How is AI Changing Higher Education? Rethinking College In The Age Of AI

AI appears to be everywhere.  In the news, on social media, on top of Google search results.  Higher education is no exception.  As artificial intelligence tools evolve, so does the college experience - but not always in the ways you might expect.

This week, Dr. Ross and student producer Ashley Worley discuss how the University of Texas at Tyler has seen AI integrated into campus life with Dr. Colleen Swain, the school's Associate Provost for Curriculum and Experiential Learning. The conversation explores the variety of ways - both good and bad - that AI has affected the average university experience.  From the integration of AI-powered assignments and teaching assistants to ever prevalent concerns about cheating, this technological advancement has found its way into the classroom.  Still, faculty and students alike wonder: what comes next?

Have more questions about the impact of AI on college life? Email us at ADRquestions@gmail.com. We'd love to hear from you!

(The production of this episode, including the text you're reading, was completed entirely by student producer Ashley Worley with no AI assistance.)

Transcript

Welcome And Why AI Matters

SPEAKER_01 0:04

Stay tuned to the Ask Dr. Ross Podcast. It's created to give you info to succeed at college. Our hosts are highly qualified. Dr. Katherine Ross is a member of the University of Texas Systems Academy of Distinguished Teachers. She's also a popular professor of 19th century English literature. Ask Dr. Ross is a community service of the University of Texas at Tiger.

SPEAKER_00 0:31

I'm here today with Dr. Colleen Swain, whose job it is to be. Well, just tell us what your job is, Dr. Swain, because that will help us understand why you're one of our experts on artificial intelligence.

SPEAKER_02 0:45

Well, it's wonderful to be here, Dr. Ross. I will say that I would not call myself an expert in AI, but I am over curriculum and experiential learning at UT Tyler. And how do we think about that as we deal with AI in higher education?

SPEAKER_00 1:13

And this is not the first time we have talked about AI on this podcast. About two years ago, with our very first producer, Nathan, we did a session on it where he was, as a student, was really thrilled about it. I was sort of cautiously optimistic. And then in the meantime, AI has taken off. We wanted to add here that Dr. Swain and I, and several other members of our administration, recently participated in a year-long AI summit? Is that what we called it? It was put on by the American Association of Colleges and Universities, AAC and U.

SPEAKER_02 1:49

It's an organization that really is focused on how do we improve experiences that students have, that faculty have. How do we provide that resource?

SPEAKER_00 2:03

It's put on a number of really wonderful university teaching programs. And so the special institute this last year was all about AI, and they brought in specialists from all around the country, but also I think they had over a hundred different universities. And Dr. Swain and I sat in on many of those episodes and came away dazzled in some cases, maybe a little more anxious in other cases. So do

Moving Past Cheating Toward Better Assessment

SPEAKER_00 2:30

you have a big picture sense of what you think about AI now and how it may impact not just UT Tyler but higher education in general, Dr. Swain?

SPEAKER_02 2:40

I think that AI is forcing us to change some of those staples that we always have had. Exams, writing papers is still important, but writing a paper to show that you know the content, it will also change, honestly, some of our practices as faculty. Already it's changing mine, I know. Absolutely. And I think what we'll see is we'll move away from this fear of what if the students are cheating? That has been around since there has been education. Since pencils. Yes, absolutely. And now faculty can really start thinking, okay, in the real world, if you need to know an answer to something, you could Google it or you could use AI to find that response. And quite frankly, businesses, workplaces expect you to do that. So now faculty have this neat freedom because in some ways it was testing you on the content that you're looking at. And now we get to think of assessments in ways that could be experiential, where you're doing something, or it could be where you're solving a scenario and you're using the knowledge that you have in this scenario. And if you do that well, that is not something that AI can generate.

SPEAKER_00 4:29

Before we get too much farther into this, I was going to ask you, Ashley, because you're our student voice here. Do you think most college students kind of know what AI is and what it can do, or do you think there's still a lot of confusion out there?

SPEAKER_03 4:41

No, I think it's definitely out of the bottle at this point. Just personally, I've seen it used a lot for varying reasons. I've seen really, really great uses of AI that actually help facilitate critical thinking, that help you to understand the material better, but it's just as easily used for the opposite, you know, to cheat on your work.

