AI, Innovation, and Teaching in MSU Meteorology with Andrew Mercer

Dr. Andrew Mercer is a meteorologist and climatologist at Mississippi State University, home to one in three of today's on-air broadcast meteorologists. As Graduate Coordinator for the Department of Geosciences and meteorology team lead, his research applies machine learning and AI to large scale atmospheric problems, from severe weather forecasting to hurricane prediction.

In this conversation with host Dr. Shannon Harmon, Andrew reflects on what it means to teach in a nationally recognized program built around both the science and the communication of weather. He shares how MSU prepares students for careers across the country, how uncertainty is woven into meteorology as a discipline, and how he is thinking about the role AI might play in the future of the field and the classroom.

For faculty wondering how to approach AI in their own classrooms, Andrew's advice is simple: treat it as a tool, meet your students where they are, and keep the focus on critical thinking.

Transcript

SPEAKER_01 0:00

I'll start by saying AI is scary. I know. I I you hear it on the news, you hear it on other podcasts and stuff all the time that AI is this looming threat. AI has to really be thought of as a tool that can be used to enhance your productivity or enhance the classroom experience or whatever it might be, as opposed to something that's coming to replace you. We're always going to need domain experts, my field specifically. There's never going to be a day where a computer is going to issue a tornado warning. It's going to be a person that has to issue that tornado warning and be accountable to that. And so that's true not just in my field, obviously, but across all domains. There's always going to be a human component. And so rather than look at AI as this looming threat, we need to think about how do we leverage it to be better at what we do.

Shannon Harmon 0:50

Welcome to Insightful Teaching, the podcast from the Center for Innovation and Teaching Excellence at Mississippi State University. I'm your host, Dr. Shannon Horman, and I'm so glad you're here. Whether you're listening on your drive to campus, between meetings, or maybe while sipping your coffee before class, welcome. Today's episode brings together two things that are shaping the future of higher education: discipline-specific excellence and artificial intelligence. So if you've been wondering what AI actually looks like in practice, not just in theory, this episode is for you. Mississippi State is home to one of the nation's leading meteorology programs, recognized for both its research and impact. And today's guest is right at the center of that work. Joining me is Dr. Andrew Mercer, a meteorologist and climatologist whose research focuses on applying machine learning and AI to complex atmospheric problems from severe weather forecasting to climate modeling. In addition to his research, he teaches a range of undergraduate and graduate courses and serves as graduate coordinator and meteorology team lead. So today we'll explore how a field like meteorology is leveraging AI in powerful ways and what that means for teaching, learning, and preparing students for the future. Andrew, welcome to Insightful Teaching. To get us started, please tell our listeners a little bit about yourself, including your role here at MSU and the kinds of courses you teach.

SPEAKER_01 2:25

Well, I've always had an interest in science and weather specifically. And I I grew up in the Texas Panhandle where we had lots of severe weather and tornadoes and hail and all the exciting things. Um, probably had a weather interest since I was very young. Uh, lived in Utah for a little while, got to deal with lake effect snow and big snowfall, and which was also a different kind of weather that I wasn't used to, and then moved to Louisiana and spent some time there before I went to college at Oklahoma. I did all three of my degrees at the University of Oklahoma, which is one of the big meteorology programs in the country, studying atmospheric processes. I kind of went into it expecting to be a private sector working behind-the-scenes meteorologist. And life, of course, sometimes takes you in different directions. And as I was working, I got really interested in research and specifically research and how it links into AI. So I was actually doing AI problems back in the early 2000s with some different um projects that I was involved with. And, you know, it's just kind of morphed from that into joining the faculty here at MSU in 2009 as an assistant professor. And then I've been here for 17 years and kind of worked my way up the ranks through the meteorology program. And now I'm um the leader for meteorology as well as a graduate coordinator.

Shannon Harmon 3:41

Awesome. One thing that I think as an alum of Mississippi State that is such a big point of pride for us is our nationally recognized meteorology program. Why do you think our program stands out?

