI getting to that level.

Rebecca: Same with visual art and design. I’m just saying.

Oliver: I agree. [LAUGHTER] And the other dimension to it is: Are humans really interested in it? I think there would be no market for two super chess computers playing chess. Maybe there’s a niche market, but it definitely would get boring. But two mediocre chess players, watching them on YouTube playing against, that can be quite a bit more interesting. So there’s also that component to the story.

John: Because the AI systems are basically mimicking what people have already done, and the question is, can they ever get beyond that point by introducing a little bit more randomness and other components? It’s an interesting question, and I think we’ll be learning [LAUGHTER] the answer to that fairly soon at the rate things are going.

Oliver: I think so too.

Rebecca: In some ways, humans are already doing that because they’re bringing whatever their interesting question is and putting it into the system and challenging the system to start combining things that maybe wouldn’t have been combined otherwise too.

Oliver: And it might sound intimidating or maybe even threatening, but at the end of the day, it’s giving us an amazing tool. Maybe the tool is a little too good, [LAUGHTER] but it can also become a new way of being creative. Photography is maybe not a great example, but an example. I mean, it’s a very technical thing, and it somewhat competed with certain types of paintings, family portraits, and stuff like that. But it also allowed for new types of art and creativity to many.

Rebecca: And it didn’t get rid of portraiture and painting. It evolved both of those spaces.

Oliver: Absolutely, you know, of course, emphasis shifts.

Rebecca: Yeah.

Oliver: That’s I think progress and just history.

Rebecca: So I know that you’ve been doing some other interesting things with AI too. I think we heard something about a boot camp. Do you want to talk about that?

Oliver: So this is an idea that we want to allow students to more actively explore AI tools, especially in chemistry for data analysis, protein structure prediction, or what have you, drug discovery. Obviously, there’s a large range. Some people are extremely AI literate. Others have been very hesitant to even touch it, which is fine. But it is a tool that you probably will not be able to completely ignore, or only at a high price. So we thought together with the college of organizing an AI bootcamp for our department, and it’s for the whole college, but it is kind of focused on science and chemistry in particular. So it’s a 30-hour boot camp, and all of our incoming new graduate students, PhD students, are attending. But it’s open to everyone in the college, also undergrads and retired faculty, everyone who wants to come, existing graduate students. And so we have a lot of guest speakers from industry, from academic research, covering the copyright issues, ethical issues, mental health issues, but also some technical issues, without going too much into math or programming, and we try to do everything very hands-on. So there are lectures, but there are also always activities where students hopefully will be immediately putting to use what they’ve learned in the last lecture, and so this is something that I think Florida State University really wants to kind of develop in the next years. Introduce more of that in different forms into the teaching to position people better to do better research, position them to get better jobs, and we’re excited about doing it. Now, no one really knows how to really do it. It’s not like we have years of experience on how to integrate AI into teaching, and there’s a lot of institutional friction as it comes to grading and all kinds of concerns. So I think our society, we as professors or everyone, essentially teachers, students will have to figure out how to really do this effectively, and so this is one big step into this. I am sure that if we repeat the bootcamp next year, we will have lots of experience, and things will have changed, and maybe we’ll have completely new ideas about how to do it. But we have to experiment with it carefully to integrate it. I think ignoring it is, from my personal point of view not a very helpful thing for the students, and yeah, you could probably isolate yourself completely from it. But I don’t know whether that’s really a fair way of preparing students for the future.

Rebecca: Sounds like a really nice immersive launch into the new academic year.

Oliver: I’m looking forward to it. I’m excited.

John: How are your colleagues reacting to the use of AI in the classroom. I know here we have people with a fairly wide mix of views. Some people are really excited and enthused about it. Others want to resist it at all costs.

Oliver: Yeah, I think it’s the same at Florida State University. I think it depends a little bit on the department and college, and I completely understand that if you are a creative writing professor or in the English department, the reality on the ground looks very different from the reality in chemistry, where we are like, “Okay, we are teaching this material. Here’s a new way of teaching it. Let’s integrate it.” Probably the biggest thing at the root of it is evaluation. You can think of all kinds of clever and not so clever ways of integrating AI into education, but evaluation becomes a real problem. Where you could before give a long homework, a take-home homework, or ask people to write an essay over two weeks. Those things have gotten very difficult as evaluation tools. Same with college admission, I would say. The classical essay that people wrote, I don’t know how much value it still has. So these are things that we all have to kind of find ways to deal with it and adjust to it and make good use of it. It’s not like we are downgrading things… maybe in the short term… but I think I’m very optimistic about the future, that this will be an empowering tool for us to learn better, to study better, to teach better, to give people access in places where they maybe don’t have access to education. It’s a time where you have a 24/7 pretty smart, maybe not perfect, but pretty smart mentor and tutor at home. That’s like stuff from science fiction, and you might have it for free or for very little money.

Rebecca: I’ve been using it for running coaching. [LAUGHTER] You know, like when you need an expert, but you don’t have an expert, there’s ways that you can kind of integrate it into different ways of learning about new things.

John: So we always end by asking, “What’s next? With AI, that’s something we’re always asking pretty much every day.