SPEAKER_00 5:02

Aaron Powell I think the phrase we've been using is cognitive offloading, letting it do your thinking for you. And you know, that's the worst use of it. Understanding how to use it ethically, how to use it not just ethically, but efficaciously. You know, so it is really to your benefit. Oftentimes I will use AI to generate problem sets or quiz questions, but I always ask it to give me twice as many as I need because I don't like about half of them. And of the others that I do like, I'm almost always going to change them some. But it gets me started. It saves me some time in that way. I was able to, through the help of a colleague at another university who's done a intensive couple years of fellowship on AI. He taught me how to make an artificial intelligence teaching assistant from my students, where I load into it just the materials for my course, just like the names of the texts, the assignments I give, nothing outside, and it's very carefully controlled so everybody's privacy is protected. But it's ended up being a very useful tool for students, you know, at

Useful Classroom AI Tools And Simulations

SPEAKER_00 6:10

three o'clock in the morning, they don't remember exactly what Dr. Ross said, type in the teaching assistant and it'll tell you, oh, well, she said this. So there are things like that that are good. But right now, when everybody's tired at the end of the school year, I think a lot of people are using it in irresponsible ways or just rushing in and not finished with their work and trying to use it. That's a big, big problem. What are some of the more inventive ways you learned about from this institute, Dr. Swain, of using it? Because what we really want it to be is adding on to what we're teaching, not detracting from it.

SPEAKER_02 6:45

Right. It's taking the knowledge and transferring it to different situations, to a different novel. Do we see the same things? So there's a lot that we can do. In terms of some of the interesting things I saw were simulations where you use this large language model, and there's many of them, and they're built for specific purposes. And the faculty member is able to create this simulation, and if you do this, it's gonna do this. So it's uh also adaptive. What a powerful way for students to practice. And what's nice about that is if you are one who needs to do it more than once, AI is very patient and is happy to do it.

SPEAKER_00 7:45

I will be very polite to you at nighttime, too. Yes. Well, one of the things I was able to do, you know, there are a lot of the novels I teach are now out of copyright, and you can load the whole document in there. And I was able to get my students to have conversations with characters in novels. Because it goes how to go back in and sort of find a place where the character talked about that. Now, sometimes it wasn't nearly as interesting a conversation as I had hoped it was going to be, but it was a start. And my guess is that was about two years ago I was doing that. There's probably 18 different agents out there now that will be able to do that. Would you tell us a little bit about what you learned about agentic AI from this institute, Dr. Swain? Because that's something that is sort of still mysterious to me a little bit.

SPEAKER_02 8:32

Well, what we're looking at there is where AI takes on a persona that is helping or guiding.

SPEAKER_00 8:43

And actually making decisions. I think you teach it how to make certain decisions. If this happens, then do this. If that happens, then do that. Is that correct?

SPEAKER_02 8:51

Yes. Where if you create your own, you're going to teach it. Well, this is my style of writing. This is how I write. Make it sound more like me.

SPEAKER_00 9:02

So in that way it becomes your agent. It speaks on your behalf.

SPEAKER_02 9:05

You can. For some, we might have a agent helping you deal with the problem. Well, it could be a relational problem. It could be an academic problem. So it could be teaching you, I just do not understand how to solve this kind of equation. I don't understand the process. I don't even know where to tell you I'm confused. And so as you respond, it's thinking about what you're doing and how to help you more. And what's also interesting is I heard a business professor who was teaching his students to make cold calls. And so he wrote a simulation to do that in an agentic format. So I mean, it felt like a person, it would respond like a person. And they had to do it at least once with AI and record themselves. And one student he said is very introverted and shy, and this was going to be a huge challenge for her. And she came up and told him, I did this, and I think he said like 15 times. But the response kept changing because it's learning what she says. So it continues to, in a way, ramp up the difficulty of the questions being asked back. And then when she went to make her cold call, she was like, that was so easy. It's a way to practice. I love when I see faculty say, try to outdo AI. Here's what I put in, and you know, I put in your reassignment. This is what AI said. So write a paper that's better than that.