SPEAKER_01 3:55

It's a very unique program. I'll start by saying meteorology in general is a very niche. Uh, there's only about 70 or 80 programs in the whole nation that do meteorology. And so having one at Mississippi is already kind of an advantage. And in addition to that, we are a very unique program because we're so focused on broadcast meteorology. We've been known for doing broadcast meteorology since the 90s. And since, you know, early to mid-2000s, we've developed more of a full-blown professional meteorology track. And now we're we kind of cater to a unique type of student that's interested in not just the um being a researcher or being a forecaster, but also people that want to do TV weather or people that maybe want to do all of it. They want to see what other options are out there. And so most of the meteorology programs in the country don't cater as much to broadcast. They look more at just the professional side or being a researcher, working as a scientist. Our component of broadcast meteorology, which is we're known for nationally, is the thing that definitely sets our program apart. We have between a third and 40% of all TV meteorologists have either taken courses from us or have a degree from us nationwide. So we're we're known all over the country for our meteorology program.

Shannon Harmon 5:09

Wow. And that speaks volumes about the types of experiences that students in that program get here at Mississippi State. They get that in front of the camera, not just learning about the data and like charting the graphs behind the scenes. They get both here.

SPEAKER_01 5:23

Right. Yeah, we actually have a TV studio in the building with us, and we have an instructor whose entire job is just teaching broadcast. We do four semesters of broadcast practicum where they're in front of the green screen and doing severe weather coverage or doing making a forecast, working with the tools that we use in industry, learning how to put together the whole, it's the whole production that uh broadcaster has to learn while also being in the classroom and learning the dynamics and the physics and all the heavy duty math and science that's involved in a meteorology curriculum.

Shannon Harmon 5:53

No wonder we're nationally known.

SPEAKER_01 5:54

Yeah.

Shannon Harmon 5:55

We are really giving um preparing them for that to be a well-rounded uh person in their field. I think that is so important and really lets us stand out. Uh so we've talked a little bit about like in front of the camera. Talk more about um what happens behind the scenes as far as weather and data sets and what that looks like.

SPEAKER_01 6:13

Weather is a data rich, but also a data poor science, which is kind of an interesting thing. Um, it's data rich in that we in the US, United States, we have the most, the highest density of observations anywhere in the world. Um we have weather stations everywhere. We have weather balloons that we launch twice a day that collect data, and then we generate weather model data that are massive, like terabytes of size, um, every single day. And we use all that to make our forecast decisions. All that to say, we still we launch weather balloons twice a day. We launch them in maybe 75 to 80 places around the country, and that's how we measure the vertical atmosphere. So we have a data problem while also having a lot of data. And so weather scientists, especially people that go on to work in graduate programs or do anything outside of just pure forecasting, they're going to be needing to know how to work with data and manipulate data and understand data structure and how data is collected and its limitations. There, so there's a massive data component of being a meteorologist.

Shannon Harmon 7:16

That is awesome. So what we provide for our students is really bridging that in front of the camera, that practice to then theory, to then research. Yes. That is key to what we do. And as a result, uh, what was the fact, Andrew? You probably know it more. I read it, it was on our uh Mississippi State uh social media. It's like we produce what's the number of national forecasters or broadcast journalists in the United States.

SPEAKER_01 7:42

Like for meteorology, we produce the top number. We're number one.

Shannon Harmon 7:47

Yes, we are number one. Wow. That's such a testament to what your program's doing. So thank you for that. Um, so we were just talking about data, which makes me think about your own research. Um, so it it sort of sits right at that intersection of meteorology and artificial intelligence.

SPEAKER_03 8:04

Right.

Shannon Harmon 8:04

Um, for listeners who may not be familiar with the field, uh, how is AI currently being used and how has it been used? Because you said you've been doing AI for a really long time.