Oliver: I think it’s probably always difficult to predict the future, in particular. It is even more difficult to predict the AI future and how it affects us, how it will technically develop, how it will be regulated or not. I think we all sense that is happening very fast, and this makes it different from earlier changes like the internet happened also fast, also led into bubbles, but it’s hear, right? And it’s our life nowadays, and we are communicating via the internet in the moment. And so it’s hard to predict. But I think in the short term, and that’s maybe one thing I’ve actually learned, when you start thinking about it, it’s very difficult to think about how will it look like next year, in five years, in 50 years. It’s very tempting to immediately smear everything together and paint out a very positive or very negative future. I think realistically, next year will look very similar to this year for most people, for most students, for most professors, and so we should just not sleep through it, and we should try to really integrate it carefully with a little bit of experimentation, but not too much. Listen to feedback from students. For my specific project, I really, like I said earlier, would love to integrate, see whether I can generate videos. A short teaching video could be very fun. I don’t want to change my whole lecture to songs and videos, but to have that as one component in the whole process, something that was impossible before. I can make an animated PowerPoint slide, but a four-minute video on my research would be, or on my teaching topic, would be difficult and very time-consuming, and so I think these things might be possible next year or the following year, and most of the other things remain to be seen.

Rebecca: It’s exciting to think about improvements or things that you want to integrate into your class that you’ve been thinking about for a long time, but just can’t take on because of the time or whatever, and that they’re only a fingertip away, perhaps.

Oliver: And it’s so playful. I mean, it’s not like I don’t have to read six books and spend a whole summer on trying to get into something new. You can open up the thing and say, “So what can I do with you?” Right? [LAUGHTER] It’s remarkable. The answers might be bad. [LAUGHTER]

John: Experimentation can be fun.

Rebecca: That always helps.

John: One nice thing about this is it provides more options for multimodal learning, and we know that when students see the same content in different modalities, it tends to reinforce learning, and it tends to make the learning a little bit deeper and to stick a little bit better. So it just gives students one more way of learning these things and making connections, and that’s pretty much always a good practice.

Oliver: I think so too. I’m really very traditional in my teaching. I write on the board. I don’t even use PowerPoint slides mostly, unless there’s something really I want to show, like a photo or a little animation or something like that. But I do know that while that is the traditional approach in math-heavy subjects, not every student necessarily responds well to it. Not because they are different; it’s just like a matter of taste. And these songs, I think, they have a good chance to stick also in general introductory general chemistry classes where we interact with the students maybe only for a semester or two, and then they go off and do something very different… really, kind of leave some positive, hopefully, impressions and memories, and maybe even some concepts that they remember.

Rebecca: Definitely. Well, thanks so much for sharing your experiments with us.

Oliver: Yeah, that was fun. We’ll see how it evolves. [LAUGHTER]

John: Yeah, thank you. This is a really fascinating experiment, and it’s nice to hear about it in more detail.

Oliver: Thanks again for the invitation. It was real fun.

[MUSIC]

John: If you’ve enjoyed this podcast, please subscribe and leave a review on Apple Podcasts or your favorite podcast service. To continue the conversation, join us on our Tea for Teaching Facebook page.

Rebecca: You can find show notes, transcripts and other materials on teaforteaching.com. Music by Michael Gary Brewer.

[MUSIC]

Show Notes

• Landy, Kathleen (2024). The Program-Level Conversations We Should be Having. Inside Higher Ed. February 28.
• Baule, Steven M. (2026). The AI-resistant classroom is a myth: Designing assessments that assume AI is present. eSchool News. March 12.
• FLX AI Hub at Finger Lakes Community College
• Dave Ghidiu (2025). Vibe Coding, Tea for Teaching podcast, Episode 409. September 3.
• Dave Ghidiu (2023). Google Apps and the LMS, Tea for Teaching podcast, Episode 278. March 1.
• The Immersive Lens video podcast – A podcast series hosted by Dave Ghidiu and Paul Engin from Finger Lakes Community College
• Gallant, T. B., & Rettinger, D. A. (2025). The Opposite of Cheating: Teaching for integrity in the age of AI (Vol. 4). University of Oklahoma Press.
• Tricia Bertram Gallant and David A. Rettinge (2025). The Opposite of Cheating. Tea for Teaching podcast. Episode 388. April 8.
• Jessmyn Neuhaus (2022). Faculty Book Clubs. Tea for Teaching podcast. Episode 269. December 28.
• Jessmyn Neuhaus (2025). Snafu Edu Tea for Teaching podcast. Episode 400. July 2. (Snafu Edu, along with The Opposite of Cheating, were the focus of our two Fall 2025 reading groups at SUNY-Oswego.)
• Deborah Ortloff

Transcript

John: When generative AI platforms first appeared on the scene, faculty had to address these challenges alone. In this episode, we discuss how program-level collaboration can help educators adapt more rapidly and effectively.

[MUSIC]

John: Thanks for joining us for Tea for Teaching, an informal discussion of innovative and effective practices in teaching and learning.

Rebecca: This podcast series is hosted by John Kane, an economist…

John: …and Rebecca Mushtare, a graphic designer…

Rebecca: …and features guests doing important research and advocacy work to make higher education more inclusive and supportive of all learners.

[MUSIC]

John: Our guest today is Kathleen Landy. She is the Provost and Vice President for Academic Affairs at Genesee Community College. Prior to this, Kathleen has served in leadership roles at teaching centers at Cornell University, Queensborough Community College, and Mercy College. She has taught in multiple modalities and also has extensive K-12 teaching experience. Kathleen also served on the POD Network Core Committee from 2022 through 2025, and currently serves as the Co-Chair of the SUNY Council on Assessment. Welcome Kathleen!

Kathleen: Thank you so much! I am delighted to be joining you today.

Rebecca: Today’s teas are… Kathleen, are you drinking tea with us today?

Kathleen: I certainly am. I’ve got a nice mug here from the Ithaca Coffee Company, and it is filled with Irish Breakfast tea because it’s March and it’s approaching St Patrick’s Day. So why not lead with that?

Rebecca: I think you’re on theme. You’re good.

Kathleen: I’m trying. Yeah, thank you.

Rebecca: How about you, John?