SPEAKER_00 10:53

And that's the truth, is if a machine can do it as well as X, you have to be X plus. You have to be better. One of the things that I worry about is that there's a real divide out there. And I bet you've seen this too, Ashley. Some students are terrified of using it for fear they'll be accused of plagiarizing or not doing their own work. There's some who have not used it, and there's various machines that are supposedly will tell you if it's been used, and they have been called forth by a professor saying, I think you used it, and they were very upset because they didn't. And what this has given rise to, I think, has been suspicion and anxiety and bad feelings all the way around. And that troubles me a lot. Have you watched students go through any of that yet, Ashley?

SPEAKER_03 11:38

Aaron Ross Powell I've not known

Plagiarism Anxiety And Teaching AI Judgment

SPEAKER_03 11:40

specific students who had that problem. I know it's a very real concern, and it's something everybody thinks about now, especially when you're writing content. You write this and there's an extra layer of self-editing that happens because you're not only going, is it spelled right? Does it express the idea? You're going, does it sound like AI wrote it? Because the detection models, as far as I know, just from everything I've looked at, and I'm by no means an expert, they're not super accurate at the moment. So the concerns are valid. It doesn't mean that that situation won't change in the future. It just means that this is the area where there needs to be more work done for AI to be a really useful and effective tool.

SPEAKER_02 12:22

Well, let me push back on that a little bit. This is about AI and higher education. And I started by saying we're going to have to change. We have some freedoms to do some things that we haven't in the past. Because we don't need to give multiple choice tests. We don't need to do, you know, truefalse, fill-in-the-blank, identify this. You need to know fundamentals. But we have the opportunity to really help students learn what is the right AI tool, what is the right large language model to pick. And co-pilot is not appropriate for everything. And yet that's the only thing we have right now at UT Tyler. So we've got to expand. You know, to me, it's a moral obligation for higher ed to change. Because our graduates have to go out in the world. And even if we're saying, oh, well, we're training people for the academy, we know the academy is changing. So we have to prepare our students to go out there. Right now, computer science graduates are the ones who are feeling this. Matter of fact, at one of our employer advisory committee meetings, a person in the software development field said, AI does our entry-level programming. Well, we used to hire fresh graduates from college. But now those graduates need to be able to read the output AI produced, the code, and to look, are there errors? Because we know there are mistakes. Yes, we're trying to teach it to quote, learn, but its foundation is scouring the Internet.

SPEAKER_00 14:10

Well, and also it's my understanding that a really skilled coder still needs to be there to check the code and make sure it's working well. And if you have all these more sophisticated programs now that will do the early level coding for you, you don't train a lot of young coders to be more sophisticated coders. And it's stealing from them the chance to sort of practice, practice, practice till they're good at it. I think, too, that one of the things that I've used it for as an English professor, I find, and I teach English literature, not American, so I have to kind of orient my students to what's going on in the British history at the time when Pride and Prejudice was written. So they understand what was going on in the world that that literature was responding to. And yet I don't have time to give them a whole history lesson. And so I found that if I say, and I have a premium level of ChatGPT that I purchased, I'll say to it, in the period in which Wordsworth wrote these two poems, I need to be able to explain to students what was happening in England and France in this Napoleonic War. Give me enough of information about that so the student can see what the poet was responding to and what were the tensions of society at the time, so that it makes more sense. And that can usually crank out a pretty good short history lesson. But there's so little nuance that that's always where I get worried. I think it can catch me too, is what I'm worried about, Dr. Swain, is that, yes, it's a tool, but would it be better just to send them to the library?

SPEAKER_02 15:49

Well, and that's goes back to the literature we know on novices versus experts. Experts have over 10,000 hours minimum to dive in and think about how they're going to organize that in their brain and how what it connects to and what it doesn't connect to. But what we have to think about is how can we help our students learn content, but also experience the content in the world. And one of those ways is through internships or clinicals, practicums, and sometimes you can't have that, so you do the simulations. What a powerful way for your student to have an internship where they go and write copy and they take disparate pieces of information and think critically about them, put them together with what's happening in our world. Those are the skills that you teach. And I was looking at the NACE. NACE stands for National Association for Colleges and Employers. And so they are always looking at, are we aligning well? And they have a spring update from their fall. And they noticed that AI and job descriptions tripled from fall to spring. So, like on LinkedIn, the job description tells you that you need AI. And I was reading a part on internships, and 60% of the employers say they are assigning interns projects that use AI tools and skills. So part of our critical thinking that we're teaching also has to be how are you thoughtfully thinking about asking a tool to do what you need it to do and then be able to analyze the output?

