Elizabeth Riede: All right, everyone, it is time for the humanity check. ⁓ So for each of you, what is one rule of great service or leadership that you wish everyone followed? Welcome back to Human at the Counter. It's been a couple months, long enough for a new semester to start, a few new platforms to roll out, and at least one AI tool to quietly enter your workflow, whether you invited it or not. The calendar may have flipped, but if you lead in hospitality, operations, or higher education, you know the pace never really slows down. It just resets and keeps moving. Here's what I'm taking away today. AI is here. The question is not if we use it, but how. In dining, in advising, in every corner of campus life, our choices show up in human moments. Who got nudged? Who felt left out? Who trusted the system? And who didn't? So here's your final humanity check. Are you using AI to replace people?
Keith: Tom, you want to go first?
Tom Van Gilder: Yeah, I'm happy to. So I was reflecting back on my days of ⁓ director of IT support. And maybe at the time when I took over that area was maybe not known as awesome customer service. I found this really cheap. Well, know, times have changed, right? But I found this really cheesy video of this guy who owned a restaurant and the name of the video was given the pickle. I...
Elizabeth Riede: Say it isn't so. or to release people to do the work that only humans can do? If you can answer that honestly, you're already ahead. Thanks for listening and remember, on a campus, as in life, the fastest path to trust is still transparency. This has been Human at the Counter. I'm Elizabeth Reedy, and remember, being present is everything. If you work in hospitality, the kind that lives in details, you know service is shaped by systems as much as smiles. ⁓ schedule, the point of sale, the inventory order, the follow-up email, and increasingly, those details are being shaped by artificial intelligence. Today, we're looking at AI through the lens of a university campus, but whether you're running a dining hall, a hospital kitchen, a multi-unit restaurant group,
Keith: Thank you.
Tom Van Gilder: kind of made my staff and now I'm the give them the pickle guy hence forward. 20 years I've been the give them the pickle guy. But, you know, it's just a message about strong customer services like critically important to the greater good for our campus. So I think that's my one unwritten rule is like just give them the pickle. Just, just, yeah.
Elizabeth Riede: You
Keith: .
Elizabeth Riede: or overseeing student services, the leadership tension is the same. How do we use technology to strengthen service without letting it quietly erode trust? I'm Elizabeth Reedy. I've led teams in hospitals, coffee shops, and now higher education dining. And here's what I've learned. Systems scale efficiency. Culture scales humanity. And when those two fall out of alignment, people feel it.
Keith: Good. Good.
Elizabeth Riede: I love that. Yeah, that's an old Bob Feralism. Yep. Yep. Yep.
Tom Van Gilder: That's it Bob Fairley, you know it, you know it.
Keith: All right. Tom's already overshadowing me. ⁓ I think my best piece of advice is listen. know, I like Tom kind of harkened back over my career, doing everything from business development all the way up through practice leadership and into my current role. Being able to listen and actively listen, I think is so critically important to leading and ⁓ more important to customer service.
Tom Van Gilder: Hey
Elizabeth Riede: I'm gonna ⁓ and fast. Culture decides whether technology feels like service or surveillance. Higher education isn't at an inflection point. Artificial intelligence isn't a pilot project anymore. It's embedding in advising platforms. It's embedded in learning tools. It's embedded in dining forecasts and help desk tickets and accessibility software. So today we're not asking, should campuses use AI? Yeah, in one of my previous episodes, we talked about a lot of people listen to answer. But if you're actively listening and listening to engage, it really takes your leadership to a whole other level. So I agree with both of those things. All right, name one thing that a human can do that no AI or robot could ever truly replicate.
Keith: Absolutely.
Elizabeth Riede: We're asking something harder.
Keith: I think a lot about a lot of the trades, right? And I know that's probably everybody's go-to, but you know, I've got two boys in their twenties and you know, I've kind of steered them more into trade related opportunities because I think, you know, I don't think there's any, well, there may be a few exceptions, but for the most part, I don't think that's going to be taken over by AI.
Elizabeth Riede: Who decides how it's used? Who protects students when it's wrong? And how do we build systems that scale without shrinking trust? Yeah, yeah. What about you, Tom?
Tom Van Gilder: Yeah, I think it's for me, it's around the, don't think a bot's ever gonna really help us decide what's fair and what's not fair. It's like, are really intended to more about what's efficient and not efficient versus where the human's really, you know, how to have that moral sense of fairness kind of stuff. That's me in my core.
Elizabeth Riede: Mmm. To unpack that, I'm joined by two leaders who sit at the center of these questions every single day. Keith Warner, Chief Information Officer at Appalachian State University, and Tom VanGilder, Deputy CIO. Leaders responsible not just for uptime in cybersecurity, but for policy, governance, and the guardrails that determine how innovation shows up in students' lived experiences. This isn't a tech episode. This is a leadership episode. I think those are both such awesome answers. think it's so funny when I talk to hospitality folks, it's usually, the answers are usually something like really truly ever love or really truly smile. And so I think there's a concreteness to your answers that I really appreciate because I think you're right. ⁓
Keith: We are IT folks, Elizabeth. just ultra pragmatic, I think. ⁓
Elizabeth Riede: So you don't love or smile. That's what you're telling me. So let's get into it. Before we jump into frameworks and policy language, I want to zoom out a little bit. Imagine it's registration week. A first generation student logs into a portal at 1147 PM. They're trying to figure out financial aid, housing, and whether they can afford a meal plan. The system suggests courses, a chat bot offers help, a predictive model flags that they may be at risk of dropping below full-time status. No, I can appreciate that. All right, if you walk into your favorite cafe, diner or restaurant right now, what is your go-to order?
Tom Van Gilder: funny. I like a, the ACP shrimp wrap at the taco truck. can't, I'm losing the name of the taco truck, but, ⁓ it's just like an awesome, my favorite go-to dish with shrimp and veggies and some rice and cheese. It's like, it's got all the goodness. It go wrong. That's it.
Elizabeth Riede: On paper, this is proactive support. Behind the scenes, models are running probability scores, systems are prioritizing outreach, alerts are firing. Can't go wrong. It's got rice and cheese in it. Can't go wrong.
Keith: I'm, I'm eating potatoes. mean, give me a good prime rib, good steak. I am, I am good. ⁓ but you know, if I went to any random, ⁓ kind of diner, I think it'd probably be a hamburger. You know, I'm pretty, I'm pretty simple.
Elizabeth Riede: There you go. There you go. All right, go along with your hamburger. If you were a condiment, which one would you be and why?
Keith: I gotta go with ketchup. I mean, it's ubiquitous, right? I mean, it goes with everything. So I've gotta go with ketchup. And Tom, if you say anything different, I'm gonna be shocked.
Elizabeth Riede: Thousands of signals produced in seconds to prevent friction before it becomes failure. But in the student's world, it's either, wow, this university anticipated my needs, or how did they know that about me? That line between helpful and unsettling ⁓ is razor thin. Love it.
Tom Van Gilder: ⁓ My wife has turned me on to the chili crunch stuff. So even though it can sometimes cause me some indigestion at the early
Elizabeth Riede: ⁓ yeah? My listeners will not see this, but yeah. A little flyby thing. Yeah.
Tom Van Gilder: ⁓ you got it. got it. Yeah. So we have found our favorite. We've tried like four or five different chili crunches and like, nope, this one that we have to special order. But it is best.
Elizabeth Riede: And here's where it gets real for operators like us. If dining uses AI to project traffic and reallocate labor, if advising uses AI to nudge enrollment decisions, if IT uses sentiment analysis to escalate tickets before complaints trend, who defines what's fair? Who audits the unintended consequences? Who makes sure that efficiency doesn't quietly outrank equity? Of course, it's always the one you have to special order. That's always the best one.
Keith: I'm gonna have to try that. What are you putting it on, Tom?
Tom Van Gilder: We put it on eggs and we eat a of them. It's got a little kick to it, but it's really good. I'll bring you one of them. Because I think that to get a special discount price, we had to order like 10 of them. So I think I've got plenty to share. ⁓
Elizabeth Riede: Pizza, noodles.
Keith: Really? Okay. It's the full, bursting. ⁓
Elizabeth Riede: Because optimization is not a neutral act. It reflects values. It reflects what's important. And that's why CIO leadership right now isn't about servers and storage, or isn't just about those things. It's about stewardship. So Keith, Tom, let's talk about how you think about that stewardship. What does responsible AI leadership actually look like on a campus? But first,
Tom Van Gilder: You
Elizabeth Riede: Alright, if you could have an unlimited lifetime supply of one snack, what would it be? Besides chili crunch, Tom.
Tom Van Gilder: Thanks.
Keith: .
Tom Van Gilder: I'm a potato chip kind of guy. I could go with my rest of my life without sweets, but potato chips are my jam.
Keith: I try to stay away from the sweets, but if I had to have one go to, it's Reese's peanut butter cups.
Elizabeth Riede: That's a good, peanut butter is a perfect food. I firmly believe that there are two perfect foods in the world. Peanut butter is one and yogurt is the other. And peanut butter and yogurt together? All over it. Oh man. It's time for the Humanity Check Questionnaire.
Tom Van Gilder: Okay.
Keith: It is. Okay. Now I'm one out of two on those.
Elizabeth Riede: All right, if a robot took over one chore in your life forever, what would it be?
Keith: I think for me it'd be laundry. That's the one thing I least look forward to every week, I think.
Elizabeth Riede: That is a shared trait, my friend.
Tom Van Gilder: Yeah. For me, it's cleaning the bathtub because I just can't stand it. And I do like like occasional hot baths. So I use it quite often, but just getting in there and cleaning the darn thing is just no bueno.