John: And I forgot that we’re approaching St Patrick’s Day, so I have English breakfast tea, which might not go as well with Ireland in the same way.

Rebecca: Immediate conflict, [LAUGHTER] there’s immediate conflict.

Kathleen: I didn’t expect this so soon in the podcast, [LAUGHTER] the acrimony. Oh my.

Rebecca:I have a blue sapphire again, John, but it’s in my new favorite mug that I got while I was in Switzerland.

John: Excellent. We’ve invited you here today to discuss your February 28th 2024 article in Inside Higher Ed that addressed institutional responses to generative AI platforms. We recently ran across the article and thought these arguments were just as appropriate today as they were two years ago when you wrote this. At that time, ChatGPT had only been available for a relatively short time, and you argued that institutional responses were generally reactive rather than proactive. Are higher ed institutions still mostly reacting to changes determined by AI providers?

Kathleen: I’d say yes and no. And I think a really important disclaimer here is that anything I say from this point on is from the perspective of a dedicated practitioner, and I’m not conducting empirical study and formal study of many institutions higher ed, right? So this is certainly based on conversations with peers and my experience and my day-to-day goings on. And so I think we are still a bit reactive and I think that also varies greatly by institution and within institutions, by school and by program. So I think it really depends. I think we’re all over the road, to be honest. That said, I think, initially, folks were dealing with this reaction, or this reactive response, in isolation and kind of left on their own to explore or to experiment, and I think that probably remains true, but perhaps more at a broader organizational scale. So I think individual departments and programs are experimenting and exploring, and so I think there’s stronger or weaker degrees of coordination of those efforts across institutions. So yes and no, I know that that’s a pretty broad answer, but I think it varies.

John: And I think that applies to most things involving AI in higher education these days.

Kathleen: That’s fair. Yes, I agree.

Rebecca: I feel like faculty really experienced a lot of transition and having to respond to a lot of different things in the last five years, which is a little tiring. This included things like the shift to remote instruction in 2020, the growth of for-profit companies like Chegg, whose business model basically encourages [LAUGHTER] academic integrity violations, and this proliferation of AI platforms beginning in late 2022, and also the upcoming deadline for digital accessibility requirements. So there’s lots of different kinds of things that faculty are having to shift and rethink around. And faculty in the SUNY system have also faced a transition in their learning management system on top of everything else during this time period. Is it really reasonable to place so much burden of addressing these systemic issues in the hands of faculty? I guess the bigger question is, what role do faculty play, and then what role do institutions and systems play?

Kathleen: It’s a great question. I think it’s a loaded question, and I’m all right with that. I think it’s a fair question. I think, no, it’s not reasonable to put all of that expectation on faculty. And I think the second part of that is, I hope it’s not the case that it is the expectation. So no circle in which I chat, or operate, or work to serve the mission of Genesee, or any place where I’ve been, it’s never been the assumption that faculty need to figure that out, they need to figure that out on their own. So, because it’s such a critical responsibility, because it’s such a mission-critical responsibility, because faculty play this critical role as stewards of teaching and learning at any institution, I think they deserve the support. So if I’m doing my job decently as a provost, I’m trying to coordinate that support. I’m trying to do what I can to encourage a couple of different things. So the day-to-day work that I do as a provost, hopefully, is mindful of three things, assessment, professional development, and, frankly, wellness. If faculty are the stewards of the teaching and learning experience on campus, I work in service of the faculty, so I’m constantly trying to think about the assessment of it all. What do we need to learn? What do the faculty need to be effective in their roles? What do the faculty need? What development do they need? What do they want to learn in order to be effective in their role? And then, of course, the wellness piece, and that’s trying to keep tabs on everything from workload to culture and departmental environments and institutional priorities and the extent to which we’re pursuing them with the spirit of shared governance and collaboration in all the things. So no, it’s not reasonable to put everything on the faculty, and if senior admin are doing our jobs correctly, we’re supporting them in that endeavor and trying to work with folks to figure out what we need to accomplish next and how we can collaboratively do that.

Rebecca: One of the things that you’re hinting at is a little bit of prioritization. When there’s a lot of things going on, everything can seem like it’s in its own bucket, and it’s equally as important, but it’s hard to put attention on everything equally. So it seems like there needs to be some sort of coordination or integration of these different kinds of initiatives: accessibility, AI, and the list goes on.

Kathleen: Yes, I think that’s where organizational structure and communication and collaboration between those and among those structures is so important. And so that gets into organizational health, and how effectively are we keeping each other informed about the things that we’re doing and the priorities that we’re setting? If we’re working in a siloed, isolating environment, it’s so counterproductive, of course, so I think the sense of institutional priority that everyone’s working with is paramount. Right behind it is the health of the working culture, the working environment. So I agree there’s an emphasis on priority there. That’s important.

John: Now you mentioned two terms there that I was just thinking about in your earlier comments. One is culture, and the other is silos. Is part of the issue, where we don’t see as much decisions made at the department or program level, the fact that we do have a culture in higher ed of faculty working in their own silos, working independently, and not agreeing on as many approaches to teaching and learning. From what I’ve seen here, most faculty are addressing the challenges of AI entirely on their own, without even discussing it that much at the departmental or the program level, in part because they’re used to developing their own courses without sharing a lot of information with others. The whole notion of academic freedom and independence is a bit of a culture in higher ed that may actually serve as a barrier to cooperative work on this topic.