SPEAKER_00 18:07

One of the things that I find interesting is we

Trust, Transparency, And Student Wellbeing

SPEAKER_00 18:10

did a survey in one of the workshops at the institute about how many of us were using it. About 80 percent. Of course, most of us were already early adopters anyway. But then when they asked, what percentage of your colleagues do you think are using it, it would drop really radically down to about 30 to 40 percent. And there's some folks that are really, really anti-AI, and I think that's unrealistic. Would you agree?

SPEAKER_03 18:34

Aaron Ross Powell Especially a lot of the basic and most accessible versions like ChatGPT, you know, the free versions, they're just really advanced search engines. In a way, in a way, there's differences. But in application, that's what I think most people are using it for. And the AI-generated results for even typing in something regular to the Google search engine right now. That's really what I think most people are using it for in their day-to-day life. And so it's kind of unrealistic to expect that to go away.

SPEAKER_00 19:06

Oh, yeah. But I don't think anybody knows for sure how it's going to end up. And so at the same time that we're trying to teach people how to speak intelligently to machines and teach machines how to learn from us, and I find that weird. I don't feel like I'm really teaching my ChatGPT how to learn from me, but it is, because I noticed, you know, for example, today I was putting in some questions, generate some practice quiz questions, and it referred it automatically to two of my courses without my saying which courses I was using it. So it already knew, you know, where to direct me. So that was really interesting. But I had a student tell me the other day, there's a group of students now who were absolutely entirely against it for environmental reasons. But what this one student said is it was addictive to her personally that she started talking to it like it was a friend. And she said, I was really getting hooked on this conversation. On the one hand, it made me really sad to think that a student didn't have enough human, real friends or family they could talk to, but also that it felt real enough. And I was wondering if either one of you have had any kind of experience with folks who are troubled by that yet?

SPEAKER_02 20:21

No, I really haven't. AI is ubiquitous in our world. And so to say I'm not using AI means you have to stay in your house and not And not turn on the internet or not do anything, really, because again, it is so ubiquitous. What we have to do is teach our students to be proficient practitioners of AI who understand ethical, disciplinary, societal, philosophical impacts of their decision to use AI.

SPEAKER_00 21:01

I've had some people say it's the end of the personal essay, that AI can write an essay so much better than most college students can. And I'm sorry to say that in some cases that not too far off from what my experience is with some, not college students, but kids right out of high school. Have you used it much in writing for yourself? Because you're a communications major, aren't you, Ashley?

SPEAKER_03 21:22

Aaron Powell Yes, and digital storytelling minor. So almost all of my work is project-oriented, is a lot of writing, is a lot of graphic design. And my personal policy is not to use it to generate anything. I use it to evaluate in some ways. So obviously for grammar, I like using AI for grammar, especially if it's late at night and there's something I may not have caught. I use it for cross-evaluating like data sets. If I've already taken a look at something, I'm like, I think my conclusion is correct, but what are some options I may have overlooked? It's great for that. It's great for finding what you may not have considered. It may or may not give you some great suggestions about stuff that you hadn't thought about before, but at least you will have considered what are the other ways somebody could think about this problem. So I've used it in that aspect. I don't use it to produce any content.

SPEAKER_00 22:16

I've had people at other schools say that they're using it very, very effectively for keeping track of student data, collecting the kind of data we as universities need to have to see how well we're doing. So I think that there's a lot of ways that it might streamline the operation of universities. But I'm trying to think how I can be sure that my students understand the risks. Because that's what worries me the most. What do you think?