Elizabeth Riede: It's not set up to be clean. You would think because it has the built-in drain that it would be, but it is not set up to be clean. You have to get in it and then you have the soap and then you have to not get it on your feet. Yeah, it's hard. All right, you're stranded on a desert island, but somehow there's a working food truck. What is it serving?
Tom Van Gilder: mean, I'll go back to that shrimp quesadilla. hope that, ⁓ yeah, hope they're okay. ⁓
Elizabeth Riede: Yeah, it's got to be there.
Keith: Probably fish tacos for me.
Elizabeth Riede: Yeah, it's good. You probably have some fresh fish around too, so that's probably good.
Keith: That's a point. I didn't think about it at that angle.
Elizabeth Riede: Alright, you could have lunch with any fictional character. Who would it be?
Keith: I guess it's my turn. This is a tough one for me. Probably ⁓ the Kevin Costner character on Field of Dreams. ⁓ That was like a, that was a magical movie for me. You know, the father son connection and I'm a big baseball guy. So ⁓ that would be mine.
Elizabeth Riede: Good one. Nice.
Tom Van Gilder: I grew up on the Huck Finn and Tom Sawyer books. so getting on the ⁓ homemade raft on the river sounds pretty amazing to me as a young kid. that would.
Elizabeth Riede: Tom, you are the second guest on this podcast that has picked that. That's great. I'm gonna have to make that happen somehow. I bet AI could make that.
Tom Van Gilder: Is that right? I'm sure, but that's one thing I don't want AI to do. I want to go do that raft myself. I'm getting too old for it now.
Keith: Mm-hmm.
Elizabeth Riede: Nah, never. Mountains keep you young. If your life had a theme song that played every time you walked into a room, what would it be?
Tom Van Gilder: I guess that's me. I guess I'm a Queen guy. So Bohemian Rhapsody is probably my go-to. just, Harkin's days of rocking that song with my kids and now I have grandkids and I'm turning them onto it as well. So like, let's do the mama mia, mama mia stuff. And they're rocking it with me. that's, yeah, they do.
Elizabeth Riede: Nice.
Keith: These answers show our age too, don't they Tom? Mine's got to be Mr. Brightside.
Elizabeth Riede: Love it. Not at all. Nice! Y'all picked two great karaoke songs. I appreciate that.
Keith: which you will never hear me do.
Elizabeth Riede: ⁓ come on now. I didn't tell you that's the end of this podcast.
Keith: I think I have a connection problem.
Elizabeth Riede: Now I know somebody in IT who could fix that for you.
Keith: Yeah, so do I, but I don't know if I'll call him with it.
Tom Van Gilder: Ha ha.
Elizabeth Riede: All right. Well, thank you guys. You passed the Humanity Check. ⁓ I think this episode is really all about the intersection of how technology and governance of technology supports the hospitality experience. I think we were talking before we started about the fact that ⁓ increasingly technology is becoming a part of our hospitality world and in ways that we didn't expect and in ways that we can plan for. ways that we didn't anticipate. And so think I'd like to start out by talking a little bit about your roles ⁓ how the government framework shape not only the work that you do, but the way that our hospitality work is shaped by those roles. ⁓ I'd love to hear a little bit from your perspective ⁓ about your roles and the work that you do.
Keith: Sure, and I'm happy to start Tom. So I serve as the Chief Information Officer here for this organization and Tom is my Chief Deputy and Chief Operating Officer for the division, which basically means if you encapsulated it is that we work ⁓ in partnership ⁓ and artificial intelligence is probably top of the list for both of us. You know, I think as we kind of launched into this collectively, it was really about How do we set up a governance framework that makes sense to drive our organization forward ⁓ in an ethical way, a responsible way, a secure way, and a sustainable way, or as much as we can from a sustainability perspective? ⁓ So we initiated a task force that was put together nearing two years ago now. ⁓ Tom had the really bright idea of having ⁓ working groups or what we call them enablement teams that are ⁓ kind of feeder working groups into the task force. And Tom can go into more detail about those, but it is, that has been a ⁓ boost for how we manage our intake of projects and use cases, policy considerations, sustainability, as I mentioned before, ⁓ and thinking about that organizationally and bringing the right stakeholders together. to think through it with us and hospitality being a large part of that. So we value those partnerships because Tom and I aren't experts on all the business processes across our organization, nor should we be expected to. So those partnerships are critical to us thinking through potential use cases and how can we get the most value from AI.
Tom Van Gilder: Yeah, so the AI enablement teams that Keith mentioned, it's which I kind of put a call out to campus and just said, hey, you we've got these various working groups that we know are important to us. ⁓ Everything from AI for teaching and learning, for example, AI for research. Now that we're a research to designated university, this made a lot of sense that we had start focusing on AI for research and. And we've start thinking through like AI literacy for our campus. And so we've got another working group. And I'm smart enough to know that I need some excellent leaders along the way to help me with all these conversations. So for like the teaching and learning, we talked with this fellow from College of Ed, ⁓ the faculty member helping us lead some really rich AI discussions around teaching and learning. ⁓ There's some faculty consternation around ⁓ AI teaching and learning, as you might imagine. across our campus, which is not uncommon across other universities at this point in time in our lives as well. But what does that look like for our campus and what are the things that we should be bringing forward and or not bringing forward in the teaching learning? So we've had good, rich conversations around things like using AI detection tools ⁓ in teaching and learning. Should they or should they not be? And I think our stance is that if they're maybe 92 % accurate is kind of the standard rate right now for some of these detection tools, for example. So if they're 92 % accurate, that means 8 % of our students, we're going to disenfranchise from the academic mission of our university and call them cheaters for lack of better words. I don't think so. And the faculty are the same thinking as well. It's like, we don't want garbage. So once it gets 100 % accurate, come talk to me kind of thing.
Elizabeth Riede: Yeah.
Tom Van Gilder: Anyway, so these working groups are working really well. We've got a good partnership with our Office of Research folks who are helping us lead that research team. our academic technologies, ⁓ academic and emerging technologies. just changed the name. I'm still getting used to it. ⁓ So academic and emerging which is a thoughtful name change, actually, because we're starting to think through all these things that are coming at us ⁓ like AI. And so we need a ⁓ ⁓ campus as to be thinking through. what this means for us as a campus.
Keith: That's a really good point too, Tom, about the literacy. I think that was one of our early understandings is that foundationally we needed to provide workshops and training and professional development opportunities to learn about AI. So we had some common understandings and I think we've done that really well actually. And I think that's really benefited us as we continue through this journey. The other thing we did was standardize on a few tools that we that we could kind of own and control the flow of data. So I think those were a couple of the really important moves that we made early on that has really helped us.
Elizabeth Riede: think you're touching on a couple of things there that I'd like to dig into a little bit. The first is around trust and the importance of trust. think in hospitality, talk about, ⁓ need trust with our customers, we need trust with our employees, and we need our employees to trust our customers, and we need our customers to trust our employees. ⁓ And I see in technology, you were talking a little bit about with the detection tools that
Keith: Hmm.
Elizabeth Riede: there's this possibility for the introduction of mistrust or distrust, right? Or maybe that there's some reservations, understandably, ⁓ about whether they should trust this new technology. ⁓ And then the other thing that you guys were talking about is literacy. ⁓ And so I think as you're thinking about where you're leading and advising on whether it's AI policy or the tools that you're using, What are some of the guardrails that you're putting in place ⁓ when you think about both opportunities, but also the potential for some of that either mistrust, distrust, misapplication, et cetera.
Tom Van Gilder: Got a great story too, and it literally just happened today. Our budget officer is taking a class and the faculty member just like got a little bit PO'd because he can tell that they're all using AI for all the problems that he's putting out for homework. And he's like, damn it, I really want you all to learn and you're not learning if you're just throwing in the AI slot back at me and that's not what this assignment's all about. So a really interesting timing on her part to... message Keith and myself about this, like, man, this is like spot on, we need to be having some talking about, you know, what it means for like ⁓ assignments and assessment in the day of AI kind of thing. So I think this is where that literacy piece is really critically important and how to get our faculty, they're still going to do that. I mean, it's so maybe it's figuring out how to best utilize the tools that are afforded to folks and how to, you know, like I said, change your assignments and assessments.
Keith: Thank you.
Tom Van Gilder: way as we figuring out these new tools that are afforded to us.
Keith: And trust, think, also carries forward just with staff as well as faculty. So, you know, we have a lot of trust with our staff with sensitive data or confidential data. So, you know, part of having standardized tools is we can control it and we're not feeding, you know, what they call larger language models, which are public models with our very important data. you know, there's a...
Elizabeth Riede: Mm-hmm.
Keith: A lot of trust that goes into that. We're investigating tools that, you know, kind of center that, but you know, if a staff or faculty member brings their own AI tool of choice and feeds those models, we're losing sight of that. So we're looking at tools that we have line of sight to that data and making sure that we know where the flow of the data is and where it's going and who's doing it. But it's complicated and oftentimes expensive to get these tools.
Elizabeth Riede: Yeah.
Keith: So it's definitely been a challenge. think guardrails are going to continue to morph over the next few years. This is such an evolving area that I think we're going to see better guardrails to Tom's point on the faculty side. And I think we're going to see better guardrails in terms of which models or platforms that you can use. The other challenge that we have now is every vendor that we deal with has some AI functionality associated.
Elizabeth Riede: Mm-hmm. Mm-hmm.
Keith: So, know, A number one is how do we assess that functionality and how good it is, but the other is how safe is it? So there's, it's a really good question with not a lot of really good solid answers yet, unfortunately.