SPEAKER_01 8:14

Right. Yeah. I I was actually in a working group in 2004 when I was at Oklahoma. And at that time, there was no meteorology AI link to speak of. Um, the the story I always like to tell is when I went to the annual meeting for meteorology, which is the American Meteorological Society annual meeting in 2007. I sat in the AMS AI working group meeting, and there were like 20 people in there. And last year, right, I guess it was the year before last, I went to the AMS meeting again, and it was two sessions of 200 plus in each going simultaneously. So AI has really taken off in our field, really, in the last five years, um, with the advent of large language models and also the uh just the availability of data and computing resources. My research specifically, I look at AI to solve um large-scale weather problems. And so I look at different things like classifying outbreak types. I've done my dissertation work was looking at differentiating between tornadic and non-tornadic severe weather outbreaks. And so we developed a probabilistic model that would actually predict this using AI back in 2009. So we were looking at it even back then. Um since then, I've done different kinds of problems. Some have looked at things like predicting hurricane rapid intensification in the Atlantic, which is really important because the major hurricanes are the ones that rapidly intensify. Yeah. And it's something that we're currently really not very good at forecasting. And so it's an opportunity to improve. I've also done a lot of pattern recognition work where we look at uh like climate scale phenomenon and how those change. I've actually got a product right now in the climate prediction center's tropical cyclone annual outlook. You know how you see like in April and May they produce uh we think there's going to be this number of hurricanes that year. Well, we actually have an AI-based tool that's in that workflow. So one of one of our um, one of my postdocs helped develop that. And now is that's currently part of the operational entity. And we're actually working with that research group right now to develop a sub-seasonal forecast where we're looking at it maybe on the order of a month or two in advance and at a specific basin, so not just the whole Atlantic. We're gonna look at the Gulf or the Caribbean or the South Atlantic versus the North Atlantic, those kinds of things.

Shannon Harmon 10:35

So that is awesome. Okay, so two things you said that I think are really important. I want to talk about them individually. Okay. So um we're gonna come back to this first one, but you just brought up like doc students and working with students. So you've been using AI and your research for a while, but you're also using it with students. So I want us to talk about the way that you not only use it for your own work, but then teach students to use it. So I want to talk about that. Okay. But I want to second or first, this is a second point, but I want to talk about this first. Is um okay, is it true that the tornadic alley has sort of shifted? You said you were in Tennessee, Oklahoma from a really good program there, that's where you learned. So that used to be sort of like, you know, n 80s and 90s, that's where all the tornadoes happened. But now in northeast Mississippi and this area, Tennessee, we've seen a lot more of tornadic activity. Um, so is that true or is that just I don't know what we think?

SPEAKER_01 11:24

It's it's complicated. Uh-huh. I'll just say that. I love that. Absolutely.

Shannon Harmon 11:29

That's why I'm not an expert in it, and you are.

SPEAKER_01 11:32

The the types and the the types of storms that produce tornadoes in Oklahoma versus in the southeast are often, I mean, they're similar um in terms of they're producing a tornado, but they often move at different speeds. Uh so like the storms in Mississippi, they're usually big storms that are moving quickly, like 60, 70 miles an hour, and they're hard to outrun. Sometimes in Oklahoma, they'll get what we call discrete storms that'll just pop up, maybe sit in a little area and then um fall apart, and that'll be the end of that. But because of the type of storms they get in Oklahoma, they actually get more in terms of the numbers of tornadoes, more often than not than we do. And that was where the original definition of tornado alley came from was how many do they get in Oklahoma? Uh, actually, Oklahoma, Kansas, and then the northeast part of Colorado were kind of the original tornado alley up into Nebraska as well. But there was kind of a maximum there. Um, but in the southeast, we don't get as many by number, but the ones we get often are longer track. So they'll they'll go much farther because they're attached to these storms that are moving much more quickly. And so more area, I guess, gets affected by tornadoes in the southeast than it does in um the plains on average. And there's there's plenty of research out there that's still being debated about that. Regardless, the hazard is important for everybody. Absolutely.

Shannon Harmon 12:54

I think that's why we have we need to have expert meteorologists to talk us through those storms, right? And I think that's that's what's so beautiful about what we're doing here. I mean, in a in recent um storm activities, you know, we've had Mississippi State students live streaming and lots of things going on that I think uh provide information to people when our electricity is out. There are other avenues that we can get information about what's going on and so we can take precaution.

SPEAKER_01 13:20

And we're trying to to innovate new ways to do that even now because the traditional TV weather model is slowly being replaced with something that's more like real time. I want access to it right now, instead of waiting till the five o'clock, six o'clock, ten o'clock news. And so we're trying to be at the cutting edge of that to figure out what other ways can our broadcast students do media to communicate these risks and challenges to um the public. Because that's really what we're trying to do is to communicate to the public and save lives.