SPEAKER_03 22:49

Aaron Ross Powell Well, if we were going to go back to what you mentioned earlier about the student who was talking to AI like it was her friend, I know that loneliness is at an all-time high for people my age. I know that's been a problem since social media became a thing. It was a problem before, it just got worse. And people have problems with connection. People my age and younger, they're growing up in an environment where you're constantly exposed to pieces of communication and exposed to what feels like other people. But because there's a barrier of a screen, people are much more mean. People are much nastier. And it makes you sort of have this perception a lot of times, if you grow up in that environment, that it's a very dangerous thing to engage socially. In my opinion, my uneducated opinion, I would think that's probably why social anxiety is such a big deal in people of my age right now, because they grew up seeing in various avenues that socializing is to some degree very risky. Well, these tools obviously aren't going to go away. And I think in terms of integrating them well into people's lives and to be able to use them in a productive way, it has to do more with, I think, addressing what is that actual root problem that's making somebody talk to an AI like it's their friend. That's what I would think is maybe focusing more on what is the loneliness issue, what is the underlying factor as opposed to is AI a good enough tool for addressing it? And then probably there's a way maybe that AI can be incorporated later. But I think finding whatever the root cause is would be the main thing to focus on.

SPEAKER_00 24:27

Yeah. Yeah. Well, you know, I think the jury is still out on how we're going to end up using it. I don't think anybody is at a university gonna look at us and say, you must use AI in every one of your classes. Although I think we're getting close to the point where every department's got to have some accountability that, you know, we have investigated how historians or mechanical engineers or accountants or computer engineers or English majors can, might, should, or should not use AI. And also just we can't predict what it's gonna do the job market, can we?

SPEAKER_02 25:02

Well, we do know right now what skills employers are asking for. And that's to identify and use AI tools that are appropriate to the task so you don't pick the wrong thing. Develop effective AI prompts to elicit quality output. They're asking for job applicants to be able to analyze and revise AI output as needed. And then some of them, I would imagine these are most in the software industry, but are to develop AI tools to increase work productivity. So we do need to really think about in our discipline, in English or in math, or in communication, what are appropriate tools to perform tasks that solve a problem. And I was just over at the Souls College of Business today, and computer science students were displaying their project in a class on AI, and they were solving different problems. One was an app that said, okay, I have this, you know, marking on my skin. What is it? And they

Experiential Learning For Real-World Readiness

SPEAKER_02 26:20

had trained their AI to identify the top 10 cancers or skin conditions. That's kind of neat. There was one where the group built a bracelet to track children.

SPEAKER_00 26:36

The lost children to find a lost child or a child wandering past the jungle gym.

SPEAKER_02 26:40

Yes. Where are they? What's going on? So there's a lot of really cool things that we can do. Part of it is our mindset. Are we approaching from a growth mindset as faculty and saying, yes, this is really different than what I grew up with? But what a neat opportunity for me to think about how can I use this tool to enable you to solve pressing problems in our discipline?

SPEAKER_00 27:13

Well, and the other thing that universities can do right now, and one of the people that participated in this institute with us is the head of IT for us, and he is indexing all sorts of different search engines and agentic AI and apps and all the different kinds of AI that are available so that we know what they do. One of the things that was a very powerful statement that I heard one of the folks said at this conference, you know, it used to be we were an economy of attention, where attention was the thing we were trying to always attract. Now it's going to be an economy of trust. And I thought, well, that's an interesting way of thinking about how this is going to start to change the world. Who do you trust? And you know, finally, I think one of our biggest jobs in college is teaching people how to have good judgment.

SPEAKER_02 28:03

Yes, and I think faculty can model that by saying, for this quiz, I ask AI to get me started. For this presentation, this part, AI was really helpful so that we are transparent, just like we ask you to be transparent with us on how you used AI. So I think a lot of that is about us saying, okay, we can do this. In terms of the people or the faculty members who well, I'm not going to use it, we are old enough to have known that the way librarians were trained has changed tremendously. We grew up with the Dewey Decimal system. In a card catalog. Yes. And I watched a group of faculty at a university. They said, We're not going to teach that. We're going to teach traditional content. Well, students vote with their feet and no one was applying to their program. And it was closed because there weren't students. And it was because they didn't update. And so this is a point where faculty aren't going to have the freedom to say, well, I'm not doing that at all. If you use Word now, it automatically is generating AI with you as it suggests in the editor. In higher education, we must get past this. We're cheating.