Tom Van Gilder: I think some of it too is us kind of thinking through and formalizing some of our processes. So what I did mention was, is we also have an AI technical team, which I think you're very familiar with, Elizabeth, kind of helping our campus with figuring out campus processes and can we do more efficient things, yada, yada, yada. But, you know, we're starting to get asked questions like, hey, could I use this for X, Y, and Z? ⁓ And like, yes, but, know, let's think about maybe putting some guardrails up, as Keith mentioned, and I call it the human in the loop kind of guardrails that maybe almost developing a contract of sorts saying, all right, we'll develop this, but we've got to have some human processes on this so that we're not just giving over to the AI to make the decisions for us kind of thing, but having some real human skin in the game, if you will, for the...
Elizabeth Riede: Mmm. We talk about that a lot in hospitality too. think there's always a touch point that should stay human. ⁓ know, even if you've automated away these places where there's friction in whatever journey, if you're mobile ordering a sandwich or if you're placing a grocery order that's being delivered to your house, at some point in that process, there should be a human. that understands the experience that you either are currently having or have had, ⁓ so that they're staying connected to that in some way. And so do you have specific kind of like no-go moments where you would say this should never be automated or this bot agent process algorithm should always hand off to a human? if it involves X, Y, or Z.
Keith: It's a good question, Tom. I don't think we've run into any particular cases yet, but that could change literally tomorrow. Tom, I know you work more closely with the technical team than I do. Have we run across any?
Tom Van Gilder: I'd say we're not at a full stop on a couple of the projects, but definitely starting to, hey, let's talk about this a little more and really think through how we do weave in the human processes in addition to the things that are being developed.
Elizabeth Riede: Well, I do want to not gloss over. I I think the example you used earlier, Tom, was a good one around the academic integrity and the code of conduct. I mean, I think when you have 8 % of your students that could potentially be misclassified, I mean, I do think that that's a moment that it's like, hey, we should probably have a human take a second look at this and make sure that that really passes the sniff test. If you have someone who is obviously engaged with integrity, and now is being flagged for some reason, you maybe we don't gloss over that. Maybe we don't take that 8%. So I do think that that's a moment that really resonated with me. I'm sure that there are others that will emerge ⁓ as we're going through these conversations.
Keith: think on the business side, I think about hiring, you know, because there could be inherent bias in some of these AI applications. ⁓ So especially if you're using AI to potentially weed out, you know, first run of candidates, I mean, that could be a fear. ⁓ So I think there's going to be several of those types of use cases that emerge, especially at our organization, as broad as it is. ⁓ But that's one that could potentially ⁓ be an issue.
Elizabeth Riede: Mm. Hmm. Yeah. Mm-hmm. Zooming back out a little bit and thinking about AI across campus or AI in education applications, where do you see AI kind of already outside of maybe unethically doing some classwork for students? Where do you see AI transforming that experience or engage or impacting that experience right now?
Tom Van Gilder: say like, ⁓ know, we, Keith and I were on a recent vendor demo presentation about ⁓ like AI early warning system kind of thing for our students and helping with student success kind of initiative. So ⁓ a student missing an LMS login or ⁓ their quiz scores are declining or whatever, ⁓ or even, you know, financial aid gaps, things like that to kind of.
Elizabeth Riede: Mmm.
Tom Van Gilder: nudge the advisors to also nudge students, especially if they're categorized in the high risk ⁓ kind of sense. So I think that's pretty interesting. ⁓ They also offered some kind of degree mapping, ⁓ helping students visualize how changing a major and or failing a class, et cetera, affects their graduation date in real time kind of thing.
Elizabeth Riede: Hmm, I love that application. think because that can be almost like speaking another language for some of our students. ⁓ I think that's a really, really good use case.
Keith: think, ⁓ know, I always kind of think about it from a day in the life perspective ⁓ a student, right? And I think advising comes to mind. I think there's plenty of opportunities around advising and making sure that ⁓ have the appropriate resources for our students. ⁓ if we ⁓ that and do that really well, it improves our retention rates as a campus. ⁓ so many opportunities, I think, with students, you know, access to clubs, you comes to mind. ⁓ How do they get better integrated into the social framework of our campus? I think there's plenty of opportunities to get more engagement with our students, which I think is the cornerstone of what we're here to do. It's not just to teach, it's for an experience. So I think there's plenty of opportunities. I think Tom mentioned some really good ones. Tom, I know you're really ⁓ leading up a lot of our accessibility efforts too. But I think that's an area as well,
Tom Van Gilder: Absolutely. So there's AI tools now that kind of automatically generate like high quality alt text for images kind of things. So we're looking at these kind of tools as well to help our campus. But really benefits our, can benefit our students too, especially if they're maybe identified as learning, having learning difficulties. So there's some really nice tools that kind of help our students that are having some learning challenges to, you know. them with their education.
Elizabeth Riede: I think that's a really interesting application because we sometimes think about, ⁓ or hear the conversation with AI, know, Keith, you mentioned with the hiring platform, this idea that AI kind of has some of the inherent or can have some of the inherent biases that exist in how it was trained. And, but I also think there's the really, really cool phenomenon that can happen where it can actually kind of challenge what's possible.
Keith: cam.
Elizabeth Riede: in a way that a human wouldn't. And I was having a conversation with one of my other guests about it just kind of broadens out the sandbox and then it becomes the human's job to kind of rein it back in and go, well, that's crazy. I'm definitely not going to do that. But I didn't even think about the fact that it could do X, Y, or Z or that that would be possible in our operation. So I think that that's a really, really cool kind of untapped potential or maybe newly tapped potential for AI. Advising is another one or kind of those early warning signs. I I think that's a conversation that we have a lot on the campus dining side is like, what is our role when we have residents who are eating with us? And we have so many touch points in dining with our students that are genuine and authentic touch points and not classroom environments. ⁓ that can provide those kind of early warning. And so I think that's another place from a hospitality perspective that AI could really provide some links and insights for us.
Keith: That's an excellent point. I haven't thought about that one. That's really good.
Elizabeth Riede: When we think about things like that, where you are looking at kind of student patterns or data or, know, how, from your perspective, what should we be thinking about to ensure that these tools stay supportive and don't feel like the university kind of surveillance state, right? ⁓
Keith: We don't want to be that.
Elizabeth Riede: I guess to kind of put it another way, like where do you draw the line between that like proactive support and finding those moments of community engagement that we know are so important without being kind of creepy helpful?
Keith: Yeah, and we don't want to be a part of this surveillance state either. ⁓ So I think a large part of the answer is working with ⁓ our executive stakeholders like yourself and knowing where those lines are. And then designing and architecting our AI solutions to kind of fit. I just don't think there's a one size fits all for these type of things. On the back end of these systems, ⁓ we need to think about retention of
Elizabeth Riede: Yeah.
Keith: that we put into chat, how long do we retain those, ⁓ does public records laws look like, how much do we have to account for that, has the law even caught up to ⁓ of AI queries and prompts. again, I keep saying this, know, but the next couple of years are going to be really informative about ⁓ types of ⁓ questions. ⁓
Elizabeth Riede: Hmm.
Tom Van Gilder: add to that. It's kind of on us to add things like university contract and adding clauses like zero training clauses where enterprise university licenses that we're using mandate that no student data is going to go train the larger, you know, vendor global model kind of thing. And our data is our data and stays within our university's kind of digital fence, if you will. And then the others, Keith mentioned the chat bots is, you know, we can make the chat bots. keep the data as long as we want them to kind of thing. So if we are having kind of memoryless mode kind of thing with our chat bots and saying, all right, every 30 days the chats are gonna be kind of reset and we're not keeping student data. So then it doesn't become that creepy AI tool kind of thing that you mentioned.
Elizabeth Riede: Yeah, I think that's really good. I think about the difference between like, how did they know that? And how did they know that? ⁓ Like, we talk about that a lot in hospitality. Like, you want to be good enough that they're like, how did they know that it was my birthday? Or how did they like, but you never want them to use the other tone of like, how did they know that it was my birthday?
Keith: Thank
Tom Van Gilder: Right?
Keith: You
Tom Van Gilder: Big data is kind of creepy.
Elizabeth Riede: It's all cute until it's not, you know? Like these are amazing anticipatory touch points until all of a sudden someone's like, wait a second. ⁓
Tom Van Gilder: I was reflecting back the other day about I was kind of doing some Amazon looking around kind of thing. then ⁓ the apps that I play on, suddenly those ads are in my app too. like, sort of used a browser that's not so cookie-fied kind of thing.
Elizabeth Riede: You Mm-hmm. ⁓ So thinking about that, ⁓ when you, or is there an example where you've intentionally chosen either not to automate or not to bring AI into a conversation or a process, even where you felt like, yeah, we could, ⁓ but maybe we shouldn't.
Keith: I think you were touching on this a little bit earlier that you may be approaching that line with a couple of asks.
Tom Van Gilder: Yeah, and that's actually just happened last week and I was meeting with one of the folks and then some of the other folks that they were having to kind of somewhat of a data steward of the data and they're like, hey, wait a minute, let's have a conversation more. So like, all right, let's maybe have some more conversations around kind of some set guardrails around AI and defining, like I said, the human part of this versus the AI part and where should they never meet.
Elizabeth Riede: Mm-hmm.
Tom Van Gilder: And if we are, I don't want AI to ever just make a decision for us and not really be thoughtful in some of these processes. So I think this kind of guardrails document, for lack of better words, is something we start leaning on a little bit more often as we're kind of going through these campus discussions and kind of thinking through what this actually looks like and almost like a contract of sorts. ⁓
Elizabeth Riede: Hmm.
Tom Van Gilder: to make sure that we're all on the same page with what we're trying to do here.
Elizabeth Riede: Yeah, I think that's really good. We get so excited that we just want to go a mile a minute. You see a potential use case or especially where it can save hours or labor or effort from an employee. And you're like, yes, I want to do that thing. But you don't, especially in operations, we don't always think about five years down the road, seven years down the road, Keith, to your point earlier, public records, who could ask for this? ⁓
Keith: In
Elizabeth Riede: We aren't always considering all those things. And so it's really good to have language that holds us accountable and holds us true to that.