Shannon Harmon 13:51

Yeah, and that's and uh and y'all do. And that even happened with our recent ice storm that we had. That it was just it was so great that uh that we are training, you guys are training um students to do that in real time. So let's talk about that. So uh what kind of um tools, access, experiences are students getting so that they are prepared and they that we are on the forefront of this innovation?

SPEAKER_01 14:13

So we're our program is set up in a way where we have a a cohort of just broadcast students, and those students will take four semesters of practicum where they're gonna go into the in front of the green screen on the first day and they're gonna just go up there and flounder a little bit, and it'll be funny. And they always make a little video at the end to show them how far they've come from the beginning to the end. But as they progress, they learn how to do a live shot. They learn how to go put together a forecast and working with graphics from one of the industry leaders for weather or weather graphics. Um, and then they'll learn how to do severe weather coverage, they'll learn how to do um just general, general purpose coverage type things. I'm um any of that kind of stuff, they're getting that exposure on the broadcasting side. All of that's happening concurrently in with what we're doing in the classroom. Now, what we do in the classroom, I teach a lot of the theory heavy courses. So I teach synoptic meteorology, which is really large-scale weather problems. So we're looking at math and physics, but we're looking at how the atmosphere evolves on the large scale. Um, and then I also teach dynamic meteorology, which is basically like a fluid dynamics course for an engineer with all the math and physics associated with that, but with an weather application. And so they'll be taking those kinds of courses at the same time that they're learning how to present things on TV. So they're having to learn the hard science while they're also learning how to take that hard science and communicate it to the general public in a way that's understandable, which is a hard challenge.

Shannon Harmon 15:44

Well, exactly. Um, it's like having all of the knowledge, but then also I don't want to say it perform, but that is in front of and keep a cool head, right? Because you have to communicate these challenging weather situations and not also be panicked that your audience is not panicking, right? So that's a skill set in and of itself that I think some people might have it and you know, and and and may not. Um and I think that's so great that you're y'all are bridging that in front of the camera and with the knowledge in the background.

SPEAKER_01 16:15

We also have a lot of laboratory experiences where we have students sit in front of a radar software and learn how to interpret parts of a radar image. Or we actually have a two-semester forecasting sequence where they learn how to make a forecast of all different kinds of wet weather, from large-scale weather to, you know, storms to mountain weather to coastal weather to all these different things so that they're equipped whenever they graduate to go do different jobs wherever they might be.

Shannon Harmon 16:44

So they're basically like simulating. Y'all are simulating what's gonna happen in a real forecast, and then they can um a question that just popped in my mind when you were saying that. So they're gonna be ready for uh their job for the workforce. Um, and and we're producing a large number of candidates across the United States. And so you guys are preparing them for the different regions, not just that's the goal. That's wonderful. Yep. Yeah. Are there certain students who are better at that different regionally? How does that skill develop?

SPEAKER_01 17:11

It's you know, students especially in what this is probably true in every discipline, but in weather, most students had some kind of a weather thing that impacted them, that made them interested in weather. And so those students tend to gravitate towards that thing in what we're talking about. More often than not, it's severe weather, like a tornado or a hurricane. So they've had some sort of an impact and they're gonna want to learn more about that thing. And then they'll gravitate towards learning more about that and then want to work in areas that are prone to those types of things. But we've had students interested in winter weather. I've had a graduate student that studied lake effect snow, which never happens here, but happens up in New York. It's um, but he just was really wanted to learn about it. And so we did that. And so um, we have a lot of those kinds of um students that just they have a thing they want to learn about and we try to equip them as best as they can.

Shannon Harmon 18:01

I love that.

SPEAKER_01 18:02

We also just want to make sure that they're they're ready if the job offer comes. So if they get an offer in Arizona, they know how to forecast desert weather, even if we're in Mississippi and we don't obviously deal with desert weather here, they need to be prepared to address whatever opportunities might come up.

Shannon Harmon 18:19

Again, coming back to that well-rounded um professional, that graduate, um, that's again such a testament to our program and why it's so good. Uh, I think that's great. Um, okay, so how are you using AI in your teaching with them? And how are they learning how to use the tools so that it doesn't replace their thinking, but it helps support their work?