Kathleen: I think that’s an incredibly sympathetic question. There is a culture of knowing things and being the authority, and I think it’s built into the fabric and the culture of higher ed in general, and because instructors, because faculty members, are masters of their discipline, they’re experts in the field, it is so wildly disconfirming to not understand the task at hand, which is, “how do I incorporate AI into my disciplinary knowledge, my instructional practice, my assessment practice, my collegial conversations with colleagues, my academic discourse with students, with colleagues, with the field?” I don’t want to get too technical, but it’s super weird. It’s hard sometimes to admit that we don’t know something. And I think, without generalizing, that’s really hard for academics. You don’t want to show that card and say, “I’m not sure how to handle this. I’m not sure what this is all about. I’m totally fearful of this thing.” In the case of faculty, it’s on multiple levels: “Is a chat bot going to do my job in five years?” It might be so disconfirming in terms of the value placed on discipline specific knowledge. “I used to be revered for my encyclopedic knowledge of x, and now it’s a 30 second prompt. So what does that mean for me?” So it’s weird. On an individual level, it can be threatening. On a professional, personal level, I would even say on a deeply personal level, in terms of how we define ourselves and ourselves of identity… I don’t mean to get too metaphysical or philosophical on this… but I think it’s really hard to make sense of this. And I think when things were first unfolding, and there were multiple platforms, ChatGPT comes immediately to mind, of course, but select platforms were rising in such prominence in 2022, 2023, folks were in isolation, which was, in fact, the impetus for this opinion piece to say, “I don’t think we have to figure this out on our own. This is a departmental endeavor. Let’s talk about the things, let’s model learning for our students and each other. Let’s figure it out together.” So yeah, I do think broadly the culture of academia can be a little isolating and individualistic, but in my lived experience, the culture of teaching is collaborative. And “hey, I’ve got a student with this challenge, or I’ve got this module or unit or facet of my course that’s a little clunky or not flowing well, or the engagement isn’t there. What do you think?” In my lived experience as a teacher, it’s always been a communicative, collaborative endeavor, where I try to get better and better and better, because I think the best teachers are lifelong learners. So there’s a humility in that that hopefully combats the resistance that’s built into the fabric of academia that sometimes takes the form of resistance to admitting that we don’t know something.

Rebecca: How do we start to move in the direction of this more proactive, scaled, systemic response to AI platforms? We’ve kind of danced around the idea that there’s these silos, not everyone’s collaborating, there is the opportunity to collaborate, teaching can be really collaborative. What needs to happen to kind of nudge it to this more proactive direction so that we can have a more systemic cultural change?

Kathleen: Well, I think the first part of your question, if I heard it correctly, is, what’s prompting this evolution, if you will, or this progress? Frankly, I think it’s because the challenge or the problem isn’t going away, it’s here. It isn’t going away, and so we kind of can’t ignore this anymore as a sector, so as more and more institutions and individuals notice this and are kind of getting our bearings in the wake of this seismic disruption, I see lots more activity within and across institutions. And I think the second part of that, if I heard it correctly, is, how do we build upon that? Is that accurate?

Rebecca: Yeah or how do we launch it forward a little faster, [LAUGHTER] or really leverage the opportunity to address it in a more systemic way?

Kathleen: I think part of it involves our metacognitive approach to this challenge and our metacognitive recognition that there’s different facets to our institutional and individual responses. So I definitely see some curricular responses, and I think this varies greatly by department. For some disciplines, it makes so much more sense to have AI-specific program level learning outcomes. It just made sense, because AI is impacting their fields, their programs in such obvious ways. That’s not true for every discipline, obviously. So there is reluctance to touch the curriculum or to revise or modify or even review the curriculum in other disciplines, and that’s okay. In other pockets of higher ed, it’s more structural. I’m seeing so many more structures, organizational structures, that are responding to this moment, to these changes. And sometimes it might take the form of an ad hoc committee, it might take the form of a Senate subcommittee, it might take a form of a… you name it. So I think there’s structural responses. Folks are starting to self organize, and I think that’s good. And I think if we’re mindful of that, and in turn, try to be planful about that and thoughtful about the structures we create in our own institutions, that’s another facet. Yet another facet involves some of the policy-related work that’s going on. Lots of institutions are drafting policies and doing so, quite rightly, collaboratively, trying to incorporate faculty perspective and working through channels of shared governance. Those are good things, I’d say, finally, another facet is the cultural piece. I see folks sharing things so much more often. Just this morning, a colleague shared an article with me from e-campus News, and it’s entitled, “The AI-resistant classroom is a myth: Designing assessments that assume AI is present,” and while I don’t know that that’s something I would offer like “this is the blueprint,” it’s thought provoking, and it’s beautiful fodder for conversation with fellow administrators and faculty and the assessment committee with whom I work closely. There’s so much out there, and I feel like folks are sharing so much more. So, I think a notable shift, and something that we can intentionally encourage, is the shift in the tenor of the conversations that we’re having with each other and that we’re prompting with colleagues, that it’s shifting so much more away from, “Oh my goodness, how we’re going to handle this? This is so worrisome, and this is something to fear” to “These are opportunities and what can we achieve? What do we need to achieve what do we want?” So, that might have been a long-winded answer, but I think it’s so multifaceted. I think taking stock of how we’re responding to this moment is an important part of continuing to respond to this moment in a productive way.

John: I may have overstated the degree to which faculty are working in silos earlier, because I know on our campus, we do have a lot of those discussions, but the discussions tend to be among those who are exploring AI as one group, and then we also have a lot of discussions among those people who just don’t want to have anything to do with AI as a different group, and they don’t always talk to each other as much as might be productive. Even within departments, I know of one department where one of our major AI advocates on campus also contains someone who is one of the most opposed to any student use of AI. And so that presents a little bit of a challenge in building some of these programmatic approaches. Do you have any suggestions on how we can bring these people together in a productive way… to at least come up with some shared principles or objectives?