Keith: And I think the importance of evaluating risk as a part of these potential opportunities and having an actual risk framework is vitally important. ⁓ You know, what gets lost in this, ⁓ Tom mentioned data stewards earlier, ⁓ and we have tons of them across our campus. Data governance is critical to any successful AI program. How good is the data? What's the quality of the data? But my biggest concern day in and day out is where is that data going? So setting up the proper risk frameworks for that as a part of the evaluation of these use cases I think is really important.
Elizabeth Riede: Yeah, something else that's come up in the work that Tom and I have been doing is data definitions and data dictionaries and making sure that you're really clearly defining that data. Because the operational definitions of that data may be very different ⁓ and may be seen very differently by hospitality professionals than by more
Keith: Mm-hmm. Right.
Elizabeth Riede: skilled and trained subject matter experts, especially when we get excited about something. ⁓ So to that point, I think from an operational perspective, for my fellow hospitalitarians out there, what are some really good questions that we should be asking ourselves about, whether that's AI, data, risk, ⁓ what are some things that we should be considering?
Keith: Thank ⁓ I'll go back to data. What data is going to be consumed? Where is that data going? But I've said this quite a bit is we're not business process owners. Hopefully we're helping enable efficiencies through AI or other technologies. But really evaluating each of your business processes just to think through is there an opportunity for better efficiency if we layer on AI? You know that I think that's where you know, we know you do you do great work and you're thinking through appropriate business cases about food waste and scheduling and and those are brilliant ideas that you know we in IT get really excited about because you as the executive owner business owner have thought through this already right and and it's an opportunity for us to have a collective win. So yeah. I think just really honing in on business processes and understanding and knowing your data.
Tom Van Gilder: And I would add to this too that ⁓ part of this contract document that I've started last week was kind of at the forefront of it is kind of starting with some core principles of equity. So AI is an assistant, but not a decision maker kind of mode of thinking. The do no harm ⁓ thought processes of system must be rigorously tested to ensure it doesn't systematically disadvantage applicants, et cetera. ⁓
Elizabeth Riede: Hmm.
Tom Van Gilder: AI should be transparent, applicants should be informed that AI is in use to assist in a review process. And then start thinking through how do we ⁓ do bias mitigation and ⁓ the data and algorithmic guardrails kind of thing. So preventing the AI from learning or perpetuating historical biases, for example. ⁓ Data blinding, what the AI should not ever see, ⁓ it's kind of...
Elizabeth Riede: Mm-hmm.
Tom Van Gilder: one of the thoughts on this contract as well. ⁓ Things like a person's name could significantly disadvantage somebody. It's proven already with Google searches that. And then as I mentioned, the human in the loop architecture at the forefront is defining what the AI is going to do and what the humans are going to be doing and ⁓ keeping that at the top of mind.
Elizabeth Riede: Mm-hmm. Yeah, we talk about it in hospitality a lot. At the core of efficiency is role clarity. And so I think understanding where is the AI's role and defining that upfront and not allowing, we wouldn't allow scope creep from one of our employees. We're certainly not going to allow it from an AI perspective. So I think those are really, really good things for us to think about. One of the other things that you touched on was that AI literacy. And I think we're talking a little bit about it. now, but to kind of go a layer deeper, what is something that you would want every student, every employee working on campus to understand about AI?
Keith: I guess I'll kick this one. I think I know I'm sounding very redundant, but it's knowing what the standardized tools are and the level of sensitivity of your data. ⁓ using non-supported, non-standardized tools puts you and the organization at risk. So we've picked the tools that we have for reasons. But most importantly is to protect
Elizabeth Riede: Hmm.
Keith: our organization and the individual users. So that's probably A number one. Number two is just the ethical use of it. You Tom's talked about it. And that's why literacy has been so important. And we run a lot of great workshops across campus and we have a great website that we stood up ⁓ that we're really proud of. So I think we're communicating well, but how do you get to the masses with this?
Elizabeth Riede: it was so interesting. We did ⁓ an all staff So it was ⁓ all layers of our operation from frontline first employees to managers to frontline employees that been here for 25 years. And ⁓ so you can imagine kind of the spectrum of ⁓ literacy that was in that room. ⁓ just a wide variety of perspectives. And ⁓ our trainer pulled up Notebook LM. And ⁓ gone through ⁓ series of points around AI already. And they pulled up Notebook LM. And you could see everyone in the room. It had been 25 minutes. And ⁓ were in various stages of engagement. ⁓ And then he put in our Surf Safe training. materials into Notebook LM. And he had it generate flashcards for this training and, you know, Surf Safe is meticulous and it's also tedious. ⁓ Some of staff would even say boring, pointless. ⁓ And so. But he, so he had to generate these flashcards and immediately everyone in the room sat up, eyes forward, they were engaged, they were interested because all of a sudden they could see themselves in the technology. And I think you're getting at the heart of, I think what makes AI literacy so important is that people can, they can only be what they can see. And I think a lot, a lot of what we struggle with in hospitality is that people can't see themselves in this technology. And so I really like that point about making it relevant and making it be a place where people can feel safe to engage.
Keith: Right. And we're trying to meet everybody where they are. Tom and I, ⁓ we've been on a campaign to visit with the leadership across campus and ⁓ those are demos largely. And I think to your point, it's more demonstration of the power and the functionality of these tools. ⁓ And it gives AI some context and some folks. think the part that gets missed in this, think too sometimes is the level of change, right? You talked about the variety of within your staff, the pace of change to one person is going to not equal the pace of change for somebody else. And I don't think it's ever been more evident than with AI. So we got to be mindful of it and respectful of it, but it is certainly a challenge.
Elizabeth Riede: Mm-hmm. I think looking more broadly, when we think about those populations and kind of the different places that people come to us from and the experiences they come to us with, to your point, it's not one size fits all. Are there populations that either benefit first or kind of gravitate to these tools more readily? And then are you seeing populations that are at risk of being, you talked a little bit, Tom, about ⁓ equity in AI, you see populations that are at risk of being disadvantaged through this process?
Tom Van Gilder: I do worry about the AI being the new digital divide kind of thing where I'm a student and I'm making, you know, poor grades and I said, but I can afford ChatGPT Pro at 20 bucks a month kind of thing. And I'm going to fund that. And suddenly I'm an A student because I'm starting to use this. And then the student that can't afford this same tool and they just, they're going to stay the C student because they are, they don't have the same tools or whatever.
Elizabeth Riede: Mm-hmm.
Tom Van Gilder: then that's just not really fair, is it? So I worry about this as a human. mean, we've got to figure out as an institution, how do we not make this the new digital divide?
Elizabeth Riede: Yeah. Yeah. I think in hospitality, already see we're managing multi-generational workforce. We see people come to the table with different skill sets, obviously, and different experiences. And our supervisors have to manage that. And so I think that's kind of validating to here, but it's also concerning, right? Because I think we see that in our own operation. You see
Keith: real.
Elizabeth Riede: the managers that can engage readily with these models and can adopt the technology faster, can get a handle on the pace of change a little faster. They tend to rise to the top even if the capability of another person might be just as high or even higher long-term.
Keith: Yeah.
Elizabeth Riede: When we think about that pace of change, what's a skill in either campus operations or technology in higher ed that mattered 10 years ago, ⁓ matters less now?
Keith: Okay. I don't know. keep going back to Tom's, know, human in the loop. think there's, I don't think the human is going to be ⁓ irrelevant in any of these processes. so I'm not thinking anything of anything else at the top of my head, Tom. I got a little bit.
Tom Van Gilder: But as we come back to like the, like the, we used to have registrar group, you know, would do manual processes of degree auditing kind of thing. that was 10 years ago. Now that we have the, some AI tools that can help them, you know, and help our students do the degree auditing, kind of AI driven degree pilot of sorts and perform 99 % of those kind of same tasks within minutes, instead of a human having to do that.
Elizabeth Riede: Mm-hmm.
Tom Van Gilder: laborious amount of hours to do the advising part. So I do think that's high up there. ⁓ Same with financial aid and kind of looking at a student's financial aid ⁓ request. And we used to have data entry people that were speed typist kind of thing. that was a data entry position. And those are becoming maybe ⁓ somewhat less important of speed typing kind of thing. ⁓ But again, I think my message is AI should be co-pilot, not the pilot kind of thing. let's just make sure that we are, that's framing our thinking moving forward. And Ollie, I saw some kind of video reel that someone sent me the other day that was like, AI is still not getting 96 % of the jobs correct anyway. let's. fearful about AI taking all the jobs away, but let's figure out how AI can augment the jobs that we do have. Maybe that's a better way to frame it.
Elizabeth Riede: Yeah. Yeah, no, I think we talk a lot about, again, there are just moments that shouldn't be given away. There are moments that should always be human. But then there are some that I think everyone would be happy to give away, like folding laundry or cleaning the bathtub, right? ⁓ are some that it's like, yes, please take that from me.
Keith: Yeah.
Tom Van Gilder: Yeah, exactly.
Elizabeth Riede: my Christmas tree of laundry, I will gladly give that away. So ⁓ your perspective ⁓ people who eat and people who experience hospitality and service out in the world, ⁓ also as people who ⁓ thinking about the ⁓ possibilities, is there a place where you see AI can improve either hospitality service, customer service, the food experience. ⁓
Keith: Tom, I know you've been working closely with Elizabeth. What comes to mind?
Tom Van Gilder: Yeah, I'm excited about the possibilities of like food waste because that really gets on my nerves. I do worry about the amount of food waste, especially when we've got folks in our own community, have food insecurity and ⁓ how can we balance that? ⁓ Can we use AI for ⁓ the good kind of things and thinking all right, ⁓ instead of
Elizabeth Riede: Mm-hmm.