SPEAKER_01 18:39

AI is it's a challenge for sure. Um, you could probably imagine that there's a lot of opportunity for um challenges with academic integrity with AI. Um, and my personal work, I don't have students write a lot of papers because we're doing more theory-heavy stuff, but I do have them work on problems. Like they have to do mathematical calculations using these complicated equations. And I always try to make it fun for them, as fun as you can make a dynamic meteorology problem. Um, I'm I'm kind of known, and you remember this when you came over to my office as the Star Wars guy in the department. So I try to theme a lot of my homeworks around Star Wars locales. So the students have to think about how I would forecast and or answer questions in this type of a situation. And usually I'll have AI help me come up with some of those ideas. Um, some of them are my own, but you know, problems you get tired of um running the same problems year after year, and then the problems are out there in the world and you want to try to come up with new things. And it as easy or difficult as it might be to write a question, like in a calculus class, it's difficult to write a dynamic meteorology question that's unique every single semester. And so an AI can help you kind of think outside the box and say, all right, I know the concept that I want the student to know from this problem set. How can I innovate a question that's like this one but different that will still make sure that they're getting the information and the knowledge they need from that question? And I've found it to be mostly helpful, but sometimes it it'll also not be helpful. And so I'll have to kind of adjust and take what it's what it's provided and work with that. And really, I just AI, I think a very important thing to think about with it is it's it's uncertainty. It's all a probabilistic model. It's all everything is related to probability, and probability naturally builds into the field of meteorology in so many ways. And so I try to communicate that uncertainty to the students from the AI perspective, from just an understanding that our weather models have uncertainty built into them and there's challenges there so that they understand the limitations of our systems. Like we say that we want to predict a temperature that is going to be 80 degrees today. It'll never be 80 degrees. It could be 80.1, it could be 80.2. Right. They're never going to be exactly right. And the weather instrument could be wrong, and the model could be wrong. And there's so many uncertainties built into our field that I feel like it's most important to communicate how uncertainty builds into meteorology, and that naturally ties into the uncertainty built into AI. It's all kind of intertwined in that way.

Shannon Harmon 21:18

I think that's so great that we think about too, though, this uncertainty is where we also then can critically think our problem solve beyond that, right? Yeah. So you you brought up that you've been using um AI in your um teaching. So thinking about what you're gonna go into your class and give them as scenarios, right? So coming up with, I love that, new um item banks, assessment item banks, new things like that, case scenario banks that you can you can then come up with. Because as you said, students will start to share ideas with each other. So the next semester students know what you were gonna ask if you don't change it. So that's so important that we're using it in this way to help our teaching, right? And instead of us um thinking about, oh my gosh, I've got to take all this time to rethink all these things, we can use it as a thought partner. And then when it doesn't give us something. That we wanted we then, but it gets us started thinking about that. Right. So we can tweak it. So it almost allows us to critically think without having to start at the critical thinking. It just gives us a a starting point. Does that make sense?

SPEAKER_01 22:09

It's a great, it's a great brainstorming tool.

Shannon Harmon 22:12

It is a great brainstorming tool. I I definitely think for our teaching and brainstorming and thinking about the lessons that we're going to go and enact.

SPEAKER_01 22:17

Right.

Shannon Harmon 22:18

And so I it sounds like you've been doing some of that. That's been working.

SPEAKER_01 22:20

Yeah, that's that's been very helpful. And I I mean teaching the some of the classes that I teach because they're so theory heavy, you know, students when they come to study meteorology, they don't want to learn the momentum equations. They want to learn how a tornado works and they want to know how to predict a tornado. And but yet they have to learn all of it, not just the parts that are the fun parts. And so I try to use the AI to to help me innovate ways to make things that would otherwise not be interesting to the students be more interesting to the students and maybe try to find ways to link ideas that are unique that maybe I haven't been linking in the classroom before to kind of provide a better cohesive narrative to how I teach the content that I teach, which is really important.

Shannon Harmon 23:01

Nice. So have you been talking with your students about you doing that uh for their lessons and for their for the work that they're doing? A little. Uh-huh.