Kathleen: I do, yeah. I think recognizing that there’s, as you just described, there’s such a range of responses to this. Some are logical, some are emotional, some are deeply discipline driven by discipline-specific thinking. But there is variability, and we know that. So I think the ways to bring people together… this is not an earth shattering statement, but… need to be really thoughtful. It can’t be one and done. It’s not an hour with cookies, although I do like a good cookie, let’s not have that come across the wrong way. But I think it’s frequency of meetings or just opportunities for discussion. I think it’s the structure. I think it’s the clear communication about the purpose of getting folks together. There’s a few different pieces to that. If there’s a group of folks who were super recalcitrant, they’re super resistant, I think there’s a number of different strategies to employ there, and one is modeling interesting or effective use of some of these platforms. I think it’s creating low-stakes experiences with platforms to facilitate greater exposure and to generate intellectual curiosity, to take the sting out of that isolating expectation that folks experiment on their own. I think communicating the intent, as I mentioned before, and I think most importantly, grounding these conversations and these experiences in the mission of your institution. There may be such variability among folks on a campus with regard to their perceived value of generative AI, but what they all have in common should be the mission. And so if I can frame conversations about, “Hey, what does this platform or this use or this potential use for AI have to do with your program? Have to do with the students’ experience?” That’s, I think, what connects us all, and that’s what I think should be the impetus for these conversations. What is best for the students, what is best for our viability as a program, to ensure that we’re still relevant, that we’re still accomplishing for students what we say we will. And I think that’s a good starting point. I think framing those experiences where you’re hoping to achieve consensus, I think framing them is really important. It’s a good start.

John: One of the things we did last year was we had 24 faculty participate in a SUNY IITG-funded initiative where they spend a little over a semester developing some activities that they could embed in the classes and experiment with, but the first day of that, we had people together all day, and it was really one of the most productive sessions I’ve seen, because at the start we saw this wide polarization, but then the people who were really opposed to AI got to see how it might be productive for them, and they got to see some examples from others of how it might be helpful for students who might be left behind without some additional support. And at the other extreme, the people who were really enthusiastic got to hear some of the concerns of the other people and see how it could be used in ways that are detrimental. And I think everyone left there with a much more balanced perspective at the end of that day. And then we met regularly throughout the semester, and it was really nice to see people coming together there. The challenge is getting them all in the room in the first place.

Kathleen: So that’s a great example and thank you for sharing that, and it prompts me to share this example: For that list that I just offered, in terms of trying to demonstrate effective use of select platforms, and trying to give concrete examples of how this might be beneficial, to try to model that. And so an example that might be worth sharing is that we are up for our middle states reaccreditation. In fact, our site visit is in a couple of weeks, and we’re trying to get the word out, “Read the study! Read the study!” We want folks to engage. It’s a really thoughtful self study. We’re proud of who we are. We’re mission centered. We do good work. We know what we’re working on. We’re humble, we hustle, all the things. However, I know that it’s a tough sell to get an entire campus community to read a 100-page document that is evidence rich, and perhaps to some, not the most dynamic document. It might not be the page turner we hope it is or that we think it is. However, I leveraged a platform to make summary podcasts for each chapter, and I shared these along with the self study, and we shared them broadly with the campus community, and it is truly trying to model this in the spirit of UDL. Here’s multiple means of representation. You can read this and then you can listen to this, and it’s intriguing and it’s entertaining and it’s good and it’s a decent use of a handy application. And I’m hoping that folks will listen to that podcast and think, “Oh, wow! Maybe I could do that in my sociology class. Maybe I can encourage students to create some summaries of some reading assignments, not to totally supplant them, but to supplement and to reinforce learning.” And so that’s one example that your example has prompted me to share. Yeah, I think there’s so much that we can do. Your example also speaks to the idea that it shouldn’t be one and done. It’s not an hour. We need enough time to engage with each other, with these platforms, to make sense of something before we can even speculate on how to best incorporate it into our curriculum, or not.

Rebecca: Part of what I’m hearing both of you say is, it’s important to have someone who’s watching the bigger conversation on campus and can call these meetings to happen. To use the teaching center to have a whole-day workshop, or have administration say, “Hey, we’re going to dedicate this day to having these conversations. We invite you to all come.” Someone has to take that first step to: A) recognize that the different populations exist, and then B) intentionally bring them together.

Kathleen: And I think bring them together in different ways and different modalities, dare say. I mentioned sharing resources before. At GCC, we have a week of professional development days in advance of each semester. And so I knew just by being, you know, a good SUNY citizen, and by being connected with colleagues and peers, I knew that SUNY Finger Lakes Community College has been doing some fabulous work with their AI hub, The FLX AI Hub. And so we welcomed folks from Finger Lakes Community College to our campus to help us and to give some really thoughtful overviews of some ways to integrate AI into the teaching and learning endeavor, or at least to consider what those integrations might look like. And one of the follow-up activities that full disclosure, I have not yet sent invites for, but I bought the books, and that’s something. So the wonderful Dave Ghidiu from Finger Lakes Community College had recommended to us a text that he loves, The Opposite of Cheating by Tricia Galant and David Rettinger. And frankly, I want to have a provost book club. I want to get folks together with no expectation that anyone show up with the answer to all the world’s problems. I want us to come together, read something thoughtful, listen to each other, and engage in that dialog that you described, John, that some folks, they have really valid concerns that everyone needs to hear and we all need to respond to. And others have done trailblazing, innovative work that everyone should hear and we should all try to emulate and so there’s such a discourse, which I think accomplishes a few things. It helps everyone improve our instructional and assessment practice. And I know this might be cheesy, but I’m a cheesy person, it stokes that fire of intellectual curiosity that drives us. We went into these fields, into this endeavor, into academia, because of that curiosity. What a moment to be in academia. Of course, it’s daunting. Of course, it sometimes feels laborious, perhaps, but we’re not alone, and we get to figure things out together and I think that really is a gift.