Tom Van Gilder: this stuff, all this stuff going to waste. And I know you've had a lot of, you know, I've had some conversation around this, but yeah, AI for good is what I would like to see happen in that front.
Keith: Yeah, think labor waste comes to mind too, right Tom? I know it wasn't, but you're thinking about that. think inventory management and inventory management systems with all of that data, the logistics I think are going to be impacted by AI. ⁓ I just never want to see the customer service component taken out, right? think, yeah, that's, ⁓ yeah, I think people still do expect to have an experience. And if they're talking to a machine,
Elizabeth Riede: Mm-hmm. Yeah, yeah.
Keith: A lot of that is lost.
Tom Van Gilder: Yeah, just because we can doesn't mean we should kind of thing. And I'll say not to not Bojangles, but I will. Yeah. I go from the drive through in five minutes, arguing with the dang food ordering system. That doesn't seem like good AI to me.
Keith: Right.
Elizabeth Riede: Yeah. ⁓ Bo-Linda, yeah. She's great, isn't she?
Keith: Yeah, Tom, I got to tell you a funny one. I keep saying thank you to the Automated Assistant.
Tom Van Gilder: You
Elizabeth Riede: Well, when there are robot overlords, they'll remember Keith was really nice to me.
Keith: Hahaha!
Tom Van Gilder: And Tom not so much.
Elizabeth Riede: Yes, I always say please and always say thank you, just in case. ⁓
Keith: We're way too southern, I think.
Tom Van Gilder: think maybe I'm the old man of the group of curmudgeon and just like I'm the tap tapping. I was like, human, give me the human.
Elizabeth Riede: Stay off my lawn. When you think about your hospitality experiences, whether they're on campus or elsewhere, you just mentioned Belinda. So that is a red flag for sure. But is there an example where you can think of that AI has added value already?
Keith: Mm, mm, mm.
Tom Van Gilder: Yeah, I mean, went to, I guess I'm trying to think of it was like one of the Hilton change or something. But anyway, it kind of remembered that I liked this, the same temperature at night and then turned down to the nice 68 degrees for me and not me coming into the room and trying to click through the wall thermostat there. So I kind of liked that. I didn't remember that I'm a 68 degree kind of guy. So I thought that was kind of cool.
Elizabeth Riede: Yeah, ⁓ that's a how did it know kind of moment.
Tom Van Gilder: Yeah, this is a little creepy, but I kind of like that it remembered that I'm a 68 degree kind of guy.
Keith: I think the personalization of the experience, think, you know, kind of that intuitive piece of it, I think is going to be helpful.
Elizabeth Riede: awesome. What excites you most about AI's potential in higher education?
Keith: I think it's going to transform teaching and learning. think the ⁓ employers are going to expect for you to be AI literate. So I think it's incumbent on us to make sure that our students are ready to hit the workforce. ⁓ But I'm excited for the efficiencies. Tom's mentioned several opportunities that we've already created very significant efficiencies that's saving the organization money. And I think
Elizabeth Riede: Hmm.
Keith: you know, that's not getting rid of people. It's just making them more efficient in the jobs that they're in. And I think that's, that's kind of the lens that Tom and I look at this from is, you know, how can we enable, enhance, improve their operations and make sure that they can focus on the right things instead of the mundane things, right?
Tom Van Gilder: Yeah, I kind of like where AI can be used as strategic kind of ⁓ more of an AI as a liberator kind of thing. So liberates me to do higher level kind of things over here. So if it's faculty and ⁓ spending four hours a night doing all this grading and it's all the same kind of grading that I know it, throwing it into AI assistant that could help me do that grading and get it down to ⁓ minutes instead of four hours, then let's...
Elizabeth Riede: Mm.
Tom Van Gilder: think through that. But on the same coin is like, then how do we teach our students to use AI effectively and not just as a way to kind of get through all the content and stop doing the higher level thinking themselves. There's a lot to this coin here.
Elizabeth Riede: Yeah, it's all a very double-edged sword. Is there anything else that you would want listeners to walk away with? I it's ⁓ such a good conversation because I think ⁓ much of our time we focus in hospitality at least ⁓
Tom Van Gilder: very much so.
Elizabeth Riede: the tactical, like, how do I make this process work? How do I do this thing? so think it's really important that we pull ⁓ other folks in to kind of challenge, what we be thinking about? What should we be considering that maybe we're not considering? What are some ⁓ potential risks that we run to help us future-proof our strategies and our operations. So is there anything that you would want our listeners to walk away with or be thinking about?
Keith: Tommy, you wanna kick it out?
Tom Van Gilder: Yeah, I mean, I guess I'm the IT guy who's maybe the very humanistic IT guy here. But I do think that, like I said, where are these areas that AI could be used to release the kind of from the routine work, but ⁓ which there's then double downing on how we act and as what are the things that AI definitely still cannot do, like moral discernment, deciding what's fair and what's not fair kind of thing. empathy, things like that, inspiration and motivation. ⁓ the human side of this, I ⁓ the AI could be critically important to help us get more efficient with the work at hand, but how do we double down then on things like accountability and efficiencies ⁓ doing it the right human way?
Keith: I would say do your research. ⁓ think there's folks that we talked about ⁓ terms of pace of change that are resistant to any new technology, whether it's ⁓ or anything else. And there's a lot of noise ⁓ the ether about ⁓ bad it is or how good it is. ⁓ so I would really encourage everybody to do their own research. ⁓ advantage of, ⁓ our organization, ⁓ any professional development opportunities to learn about it. and be open-minded about it. And be open-minded even to the point of, I might be able to utilize this in this particular area, and it's not as scary as I thought it was. Yeah, think Tom and I deal with both ends of the spectrum, right? The ones that are highly resistant and never will want to introduce it to their organization to how fast can we go and how much can we get done in the next 20 minutes. you know, that, but we're really thinking about the folks in the middle. and how can we really support them? So my big ask would be do your own personal research, come to your own conclusions about the use of AI, ethical use of AI, the sustainability ⁓ components of AI. I think all of those are ⁓ well-researched areas, but if you're just getting your notes from social media, it's probably biased, right? every organization's gonna have their own kind of tenor around AI. So I think understanding your organizational culture and willingness to change, think, is ⁓ something to be understood as well.
Elizabeth Riede: Yeah, I think we talk a lot about the, I guess all the balls that you have in the air ⁓ when you're running any business, but especially in hospitality, you you have the employees opinions, you have the guests opinions, you have your clients opinions or your organization's opinions. And it's juggling all of those through the lens of whatever decision you're trying to make. So I can definitely appreciate that not every organization is going to be Let's see what we can innovate. Let's see what we can do today. And I think it's important that even, but your customers might be. ⁓ Why haven't you done the AI ⁓ voice at the drive-through already? ⁓ And then Tom's here to tell them why.
Keith: Yeah, exactly. And he's not going to tell them thank you either.
Tom Van Gilder: It would just work.
Elizabeth Riede: Well, Keith, Tom, thank you so much for joining me in this conversation today. I really appreciate your perspective and I appreciate your willingness to come and share your role, your responsibility and a really valuable perspective for us in hospitality. So thank you.
Keith: Thanks, Elizabeth. Thanks for having us.
Tom Van Gilder: Thanks for having us.
Keith: Tom, you want to go first?
Tom Van Gilder: Yeah, I'm happy to. So I was reflecting back on my days of ⁓ director of IT support. And maybe at the time when I took over that area was maybe not known as awesome customer service. I found this really cheap. Well, know, times have changed, right? But I found this really cheesy video of this guy who owned a restaurant and the name of the video was given the pickle. I...
Elizabeth Riede: Say it isn't so. or to release people to do the work that only humans can do? If you can answer that honestly, you're already ahead. Thanks for listening and remember, on a campus, as in life, the fastest path to trust is still transparency. This has been Human at the Counter. I'm Elizabeth Reedy, and remember, being present is everything. If you work in hospitality, the kind that lives in details, you know service is shaped by systems as much as smiles. ⁓ schedule, the point of sale, the inventory order, the follow-up email, and increasingly, those details are being shaped by artificial intelligence. Today, we're looking at AI through the lens of a university campus, but whether you're running a dining hall, a hospital kitchen, a multi-unit restaurant group,
Keith: Thank you.
Tom Van Gilder: kind of made my staff and now I'm the give them the pickle guy hence forward. 20 years I've been the give them the pickle guy. But, you know, it's just a message about strong customer services like critically important to the greater good for our campus. So I think that's my one unwritten rule is like just give them the pickle. Just, just, yeah.
Elizabeth Riede: You
Keith: .
Elizabeth Riede: or overseeing student services, the leadership tension is the same. How do we use technology to strengthen service without letting it quietly erode trust? I'm Elizabeth Reedy. I've led teams in hospitals, coffee shops, and now higher education dining. And here's what I've learned. Systems scale efficiency. Culture scales humanity. And when those two fall out of alignment, people feel it.
Keith: Good. Good.
Elizabeth Riede: I love that. Yeah, that's an old Bob Feralism. Yep. Yep. Yep.
Tom Van Gilder: That's it Bob Fairley, you know it, you know it.
Keith: All right. Tom's already overshadowing me. ⁓ I think my best piece of advice is listen. know, I like Tom kind of harkened back over my career, doing everything from business development all the way up through practice leadership and into my current role. Being able to listen and actively listen, I think is so critically important to leading and ⁓ more important to customer service.
Tom Van Gilder: Hey
Elizabeth Riede: I'm gonna ⁓ and fast. Culture decides whether technology feels like service or surveillance. Higher education isn't at an inflection point. Artificial intelligence isn't a pilot project anymore. It's embedding in advising platforms. It's embedded in learning tools. It's embedded in dining forecasts and help desk tickets and accessibility software. So today we're not asking, should campuses use AI? Yeah, in one of my previous episodes, we talked about a lot of people listen to answer. But if you're actively listening and listening to engage, it really takes your leadership to a whole other level. So I agree with both of those things. All right, name one thing that a human can do that no AI or robot could ever truly replicate.