SPEAKER_01 23:08

Um, I kind of I kind of keep that on the down low. I guess it's not anymore now if it's gonna be on this, but um, but I I do try to keep it on the down low a little bit because I it in my mind it helps to just for them to understand or to see the uh different ideas and to use I can use this as a tool without them thinking that I'm just relying on AI to do the work.

Shannon Harmon 23:30

I think it's really I think it's really important uh for us to be transparent and say, hey, I had this come up with, you know, help me as my thought partner so that I could give y'all scenarios that are real world. Right. Because I don't know right now what the what the weather in Arizona is, but you need to be exposed to that.

SPEAKER_00 23:45

Sure.

Shannon Harmon 23:45

And so uh Claude helped me do that today. Yep. And I I actually think that's really powerful. It is. Um, and especially where we're going in higher ed. Um, and so thinking about students need to know how to use the tools for their work, right, in the future. So then um, if we tell them we're using them for work, then let's now talk about how it's appropriate in your classes.

SPEAKER_01 24:06

Yes, and that's that's tricky um because it the the temptation is, well, I can just put this problem into the AI and it'll tell me the answer. And that that might be true, but they're that this is a challenge that I have as an instructor to try to come up with scenarios that the AI wouldn't be able to help the student with necessarily, or if it did, it wouldn't replace their ability to think critically. And so I always tell my students, and this is kind of where the rubber meets the road in a lot of my classes, is when we go to do a test, the test is going to be not just you regurgitating facts to me. It's going to be you take the facts that I've taught you in the classroom and you convert that into an understanding of some scenario that I'm going to provide you. So I'm going to say, here is this situation, here is this concept, figure out how those two link together, which is a critical thinking exercise, as opposed to uh relying on an AI to recreate it for you exercise. Um, so I definitely agree with that. I also agree that using AI as a tool to help you understand better is certainly a value. And I as we progress into our um the next generation of the meteorology program, we're working on developing coursework specifically in AI, where we're going to be looking at how students can interact with the AI, but not just interact with it like the large language models like Chat GPT and Claude and those, but instead working with developing your own and working with the data to understand how it works and really getting into the nuts and bolts of how it behaves because this is the future of in meteorology, but not just in meteorology everywhere is the future. And our students need to be prepared for that.

Shannon Harmon 25:49

Yeah. I think that's really crucial that we think about us as faculty on campuses um, how do we uh refine or reimagine our programming so that it does meet the needs of our students for what their future careers are going to look like, instead of maybe thinking, I'm gonna continue to teach this class because it's so great and I love it. And that might be the case. But how are we constantly refining and updating our curriculum, academic maps, um, implementing new courses? Right. Yeah, all of that. And I know that comes with a set of challenges. It does. Right.

SPEAKER_01 26:20

Well time is the big one, right? Finding time to develop all of these things. Absolutely. But but it's important.

Shannon Harmon 26:27

And I think it's such a testament to then um how we prepare students and we are the nationally recognized program, right? That's that's huge, such a testament to the work that that we do as faculty. That I don't know um just everybody out there knows what goes on behind the scenes to make programming, make academic programs appropriate and cutting edge. Yes. Um so kudos to you guys and and your colleagues um over there. I appreciate it. Yeah, you're welcome. I think that's really cool. I love what you said too about um experiences and students getting to experience different things. You know, I'm a big um fan of failing forward. And so sometimes we have to try things out and they fail, and then we have to move forward from there. Um, and so in your work, uh, how has that been been playing out with your students and these processes?

SPEAKER_01 27:14

The idea of failing is a day-to-day occurrence in meteorology. You know, we always hear we go out in the general public and someone will come up and say, Oh, you're a weather person. You're it's so it must be nice to be able to have a job and be wrong half the time. And we hear this all the time.

Shannon Harmon 27:30

Yep, y'all get blamed for everything.