John: And I want to include a couple of shout outs, one to Dave Ghidiu, who has been on the podcast in the past, and also to the authors of The Opposite of Cheating. We’ll include links to the podcast with both Dave and with the authors of The Opposite of Cheating in the show notes for this episode. And also, we did a reading group on our campus last semester, and we’ve even had a member of your campus participate in the reading group, a member of your math department who was a really active and enthusiastic participant in that reading group in the fall semester of this year. We’ve been running generally one or two reading groups a semester for faculty for the last eight or nine years now, and they’ve been really productive. And we’ve done them in conjunction with other campuses. For about six years, we were doing them with SUNY Plattsburgh when Jessamyn Neuhaus was there. That broke down after she left and moved to Syracuse, although we did read her book after her move to Syracuse. They’ve been really productive ways of bringing people together and sharing their teaching practices and building this sort of community. We just wish we had more people in them.

Kathleen: It’s how things grow, and I think it’s so multifaceted. I mentioned our accreditation process. I’ve partnered with Deborah Ortloff, also at Finger Lakes Community College, to really explore and learn from her ways in which generative AI can help in the accreditation process. There’s so much that we can benefit from, and it’s really eye opening, and I think that initial exposure is really important, and then that low-stakes period of experimentation and just making sense of things with people and not doing that in isolation, great things can happen. And I’m not trying to sound like Pollyanna, I’m not trying to sound like there aren’t some daunting pieces to this, but we have great colleagues to figure things out with. That’s a joy. So why not embrace that?

John: I wish we had access to NotebookLM when we last did our accreditation, so that we could generate those podcasts, because when we did ours a few years earlier, I had to do some podcast recordings with some of the people leading that and then edit them all and post them to provide the same multimodal experience. But if those documents could have just been uploaded and generated and shared, it would have saved me many, many hours of work. But again, having those audio versions is really good, because we know multimodal presentation helps with learning for faculty as well as for students.

Kathleen: That’s right and dare say, it’s fun. It’s intriguing. I was delighted to hear those because I was just in awe that, within minutes, this was generated, and it was entertaining. I have to say it was great.

Rebecca: One of the things that we’ve hinted at but haven’t directly talked about is perhaps the need for policies to be developed related to AI. Some of this might happen at the institutional level, and some maybe at the programmatic level, or course level. Can you talk a little bit about how you see what kinds of policies should be developed at what level? What kinds of things should be in, like an institutional policy, and what might be more at a programmatic level?

Kathleen: So at the institution level, and I hope this doesn’t sound in any way like evasive, or, dare say, even cheeky, but I think a really good institutional policy is to have a course level policy. I think there’s a danger, as we all know, of the potential for institution- level policies to become overly prescriptive and curtail academic freedom. That’s to be avoided, obviously, but I think having clear expectations that departments, programs, individual courses, establish and communicate clear expectations for the students in those courses and programs. I think that’s essential, because not having a policy renders everybody guessing, and then the rules aren’t clear, and then the expectations aren’t clear, and most importantly, the learning isn’t clear, because part of the, and this is a segue to the course level consideration, for course-level considerations, or even assignment level for that matter. Generally, folks will fall into one of three policies, either we allow AI, we allow selective or limited use of AI, or we don’t allow it. Each of those is fine. I think it’s reasonable for a department to get together and discuss, “what is our department-level policy?” From my perspective, it’s perfectly reasonable to say every instructor of every course within this program needs to have a clear policy and communicate the rationale for that policy to students, because if students understand the rationale behind a policy, if they understand the policy, they’re more likely to follow it and to adhere to that policy and to be guided by that policy. So, I think at the institution level, expectations that are specific yet flexible are prudent, with the caveat being that they shouldn’t be overly prescriptive. At the department level, I think clear expectations that provide adequate guidance to faculty with regard to what they need to include in their syllabus or their course descriptions or their assignment directions, it has to be adequately specific yet flexible. Same principle there.

John: One of the challenges that students face is that they may end up with very different requirements, and at least putting it in the syllabus and describing and having a conversation with students is a start. But you also mentioned that, perhaps, whenever faculty give assignments, they should list those things, because students don’t always go back to the syllabus, and it could be confusing when you’re faced with five different policies in five different courses, or even in a course… AI may be allowed for some types of assignments, but not others. And that’s a lot for students to keep track of unless it’s being made explicit at each stage. So should faculty also include whether AI is used, not allowed, or is used in a specific manner, in every assignment where there may be a potential for use?

Kathleen: I don’t see a downside to that. So a lot of things go back to the mission and what’s best for students. If learning is the goal, then everything we can do to facilitate that learning is worth pursuing. So, with sympathy, if a student is taking five different courses with five different course-level AI policies, why wouldn’t we communicate this clearly and frequently to students? I think there are things that departments, or dare say, even institutions, can do. I don’t think we’re quite there yet, but I think there are things that we can do, such as having a template, let’s say, for course-level policies or sample statements, so that there’s a consistency with what students encounter, even if there’s variability in the degree of use that’s permitted across courses. Hopefully, if things look similar, if things are consistent in how they’re communicated and how frequently, I think that helps. I also think it is up to individual instructors to determine what’s best for their course, and I do think it’s a skill set that students need to develop, to literally or metaphorically read the room and to adapt, because that goes for us too, as professionals, as teachers, we have to learn when it’s appropriate to use or not to use or to use in a limited way, AI. So there’s also the metacognitive skill of helping students practice paying attention to the rules as they navigate their programmatic experience, so it’s layered and it’s nuanced. But to answer your question, I think there’s potential for things to get pretty confusing for students who are taking multiple courses with varying policies on AI, but we have to help them keep track of that and reminders help, presenting expectations in multiple places helps. It’s not enough to say it once in class and expect that everyone retain that forever. We have to document it. We have to share it. We have to communicate it multiple ways.