Keith: Absolutely.
Elizabeth Riede: We're asking something harder.
Keith: I think a lot about a lot of the trades, right? And I know that's probably everybody's go-to, but you know, I've got two boys in their twenties and you know, I've kind of steered them more into trade related opportunities because I think, you know, I don't think there's any, well, there may be a few exceptions, but for the most part, I don't think that's going to be taken over by AI.
Elizabeth Riede: Who decides how it's used? Who protects students when it's wrong? And how do we build systems that scale without shrinking trust? Yeah, yeah. What about you, Tom?
Tom Van Gilder: Yeah, I think it's for me, it's around the, don't think a bot's ever gonna really help us decide what's fair and what's not fair. It's like, are really intended to more about what's efficient and not efficient versus where the human's really, you know, how to have that moral sense of fairness kind of stuff. That's me in my core.
Elizabeth Riede: Mmm. To unpack that, I'm joined by two leaders who sit at the center of these questions every single day. Keith Warner, Chief Information Officer at Appalachian State University, and Tom VanGilder, Deputy CIO. Leaders responsible not just for uptime in cybersecurity, but for policy, governance, and the guardrails that determine how innovation shows up in students' lived experiences. This isn't a tech episode. This is a leadership episode. I think those are both such awesome answers. think it's so funny when I talk to hospitality folks, it's usually, the answers are usually something like really truly ever love or really truly smile. And so I think there's a concreteness to your answers that I really appreciate because I think you're right. ⁓
Keith: We are IT folks, Elizabeth. just ultra pragmatic, I think. ⁓
Elizabeth Riede: So you don't love or smile. That's what you're telling me. So let's get into it. Before we jump into frameworks and policy language, I want to zoom out a little bit. Imagine it's registration week. A first generation student logs into a portal at 1147 PM. They're trying to figure out financial aid, housing, and whether they can afford a meal plan. The system suggests courses, a chat bot offers help, a predictive model flags that they may be at risk of dropping below full-time status. No, I can appreciate that. All right, if you walk into your favorite cafe, diner or restaurant right now, what is your go-to order?
Tom Van Gilder: funny. I like a, the ACP shrimp wrap at the taco truck. can't, I'm losing the name of the taco truck, but, ⁓ it's just like an awesome, my favorite go-to dish with shrimp and veggies and some rice and cheese. It's like, it's got all the goodness. It go wrong. That's it.
Elizabeth Riede: On paper, this is proactive support. Behind the scenes, models are running probability scores, systems are prioritizing outreach, alerts are firing. Can't go wrong. It's got rice and cheese in it. Can't go wrong.
Keith: I'm, I'm eating potatoes. mean, give me a good prime rib, good steak. I am, I am good. ⁓ but you know, if I went to any random, ⁓ kind of diner, I think it'd probably be a hamburger. You know, I'm pretty, I'm pretty simple.
Elizabeth Riede: There you go. There you go. All right, go along with your hamburger. If you were a condiment, which one would you be and why?
Keith: I gotta go with ketchup. I mean, it's ubiquitous, right? I mean, it goes with everything. So I've gotta go with ketchup. And Tom, if you say anything different, I'm gonna be shocked.
Elizabeth Riede: Thousands of signals produced in seconds to prevent friction before it becomes failure. But in the student's world, it's either, wow, this university anticipated my needs, or how did they know that about me? That line between helpful and unsettling ⁓ is razor thin. Love it.
Tom Van Gilder: ⁓ My wife has turned me on to the chili crunch stuff. So even though it can sometimes cause me some indigestion at the early
Elizabeth Riede: ⁓ yeah? My listeners will not see this, but yeah. A little flyby thing. Yeah.
Tom Van Gilder: ⁓ you got it. got it. Yeah. So we have found our favorite. We've tried like four or five different chili crunches and like, nope, this one that we have to special order. But it is best.
Elizabeth Riede: And here's where it gets real for operators like us. If dining uses AI to project traffic and reallocate labor, if advising uses AI to nudge enrollment decisions, if IT uses sentiment analysis to escalate tickets before complaints trend, who defines what's fair? Who audits the unintended consequences? Who makes sure that efficiency doesn't quietly outrank equity? Of course, it's always the one you have to special order. That's always the best one.
Keith: I'm gonna have to try that. What are you putting it on, Tom?
Tom Van Gilder: We put it on eggs and we eat a of them. It's got a little kick to it, but it's really good. I'll bring you one of them. Because I think that to get a special discount price, we had to order like 10 of them. So I think I've got plenty to share. ⁓
Elizabeth Riede: Pizza, noodles.
Keith: Really? Okay. It's the full, bursting. ⁓
Elizabeth Riede: Because optimization is not a neutral act. It reflects values. It reflects what's important. And that's why CIO leadership right now isn't about servers and storage, or isn't just about those things. It's about stewardship. So Keith, Tom, let's talk about how you think about that stewardship. What does responsible AI leadership actually look like on a campus? But first,
Tom Van Gilder: You
Elizabeth Riede: Alright, if you could have an unlimited lifetime supply of one snack, what would it be? Besides chili crunch, Tom.
Tom Van Gilder: Thanks.
Keith: .
Tom Van Gilder: I'm a potato chip kind of guy. I could go with my rest of my life without sweets, but potato chips are my jam.
Keith: I try to stay away from the sweets, but if I had to have one go to, it's Reese's peanut butter cups.
Elizabeth Riede: That's a good, peanut butter is a perfect food. I firmly believe that there are two perfect foods in the world. Peanut butter is one and yogurt is the other. And peanut butter and yogurt together? All over it. Oh man. It's time for the Humanity Check Questionnaire.
Tom Van Gilder: Okay.
Keith: It is. Okay. Now I'm one out of two on those.
Elizabeth Riede: All right, if a robot took over one chore in your life forever, what would it be?
Keith: I think for me it'd be laundry. That's the one thing I least look forward to every week, I think.
Elizabeth Riede: That is a shared trait, my friend.
Tom Van Gilder: Yeah. For me, it's cleaning the bathtub because I just can't stand it. And I do like like occasional hot baths. So I use it quite often, but just getting in there and cleaning the darn thing is just no bueno.
Elizabeth Riede: It's not set up to be clean. You would think because it has the built-in drain that it would be, but it is not set up to be clean. You have to get in it and then you have the soap and then you have to not get it on your feet. Yeah, it's hard. All right, you're stranded on a desert island, but somehow there's a working food truck. What is it serving?
Tom Van Gilder: mean, I'll go back to that shrimp quesadilla. hope that, ⁓ yeah, hope they're okay. ⁓
Elizabeth Riede: Yeah, it's got to be there.
Keith: Probably fish tacos for me.
Elizabeth Riede: Yeah, it's good. You probably have some fresh fish around too, so that's probably good.
Keith: That's a point. I didn't think about it at that angle.
Elizabeth Riede: Alright, you could have lunch with any fictional character. Who would it be?
Keith: I guess it's my turn. This is a tough one for me. Probably ⁓ the Kevin Costner character on Field of Dreams. ⁓ That was like a, that was a magical movie for me. You know, the father son connection and I'm a big baseball guy. So ⁓ that would be mine.
Elizabeth Riede: Good one. Nice.
Tom Van Gilder: I grew up on the Huck Finn and Tom Sawyer books. so getting on the ⁓ homemade raft on the river sounds pretty amazing to me as a young kid. that would.
Elizabeth Riede: Tom, you are the second guest on this podcast that has picked that. That's great. I'm gonna have to make that happen somehow. I bet AI could make that.
Tom Van Gilder: Is that right? I'm sure, but that's one thing I don't want AI to do. I want to go do that raft myself. I'm getting too old for it now.
Keith: Mm-hmm.
Elizabeth Riede: Nah, never. Mountains keep you young. If your life had a theme song that played every time you walked into a room, what would it be?
Tom Van Gilder: I guess that's me. I guess I'm a Queen guy. So Bohemian Rhapsody is probably my go-to. just, Harkin's days of rocking that song with my kids and now I have grandkids and I'm turning them onto it as well. So like, let's do the mama mia, mama mia stuff. And they're rocking it with me. that's, yeah, they do.
Elizabeth Riede: Nice.
Keith: These answers show our age too, don't they Tom? Mine's got to be Mr. Brightside.
Elizabeth Riede: Love it. Not at all. Nice! Y'all picked two great karaoke songs. I appreciate that.
Keith: which you will never hear me do.
Elizabeth Riede: ⁓ come on now. I didn't tell you that's the end of this podcast.
Keith: I think I have a connection problem.
Elizabeth Riede: Now I know somebody in IT who could fix that for you.
Keith: Yeah, so do I, but I don't know if I'll call him with it.
Tom Van Gilder: Ha ha.
Elizabeth Riede: All right. Well, thank you guys. You passed the Humanity Check. ⁓ I think this episode is really all about the intersection of how technology and governance of technology supports the hospitality experience. I think we were talking before we started about the fact that ⁓ increasingly technology is becoming a part of our hospitality world and in ways that we didn't expect and in ways that we can plan for. ways that we didn't anticipate. And so think I'd like to start out by talking a little bit about your roles ⁓ how the government framework shape not only the work that you do, but the way that our hospitality work is shaped by those roles. ⁓ I'd love to hear a little bit from your perspective ⁓ about your roles and the work that you do.