SPEAKER_00 29:37

I think at this point, the big question will be what happens the next five to ten years in higher education and really in the whole business economy with AI. I do want to put a kind of a plea out there to all faculty, is not be too hard on students as they're learning about it, because it's very tempting. It's very confusing. And I know for me, a lot of times, one of the hardest things for students to learn how to do is to write in a more sophisticated way. And you plug in your idea and and you say, AI, write this so Dr. Ross will like it. It'll make a pretty decent sound in essay. And I don't blame them for wanting me to like their writing, as long as it was their idea for starters. And so I try to wade in there and help them see where is it really theirs and where is their voice? And also to help them recognize that their own voice is a good one, and that their own voice is more important than any machine, and that we want them to learn how to use it.

SPEAKER_03 30:37

I had meant to ask you about this earlier. Could you tell us a little bit about what your responsibilities are as the director for experiential learning? Is that the correct?

SPEAKER_02 30:45

Yeah, well, I oversee experiential learning, and we can get to knowledge in multiple ways. We learn in the classroom, but we also learn outside the classroom. But we tend to learn more by doing and being active. And this is going to be where we really look at what experiences are we providing students within their discipline to learn by doing. Now we have a ton of this. As I mentioned earlier today, I was at a computer science showcase. Those students were doing. They were trying to solve a real-world problem with the programming that they had learned and what they had learned about AI. And I think two weeks ago, I was out on the North campus as the criminal justice students were exploring the crime scene that the faculty had put together. I went and got in the flight simulator where we teach nurses how to be a flight nurse? A flight nurse. And let me tell you, those places are really small. And, you know, they make the helicopter shake and do all kinds of things. They have real-world simulation. We have a great deal of this on campus. We don't document it well. We don't showcase that at UT Tyler, you are not just coming to take two tests a semester and go. We're working to get you to the end goal, which is always graduation, but hopefully you just don't graduate and do nothing. We want you to graduate and go make a difference in whatever way speaks to your why. Why are you here?

SPEAKER_00 32:45

I was talking with my students today. We were talking about looking back, what worked and what didn't work. And they almost always said it was when we talked with each other, when we did stuff, and not just sitting there listening to the professors. I think that AI is going to challenge us in lots of other ways too. But one will be to make sure that we're giving people meaningful experiences. So, Ashley, are there other questions you have for our wonderful guest here?

SPEAKER_03 33:13

Maybe we could just wrap it up with the best way that both of you have seen AI shaping how people teach now in higher education.

SPEAKER_02 33:23

What I like the most about AI is that it forces you to think differently on how you are going to assess the student's learning outcome. I'm not limited anymore. I can ask them to create

Creative Assignments And Closing Invitation

SPEAKER_02 33:40

an infographic. I can ask them to write a song. So there's so many things that I can do now that if I'm thoughtful and purposeful about linking my student outcomes in my assessments, where I give students choice, where students can also write a reflective paper about why did I choose providing that rationale. That's the thing I like most so far. Now, again, we're always learning, but that's what excites me the most, where I don't see students running to the bookstore to buy a blue book, or I don't see a student running to the bookstore to buy the Scantron. That we're all thinking differently. It doesn't mean that every course is always going to be using AI, at least for now. I do think eventually every course will be using AI because it will be so embedded within our society that we can't imagine it without it.

SPEAKER_00 34:50

Well, I'd have to say I don't know that I'm using it that creatively yet. I feel as though I'm using it for efficiency. And one of the single best things is this internal AI teaching assistant that I have. I cannot tell you how many times 15 different students will ask me the same question that is laid out in my syllabus. The other thing was, I think, also setting up my questions so that I'm asking more challenging questions. I find that's a useful thing to make sure it's a check on me. But the other thing is before you get started on these research projects, it's a really good, what's the lay of the land? Okay, so if you're gonna read North and South, which takes place in the Victorian period, and you know when the main characters leave the Anglican Church, what is the Anglican Church and why would anybody leave the Anglican Church? And that's there are shelves of books written about that, but this can get you started and saying, okay, here's where you want to start. So I think we just have to keep being creative and remembering finally that it's a very challenging time. You know, it really is. And it's a stressful time. And so, in addition to everything else, we've got to learn to be real careful and kind with each other.

SPEAKER_03 36:00

Well, thank you so much for talking with us. I also want to take a minute to thank our listeners. If you have any questions about how AI is shaping higher education, and you have any you'd like specifically to direct to Dr. Ross or to Dr. Swain, you can always leave a comment under our YouTube video or email us at adrquestions at gmail.com.