Rebecca: Your description of templates is making me think about Creative Commons licenses and the human readable versions of those licenses, where you can kind of mix and match permissions [LAUGHTER] or levels of what’s allowed or not allowed, and I could imagine some AI policies following a similar format. “You can use AI as long as you provide documentation that you used it,” or like whatever the rule might be, but I could imagine something like that being really useful for students to keep it at that human readable clear way.

Kathleen: And I would never use the word “Mad Lib”, but it’s kind of, you know, “for this course, it is my expectation that generative AI is used to: fill in the blank.And this is why.” I mean, we might want a more sophisticated term than “Mad Lib”, but template works, I think, if students know where to find this information. We’re actually doing this now at Genesee. So I meet routinely with one of our student government association officers, and he and I are collaboratively working on a draft of guidelines for conversations about breaches in academic integrity that involve AI, how to handle it. And so our hope is to, hopefully in the next three weeks,… well let’s say four weeks, hopefully in April… to gather a group of faculty and students so that we can learn from each other and refine these guidelines and socialize… I don’t love that verb… but socialize these guidelines. We’ve been engaging with folks from our Academic Senate Committee on academic integrity, our standards committee, and so we’re collaboratively trying to create a resource that will help folks engage in conversations. What to do if you suspect a student has used AI when they shouldn’t have? What should students ask at the beginning of a class of a faculty member if they’re not sure what the rules are? And so having this resource and the associated conversations that will hopefully be empowering for both faculty and students, that’s part of the cultural piece, it’s part of the resources piece, it’s part of the sharing piece. It’s all these facets of how we can make the student experience more enriching and more in line with what we’re hoping to accomplish. It was a bit of a tangent, but I think the idea of bringing more folks in to consider how to mitigate the confusion that students might experience. That’s a good thing, including the students, especially the students.

Rebecca: One of the biggest challenges associated with generative AI use in the classroom, or with student use of it, is the reliability of online assessments. And really, this is for courses, really in any modality, because we all use online assessments for things. What approaches should faculty and institutions take to ensure that faculty and students are able to reliably assess the progress of student learning?

Kathleen: It’s an excellent question, and it’s one that has prompted so much conversation on my own campus, I know on other campuses. I have a few responses to this. One, what comes to mind immediately is the prominence of tech responses to this challenge. So whether it’s proctoring and anti-cheating tech platforms, I think they have their place. It depends on the day. I have some mixed opinions on the ultimate efficacy and effectiveness of one upping, I think. Sometimes I’ve heard the term “arms race” used, and I don’t love that for a couple of reasons, but I can see how if there’s platform A, platform A-Prime comes out three months later, and then it’s a race sometimes. So I think there could probably be, and likely is its own podcast to address some of the tech platforms and the responses to that, but if I may focus on what also comes to mind… For me, surprisingly, I think there’s a personal approach to mitigating the likelihood of cheating. That comes down to relationships and communication. So I think those are at the heart of my philosophy, philosophy of teaching, philosophy of assessment. Whether there’s a tech intervention or a tech solution or no tech solution. I want students to appreciate why they shouldn’t cheat in my class, and if I’ve done my job well, if I’ve communicated that well, I want students to second guess that decision. I want students to be less willing to cheat because I have already done a good job of communicating why this learning is essential and why they shouldn’t circumvent the learning by cutting a corner or by using an AI-generated form of assistance when they shouldn’t. The other piece of that is the relationship. I want my students, and sometimes, I grant you, this is really tough in a larger lecture, I get it. So class size is a factor here, but I want my students to feel the faith that I have in them. I want them to not want to cheat [LAUGHTER] in my class, and I’m not trying, once again, to sound like Pollyanna or that I’m the eternal optimist here, but I know students cheat. I know there will always be a few who try to cheat, but I want my relationship with the class to be an impediment to that instinct. [LAUGHTER] I want folks to second guess like, “Wow, Kathleen has such faith in me and I don’t want to compromise that.” I want the relationship to be its own disincentive. And I want the communication of, “Hey, this is why I’m asking you not to cheat on this.” Can you cheat? Probably. Sure. But should you? No, of course not, and here’s why. So I think there’s a surprisingly personal part of my own teaching philosophy that informs this part of that answer. I also think there’s a real place for authentic assessment, and of course, modality may play a factor here, but to the extent to which learning is applied, authentic, that can mitigate the likelihood that students would outsource it to an AI platform. The same goes for metacognition and the assessment thereof. If I have an assessment or an assignment that takes a look at process over product, or if I have something that takes a look at students’ own metacognitive reflections on how they did, it calls for more self assessment, calls for more personal reflection. I think that also mitigates. It could be a whole ‘nother podcast to get into. I’m not going to outsource to a platform an articulation of how I feel about how [LAUGHTER] the process went. So I think there’s the communication piece, the assignment design piece, and the potential for a tech solution as well. I think those three things are relevant and connected to my pretty long answer to that question. I think it’s multifaceted.