Keith: Sure, and I'm happy to start Tom. So I serve as the Chief Information Officer here for this organization and Tom is my Chief Deputy and Chief Operating Officer for the division, which basically means if you encapsulated it is that we work ⁓ in partnership ⁓ and artificial intelligence is probably top of the list for both of us. You know, I think as we kind of launched into this collectively, it was really about How do we set up a governance framework that makes sense to drive our organization forward ⁓ in an ethical way, a responsible way, a secure way, and a sustainable way, or as much as we can from a sustainability perspective? ⁓ So we initiated a task force that was put together nearing two years ago now. ⁓ Tom had the really bright idea of having ⁓ working groups or what we call them enablement teams that are ⁓ kind of feeder working groups into the task force. And Tom can go into more detail about those, but it is, that has been a ⁓ boost for how we manage our intake of projects and use cases, policy considerations, sustainability, as I mentioned before, ⁓ and thinking about that organizationally and bringing the right stakeholders together. to think through it with us and hospitality being a large part of that. So we value those partnerships because Tom and I aren't experts on all the business processes across our organization, nor should we be expected to. So those partnerships are critical to us thinking through potential use cases and how can we get the most value from AI.
Tom Van Gilder: Yeah, so the AI enablement teams that Keith mentioned, it's which I kind of put a call out to campus and just said, hey, you we've got these various working groups that we know are important to us. ⁓ Everything from AI for teaching and learning, for example, AI for research. Now that we're a research to designated university, this made a lot of sense that we had start focusing on AI for research and. And we've start thinking through like AI literacy for our campus. And so we've got another working group. And I'm smart enough to know that I need some excellent leaders along the way to help me with all these conversations. So for like the teaching and learning, we talked with this fellow from College of Ed, ⁓ the faculty member helping us lead some really rich AI discussions around teaching and learning. ⁓ There's some faculty consternation around ⁓ AI teaching and learning, as you might imagine. across our campus, which is not uncommon across other universities at this point in time in our lives as well. But what does that look like for our campus and what are the things that we should be bringing forward and or not bringing forward in the teaching learning? So we've had good, rich conversations around things like using AI detection tools ⁓ in teaching and learning. Should they or should they not be? And I think our stance is that if they're maybe 92 % accurate is kind of the standard rate right now for some of these detection tools, for example. So if they're 92 % accurate, that means 8 % of our students, we're going to disenfranchise from the academic mission of our university and call them cheaters for lack of better words. I don't think so. And the faculty are the same thinking as well. It's like, we don't want garbage. So once it gets 100 % accurate, come talk to me kind of thing.
Elizabeth Riede: Yeah.
Tom Van Gilder: Anyway, so these working groups are working really well. We've got a good partnership with our Office of Research folks who are helping us lead that research team. our academic technologies, ⁓ academic and emerging technologies. just changed the name. I'm still getting used to it. ⁓ So academic and emerging which is a thoughtful name change, actually, because we're starting to think through all these things that are coming at us ⁓ like AI. And so we need a ⁓ ⁓ campus as to be thinking through. what this means for us as a campus.
Keith: That's a really good point too, Tom, about the literacy. I think that was one of our early understandings is that foundationally we needed to provide workshops and training and professional development opportunities to learn about AI. So we had some common understandings and I think we've done that really well actually. And I think that's really benefited us as we continue through this journey. The other thing we did was standardize on a few tools that we that we could kind of own and control the flow of data. So I think those were a couple of the really important moves that we made early on that has really helped us.
Elizabeth Riede: think you're touching on a couple of things there that I'd like to dig into a little bit. The first is around trust and the importance of trust. think in hospitality, talk about, ⁓ need trust with our customers, we need trust with our employees, and we need our employees to trust our customers, and we need our customers to trust our employees. ⁓ And I see in technology, you were talking a little bit about with the detection tools that
Keith: Hmm.
Elizabeth Riede: there's this possibility for the introduction of mistrust or distrust, right? Or maybe that there's some reservations, understandably, ⁓ about whether they should trust this new technology. ⁓ And then the other thing that you guys were talking about is literacy. ⁓ And so I think as you're thinking about where you're leading and advising on whether it's AI policy or the tools that you're using, What are some of the guardrails that you're putting in place ⁓ when you think about both opportunities, but also the potential for some of that either mistrust, distrust, misapplication, et cetera.
Tom Van Gilder: Got a great story too, and it literally just happened today. Our budget officer is taking a class and the faculty member just like got a little bit PO'd because he can tell that they're all using AI for all the problems that he's putting out for homework. And he's like, damn it, I really want you all to learn and you're not learning if you're just throwing in the AI slot back at me and that's not what this assignment's all about. So a really interesting timing on her part to... message Keith and myself about this, like, man, this is like spot on, we need to be having some talking about, you know, what it means for like ⁓ assignments and assessment in the day of AI kind of thing. So I think this is where that literacy piece is really critically important and how to get our faculty, they're still going to do that. I mean, it's so maybe it's figuring out how to best utilize the tools that are afforded to folks and how to, you know, like I said, change your assignments and assessments.
Keith: Thank you.
Tom Van Gilder: way as we figuring out these new tools that are afforded to us.
Keith: And trust, think, also carries forward just with staff as well as faculty. So, you know, we have a lot of trust with our staff with sensitive data or confidential data. So, you know, part of having standardized tools is we can control it and we're not feeding, you know, what they call larger language models, which are public models with our very important data. you know, there's a...
Elizabeth Riede: Mm-hmm.
Keith: A lot of trust that goes into that. We're investigating tools that, you know, kind of center that, but you know, if a staff or faculty member brings their own AI tool of choice and feeds those models, we're losing sight of that. So we're looking at tools that we have line of sight to that data and making sure that we know where the flow of the data is and where it's going and who's doing it. But it's complicated and oftentimes expensive to get these tools.
Elizabeth Riede: Yeah.
Keith: So it's definitely been a challenge. think guardrails are going to continue to morph over the next few years. This is such an evolving area that I think we're going to see better guardrails to Tom's point on the faculty side. And I think we're going to see better guardrails in terms of which models or platforms that you can use. The other challenge that we have now is every vendor that we deal with has some AI functionality associated.
Elizabeth Riede: Mm-hmm. Mm-hmm.
Keith: So, know, A number one is how do we assess that functionality and how good it is, but the other is how safe is it? So there's, it's a really good question with not a lot of really good solid answers yet, unfortunately.
Tom Van Gilder: I think some of it too is us kind of thinking through and formalizing some of our processes. So what I did mention was, is we also have an AI technical team, which I think you're very familiar with, Elizabeth, kind of helping our campus with figuring out campus processes and can we do more efficient things, yada, yada, yada. But, you know, we're starting to get asked questions like, hey, could I use this for X, Y, and Z? ⁓ And like, yes, but, know, let's think about maybe putting some guardrails up, as Keith mentioned, and I call it the human in the loop kind of guardrails that maybe almost developing a contract of sorts saying, all right, we'll develop this, but we've got to have some human processes on this so that we're not just giving over to the AI to make the decisions for us kind of thing, but having some real human skin in the game, if you will, for the...
Elizabeth Riede: Mmm. We talk about that a lot in hospitality too. think there's always a touch point that should stay human. ⁓ know, even if you've automated away these places where there's friction in whatever journey, if you're mobile ordering a sandwich or if you're placing a grocery order that's being delivered to your house, at some point in that process, there should be a human. that understands the experience that you either are currently having or have had, ⁓ so that they're staying connected to that in some way. And so do you have specific kind of like no-go moments where you would say this should never be automated or this bot agent process algorithm should always hand off to a human? if it involves X, Y, or Z.
Keith: It's a good question, Tom. I don't think we've run into any particular cases yet, but that could change literally tomorrow. Tom, I know you work more closely with the technical team than I do. Have we run across any?
Tom Van Gilder: I'd say we're not at a full stop on a couple of the projects, but definitely starting to, hey, let's talk about this a little more and really think through how we do weave in the human processes in addition to the things that are being developed.
Elizabeth Riede: Well, I do want to not gloss over. I I think the example you used earlier, Tom, was a good one around the academic integrity and the code of conduct. I mean, I think when you have 8 % of your students that could potentially be misclassified, I mean, I do think that that's a moment that it's like, hey, we should probably have a human take a second look at this and make sure that that really passes the sniff test. If you have someone who is obviously engaged with integrity, and now is being flagged for some reason, you maybe we don't gloss over that. Maybe we don't take that 8%. So I do think that that's a moment that really resonated with me. I'm sure that there are others that will emerge ⁓ as we're going through these conversations.
Keith: think on the business side, I think about hiring, you know, because there could be inherent bias in some of these AI applications. ⁓ So especially if you're using AI to potentially weed out, you know, first run of candidates, I mean, that could be a fear. ⁓ So I think there's going to be several of those types of use cases that emerge, especially at our organization, as broad as it is. ⁓ But that's one that could potentially ⁓ be an issue.
Elizabeth Riede: Mm. Hmm. Yeah. Mm-hmm. Zooming back out a little bit and thinking about AI across campus or AI in education applications, where do you see AI kind of already outside of maybe unethically doing some classwork for students? Where do you see AI transforming that experience or engage or impacting that experience right now?
Tom Van Gilder: say like, ⁓ know, we, Keith and I were on a recent vendor demo presentation about ⁓ like AI early warning system kind of thing for our students and helping with student success kind of initiative. So ⁓ a student missing an LMS login or ⁓ their quiz scores are declining or whatever, ⁓ or even, you know, financial aid gaps, things like that to kind of.
Elizabeth Riede: Mmm.
Tom Van Gilder: nudge the advisors to also nudge students, especially if they're categorized in the high risk ⁓ kind of sense. So I think that's pretty interesting. ⁓ They also offered some kind of degree mapping, ⁓ helping students visualize how changing a major and or failing a class, et cetera, affects their graduation date in real time kind of thing.