SPEAKER_01 27:31

We do, yep. We rarely get the credit when the weather is good, but we always get the criticism when the weather is bad. Um, and that's that comes with the territory. But you know, the the atmosphere is it's kind of an interesting analogy. But if you kind of imagine your bathtub and you're draining water out of the bathtub and all the little chaotic flow that's happening in your bathtub, that's the atmosphere wrapped around the earth. And we're trying to predict that. And it's difficult, and there are going to be challenges, and the atmosphere will, you'll think, oh, I've got this nailed, and something will happen that you were not expecting to happen. And that's just the nature of being a meteorologist. And so we're used to failure. We're used to learning from our mistakes. If anything, that's is just part of being a forecaster. If if you're not learning from your mistakes, then you're not going to improve as a meteorologist. Um, for me specifically, since I don't teach a lot of the um hands-on forecasting classes, I actually do have a graduate statistics course where we do some basic AI work in that class. But then also I give them pretty much hourly, not hourly, but maybe two or three times a class period little tasks where we're gonna talk about some concept and then here's some weather data, now you try it. And they'll sit there and and struggle with it for a few minutes and maybe they'll get it, maybe they won't. And then we'll come together as a group and go over it and say, all right, this is what is correct. Figure out what you did right, figure out what you did wrong, so that they can kind of learn from their mistakes in the classroom in safety. There's no grade penalty, they're just engaging with it and they're practicing it and they're doing it. And the more of that that you do in general, the better they're gonna be at what they do. Absolutely. So yeah. And obviously that translates not just to statistics and programming, but also into weather. Yeah. The more weather scenarios you see, the better you are at it.

unknown 29:21

Yeah.

Shannon Harmon 29:21

Well, and I mean, I think that's a that's just a life lesson too. Sure. The more you're exposed to something, the more you the less you give up, the more you keep pushing through and have perseverance perseverance, uh, the more you're gonna have a greater contribution and learn and and how, you know, so it's how you react to those failures, right?

SPEAKER_01 29:35

I think it's meteorologists have to be humble because they will be humbled by the atmosphere every time.

Shannon Harmon 29:42

Well, and you know, the population of folks who, like you said, and and and the general population, yes. But we will let them know about it. Um, okay, so I'm hearing a lot of innovative things throughout our whole talk. Again, I I think I've said it, but I'm gonna say it again. This is such a testament, Andrew, to you being here and in this program. And it's so easy for um at this point in your career to just keep doing the same old things, right? Um, and uh, you know, because you're mid career faculty, it's it's it's easy to be like, you know, I got tenure, I got this, I got that, I don't have to keep plugging it away. So why do you think you continue to want to sustain this innovation mindset or continue to do this work? Because that really is a testament to you.

SPEAKER_01 30:26

I'll I'll say, and I actually told my wife this a couple months ago. I am probably more excited about research right now than I ever have been just because of the explosion of AI. Yeah, I've been doing AI for 20 years. I've been telling people AI is cool, it's the future, and funding agencies have been, I don't know if I'm totally sold on that. But now all of a sudden, everybody wants AI. And so I just I see so many opportunities to apply the skill set that I've developed for the last two decades. And I really want to jump on that train, as it were, make sure that I'm getting an opportunity to really take the research that I've always wanted to do to the next level by leveraging the opportunities that exist now.

Shannon Harmon 31:08

I couldn't agree more. And as a research one, we have that ability. You know, I think it's so cool to do the research um on AI tools, but also then um to research our classrooms and how our students, you know, if we look at our own teaching and learning and take that scholarship of teaching and learning approach with research, that could be a really cool thing. And I think we can lead that um basically where we're positioned in many of our academic programs to be the leaders about how AI uh affects teaching and learning. Sure. Yeah.

SPEAKER_01 31:36

Yeah. And I think that's I can envision more of that as time goes on, just seeing AI showing up in every classroom. But we just it's so new. At least the large language model part of it is so new. We're still trying to figure out how what's the best way to use it so that it's not taking away from the educational experience, but instead it's enhancing it.

Shannon Harmon 31:57

And I think that's where we can do so, yeah, do some research to say that. I think that's so cool. Thank you. Okay, so uh um, I always like to end our podcast with some advice. So, what advice would you give faculty, whether it's about uh AI or walking back into their class tomorrow or staying innovative? Uh, what's a couple of uh pieces of advice you can give people listening to this podcast today?