John: And those are all great solutions. One of the things, though, that many people are increasingly concerned with is the growing availability of agentic platforms that can log into a course management system, check to see what work is due, complete that work and log out without any intervention other than the student giving it the URL, the login ID, and the password. And at the other extreme, there’s also concern that faculty could do the same thing. That these same platforms could allow faculty to set up an agent that would go in, check within the learning management system to see if there’s work to be graded, grade that work, provide notes in the feedback column in addition to a letter grade, and then log out again without any intervention from faculty. And we’re starting to see a little bit of that happening, and that’s an area that the learning management platforms, as well as the AI platforms, should be working on a bit more actively. In the meanwhile, we’ve got a lot of concerns coming up and having discussions with students, working with colleagues, and coming up with the best institutional responses, but sometimes it’s going to have to go a bit further than that, I think, to resolve some of these new challenges, because both sides of this are racing. and I work with AI quite a bit, and every time I go on social media, I get ads for these things that can do all the work for me in my courses, where I just give them some money each month, and it will do everything and guarantee that I got a high grade with an extremely low chance of being caught doing that. So students are faced with a lot of temptation, and to the extent to which we can build relationships, that will help.

Kathleen: Yes, and I think to your point, our committing to staying on top of how things are escalating and how platforms are becoming increasingly sophisticated, or, dare say, elusive when it comes to detection, etc., I think our commitment to staying on top of that is paramount. We have to know what we’re dealing with here in order to even have a hope of designing assignments and learning experiences that can circumvent that. So it’s daunting. I think I would find me tedious if I came on here [LAUGHTER] and I thought I had all the answers. I do not, but I do know that groups of folks getting together who are committed and passionate and knowledgeable and curious, I think us figuring it out together, we have a far greater likelihood of achieving what we’re hoping to, but it is daunting, it is. I would never suggest otherwise. For me, the relationship piece is huge, and the way I can structure reminders about what’s expected, whether it’s reiterating PLOs up to and including ethics. There’s a lot of talk about ethics and our shared responsibility for instilling a deep sense of professional ethics, research, methodological ethics, personal ethics, personal integrity. Where does that live in our curriculum? Which goes back to the original thought here that we have to take a look at the ways in which AI is impacting our fields, our teaching, our assessment, and do what we can to be proactive and review, refine, revise our curricula, our assessment practice, in a manner that will empower students, because if we don’t, they are vulnerable to what we didn’t teach them, what they didn’t get out of their learning experience during their time with us. And so it’s mission critical. There’s nothing more important than saying, “Okay, I need to know, as a teacher in field X, I need to know how AI is impacting my role as a teacher, but also my field, my discipline,” and if I can’t communicate that to students, if I’m not helping them navigate that with me, that’s awkward. That’s a failed promise, and nobody wants that. I have such great faith in my colleagues. I’ve never worked with a more committed faculty, a more committed campus community. It’s beautiful. And so this is mission critical. We got to think about how we can best serve students by responding to this moment in a multifaceted, planful way.

Rebecca: One of the ways to respond that we haven’t directly addressed, but might be worth mentioning in this context is we can get insulated in departments, get insulated within our institution. But there’s a real need for the higher ed sector to work together to promote state/federal regulations around some of these things too. We talked about incentive for students to leverage [LAUGHTER] these tools to skate by, but there’s also a real incentive for these companies to keep finding ways to circumvent [LAUGHTER] any of the security in place. So I think we don’t talk about that enough, our role as educators and the ways that we can participate in moving potential regulations forward. I think that what I’m suggesting is like, we don’t really talk about that much, but it seems as though there’s a role we could play.

Kathleen: I think so. And I know I’m using this word a lot in this conversation, but I think it’s a multifaceted approach. So one of the appeals I would make is that this is an equity issue. Increasingly, some will have great access and exposure to AI platforms, and will benefit from that, and will learn how to use these platforms ethically in a manner that serves themselves and society well. Others will be boxed out of that for any number of reasons, up to and including socioeconomic status. So if we can have value laden conversations, in the best sense of the term, we can keep this personal in a weird way if we cannot lose sight of the human impact of technological advancement here, I think that’s the name of the game when it comes to advocacy. And then we all just work at scale, whether it’s organizational or societal. We influence the circles in which we operate to generate smart conversations about, “Hey, what do students need to learn? What do students need to be empowered to be able to do or not do with or without AI?” [LAUGHTER] For me, the advocacy piece takes the form of what can we do to stave off vulnerability, either at the individual or organizational level, and for the individual level, it’s often on behalf of the students. How can we advocate on behalf of students? So, I don’t know if that fully answers your question, but I see a lot of things structurally, and I see different scales and different facets, different points of entry, if you will, for the conversation. And so equity would be one of my points of entry for the advocacy piece.

John: And as we’re all wondering, when we’re talking about AI, what’s next?

Kathleen: That’s a great question. I’m going to just pop that into ChatGPT, if you don’t mind, I’m just going to “what’s next.” [LAUGHTER] Oh, boy. Honestly, I think connecting with more and more folks and staying connected and having the conversations and taking a deep breath and figuring it out together. I think that’s what’s next. I think it’s about incremental change and incremental responses and just huddling up and doing what we can to honor the mission of the institutions we serve. That’s what’s next, I think.

John: It sounds like a great approach, and I think we need to do a lot more of that. Thank you for joining us. And we’re looking forward to that next conversation.

Kathleen: That sounds lovely. This was so delightful. It was really a treat to talk with you both. Thank you.

Rebecca: And we’re looking forward to the next time to huddle up, for sure. [LAUGHTER]

Kathleen: Likewise, thank you.

[MUSIC]

John: If you’ve enjoyed this podcast, please subscribe and leave a review on Apple Podcasts or your favorite podcast service. To continue the conversation, join us on our Tea for Teaching Facebook page.

Rebecca: You can find show notes, transcripts and other materials on teaforteaching.com. Music by Michael Gary Brewer.

John: Editing assistance provided by Madison Lee.

[MUSIC]