SPEAKER_01 32:19

So I'll I'll start by saying AI is scary. I know I I you hear it on the news, you hear it on other podcasts and stuff all the time that AI is this looming threat. The internet was kind of perceived as a looming threat.

SPEAKER_03 32:32

Yes.

SPEAKER_01 32:32

Um I'm sure before I was in school, even the calculator was perceived as a looming threat. And so AI has to really be thought of as a tool that can be used to enhance your productivity or enhance the classroom experience or whatever it might be, as opposed to something that's coming to replace you. Um we're always going to need domain experts. We're always in whatever field it is, and my field specifically, there's never gonna be a day where a computer is gonna issue a tornado warning. It's gonna be a person that has to issue that tornado warning and be accountable to that. And so they have to be trained on how to work in that environment and understand it. And so that's true not just in my dome, my field, obviously, but across all domains. There's always gonna be a human component. And so rather than look at AI as this looming threat, we need to think about how do we leverage it to be better at what we do, be more efficient at what we do, and how what can it provide us that we were unable to provide? It's like you and I said, it's it can work well as a brainstorming tool. Um, I could run down the hall and brainstorm with a colleague for a while and we would come up with some cool ideas. I can also sit in front of an AI and brainstorm with it and come up with some different ideas, some of which will probably not be very good, but some might be interesting. And you could use it to really kind of build up your research and do it in an efficient way, which I think is a really great opportunity. Other other piece of advice I think is just meet meet your students where they are, know where, know where the technology is, know that you know, five years from now, students that are going to be coming into college will have basically grown up on AI, at least in their, you know, their K upper part of their K-12, upper um education. And so they'll be they'll be masters at it, just like they are masters at everything else, technological for their generation. And so we we have to be able to meet them where they are. We have to figure out how to use these tools in the classroom so that they can continue to learn the way that they've been taught to learn for the last half decade or decade or however long it will be. I AI is not going anywhere. So we need to figure out how to use it to our advantage. And whatever that looks like, I think is still to be determined, but we're working on it.

Shannon Harmon 34:46

Yeah. I love that. I think keeping the students and what's in their best interest in mind is always a good thing, right? We have to evolve as our context of the learners that come to us evolve. And uh I I think that is, yeah, I think if we'll keep that in mind, we'll we'll do just fine. Sure. And so many of us on this campus and beyond are doing that. And and I so I think it's very important that we recognize it and we talk about it. Sure. Yeah. And we say that it's okay if even if we have to, you know, fail forward in the process uh or be scared in the process, that's okay too. Of course. Uh, but we're gonna move forward.

SPEAKER_03 35:17

Right.

Shannon Harmon 35:18

Um and that's what's gonna make uh things like your program nationally recognized. That's gonna make Mississippi State the place where people want to come to school. Um that's gonna make faculty stand out. You know, you said it's about your kiddos like you because you are Star Wars, but whatever it is, right? Like that's gonna motivate them to come and learn from us um and just make them better um and and our programs better as well. Right. Yeah. Well, thank you for sharing that. I appreciate that advice. Uh as we close today's episode, I'm reminded that innovative teaching doesn't happen uh just by ourselves. It's deeply connected, as you said, Andrew, to our domains or the disciplines that we teach um and and the tools that are shaping uh the way we teach and learn in the future. Uh what you shared today highlights this really important shift that I've been talking about with other colleagues across this campus. AI isn't just something we need to respond to. It's something we can intentionally integrate into how students learn, think, and prepare for their future work and making them workforce ready. Most importantly, this conversation reinforces that even as technology evolves, our role as educators remains the same. That's to help students think critically and apply what they know in complex and real world context. Just like you said, Andrew, even though they're not in Arizona, you're helping prepare them for weather in Arizona. At the Center for Innovation and Teaching Excellence, we believe that innovation is not a one-time decision. It's an ongoing process of inquiry, iteration, and care. And most importantly, it's work you don't have to do alone. As you reflect on today's conversation, I invite you to consider this question. How might you help your students engage with AI in ways that deepen, not replace, their learning? Thank you for joining us on Insightful Teaching. If today's episode resonated with you, please share it with a colleague. Until next time, keep teaching with an insightful purpose, keep innovating with an insightful heart, and keep being the reason someone believes they belong here.