Elizabeth Riede: Hmm, I love that application. think because that can be almost like speaking another language for some of our students. ⁓ I think that's a really, really good use case.
Keith: think, ⁓ know, I always kind of think about it from a day in the life perspective ⁓ a student, right? And I think advising comes to mind. I think there's plenty of opportunities around advising and making sure that ⁓ have the appropriate resources for our students. ⁓ if we ⁓ that and do that really well, it improves our retention rates as a campus. ⁓ so many opportunities, I think, with students, you know, access to clubs, you comes to mind. ⁓ How do they get better integrated into the social framework of our campus? I think there's plenty of opportunities to get more engagement with our students, which I think is the cornerstone of what we're here to do. It's not just to teach, it's for an experience. So I think there's plenty of opportunities. I think Tom mentioned some really good ones. Tom, I know you're really ⁓ leading up a lot of our accessibility efforts too. But I think that's an area as well,
Tom Van Gilder: Absolutely. So there's AI tools now that kind of automatically generate like high quality alt text for images kind of things. So we're looking at these kind of tools as well to help our campus. But really benefits our, can benefit our students too, especially if they're maybe identified as learning, having learning difficulties. So there's some really nice tools that kind of help our students that are having some learning challenges to, you know. them with their education.
Elizabeth Riede: I think that's a really interesting application because we sometimes think about, ⁓ or hear the conversation with AI, know, Keith, you mentioned with the hiring platform, this idea that AI kind of has some of the inherent or can have some of the inherent biases that exist in how it was trained. And, but I also think there's the really, really cool phenomenon that can happen where it can actually kind of challenge what's possible.
Keith: cam.
Elizabeth Riede: in a way that a human wouldn't. And I was having a conversation with one of my other guests about it just kind of broadens out the sandbox and then it becomes the human's job to kind of rein it back in and go, well, that's crazy. I'm definitely not going to do that. But I didn't even think about the fact that it could do X, Y, or Z or that that would be possible in our operation. So I think that that's a really, really cool kind of untapped potential or maybe newly tapped potential for AI. Advising is another one or kind of those early warning signs. I I think that's a conversation that we have a lot on the campus dining side is like, what is our role when we have residents who are eating with us? And we have so many touch points in dining with our students that are genuine and authentic touch points and not classroom environments. ⁓ that can provide those kind of early warning. And so I think that's another place from a hospitality perspective that AI could really provide some links and insights for us.
Keith: That's an excellent point. I haven't thought about that one. That's really good.
Elizabeth Riede: When we think about things like that, where you are looking at kind of student patterns or data or, know, how, from your perspective, what should we be thinking about to ensure that these tools stay supportive and don't feel like the university kind of surveillance state, right? ⁓
Keith: We don't want to be that.
Elizabeth Riede: I guess to kind of put it another way, like where do you draw the line between that like proactive support and finding those moments of community engagement that we know are so important without being kind of creepy helpful?
Keith: Yeah, and we don't want to be a part of this surveillance state either. ⁓ So I think a large part of the answer is working with ⁓ our executive stakeholders like yourself and knowing where those lines are. And then designing and architecting our AI solutions to kind of fit. I just don't think there's a one size fits all for these type of things. On the back end of these systems, ⁓ we need to think about retention of
Elizabeth Riede: Yeah.
Keith: that we put into chat, how long do we retain those, ⁓ does public records laws look like, how much do we have to account for that, has the law even caught up to ⁓ of AI queries and prompts. again, I keep saying this, know, but the next couple of years are going to be really informative about ⁓ types of ⁓ questions. ⁓
Elizabeth Riede: Hmm.
Tom Van Gilder: add to that. It's kind of on us to add things like university contract and adding clauses like zero training clauses where enterprise university licenses that we're using mandate that no student data is going to go train the larger, you know, vendor global model kind of thing. And our data is our data and stays within our university's kind of digital fence, if you will. And then the others, Keith mentioned the chat bots is, you know, we can make the chat bots. keep the data as long as we want them to kind of thing. So if we are having kind of memoryless mode kind of thing with our chat bots and saying, all right, every 30 days the chats are gonna be kind of reset and we're not keeping student data. So then it doesn't become that creepy AI tool kind of thing that you mentioned.
Elizabeth Riede: Yeah, I think that's really good. I think about the difference between like, how did they know that? And how did they know that? ⁓ Like, we talk about that a lot in hospitality. Like, you want to be good enough that they're like, how did they know that it was my birthday? Or how did they like, but you never want them to use the other tone of like, how did they know that it was my birthday?
Keith: Thank
Tom Van Gilder: Right?
Keith: You
Tom Van Gilder: Big data is kind of creepy.
Elizabeth Riede: It's all cute until it's not, you know? Like these are amazing anticipatory touch points until all of a sudden someone's like, wait a second. ⁓
Tom Van Gilder: I was reflecting back the other day about I was kind of doing some Amazon looking around kind of thing. then ⁓ the apps that I play on, suddenly those ads are in my app too. like, sort of used a browser that's not so cookie-fied kind of thing.
Elizabeth Riede: You Mm-hmm. ⁓ So thinking about that, ⁓ when you, or is there an example where you've intentionally chosen either not to automate or not to bring AI into a conversation or a process, even where you felt like, yeah, we could, ⁓ but maybe we shouldn't.
Keith: I think you were touching on this a little bit earlier that you may be approaching that line with a couple of asks.
Tom Van Gilder: Yeah, and that's actually just happened last week and I was meeting with one of the folks and then some of the other folks that they were having to kind of somewhat of a data steward of the data and they're like, hey, wait a minute, let's have a conversation more. So like, all right, let's maybe have some more conversations around kind of some set guardrails around AI and defining, like I said, the human part of this versus the AI part and where should they never meet.
Elizabeth Riede: Mm-hmm.
Tom Van Gilder: And if we are, I don't want AI to ever just make a decision for us and not really be thoughtful in some of these processes. So I think this kind of guardrails document, for lack of better words, is something we start leaning on a little bit more often as we're kind of going through these campus discussions and kind of thinking through what this actually looks like and almost like a contract of sorts. ⁓
Elizabeth Riede: Hmm.
Tom Van Gilder: to make sure that we're all on the same page with what we're trying to do here.
Elizabeth Riede: Yeah, I think that's really good. We get so excited that we just want to go a mile a minute. You see a potential use case or especially where it can save hours or labor or effort from an employee. And you're like, yes, I want to do that thing. But you don't, especially in operations, we don't always think about five years down the road, seven years down the road, Keith, to your point earlier, public records, who could ask for this? ⁓
Keith: In
Elizabeth Riede: We aren't always considering all those things. And so it's really good to have language that holds us accountable and holds us true to that.
Keith: And I think the importance of evaluating risk as a part of these potential opportunities and having an actual risk framework is vitally important. ⁓ You know, what gets lost in this, ⁓ Tom mentioned data stewards earlier, ⁓ and we have tons of them across our campus. Data governance is critical to any successful AI program. How good is the data? What's the quality of the data? But my biggest concern day in and day out is where is that data going? So setting up the proper risk frameworks for that as a part of the evaluation of these use cases I think is really important.
Elizabeth Riede: Yeah, something else that's come up in the work that Tom and I have been doing is data definitions and data dictionaries and making sure that you're really clearly defining that data. Because the operational definitions of that data may be very different ⁓ and may be seen very differently by hospitality professionals than by more
Keith: Mm-hmm. Right.
Elizabeth Riede: skilled and trained subject matter experts, especially when we get excited about something. ⁓ So to that point, I think from an operational perspective, for my fellow hospitalitarians out there, what are some really good questions that we should be asking ourselves about, whether that's AI, data, risk, ⁓ what are some things that we should be considering?
Keith: Thank ⁓ I'll go back to data. What data is going to be consumed? Where is that data going? But I've said this quite a bit is we're not business process owners. Hopefully we're helping enable efficiencies through AI or other technologies. But really evaluating each of your business processes just to think through is there an opportunity for better efficiency if we layer on AI? You know that I think that's where you know, we know you do you do great work and you're thinking through appropriate business cases about food waste and scheduling and and those are brilliant ideas that you know we in IT get really excited about because you as the executive owner business owner have thought through this already right and and it's an opportunity for us to have a collective win. So yeah. I think just really honing in on business processes and understanding and knowing your data.
Tom Van Gilder: And I would add to this too that ⁓ part of this contract document that I've started last week was kind of at the forefront of it is kind of starting with some core principles of equity. So AI is an assistant, but not a decision maker kind of mode of thinking. The do no harm ⁓ thought processes of system must be rigorously tested to ensure it doesn't systematically disadvantage applicants, et cetera. ⁓
Elizabeth Riede: Hmm.
Tom Van Gilder: AI should be transparent, applicants should be informed that AI is in use to assist in a review process. And then start thinking through how do we ⁓ do bias mitigation and ⁓ the data and algorithmic guardrails kind of thing. So preventing the AI from learning or perpetuating historical biases, for example. ⁓ Data blinding, what the AI should not ever see, ⁓ it's kind of...
Elizabeth Riede: Mm-hmm.
Tom Van Gilder: one of the thoughts on this contract as well. ⁓ Things like a person's name could significantly disadvantage somebody. It's proven already with Google searches that. And then as I mentioned, the human in the loop architecture at the forefront is defining what the AI is going to do and what the humans are going to be doing and ⁓ keeping that at the top of mind.
Elizabeth Riede: Mm-hmm. Yeah, we talk about it in hospitality a lot. At the core of efficiency is role clarity. And so I think understanding where is the AI's role and defining that upfront and not allowing, we wouldn't allow scope creep from one of our employees. We're certainly not going to allow it from an AI perspective. So I think those are really, really good things for us to think about. One of the other things that you touched on was that AI literacy. And I think we're talking a little bit about it. now, but to kind of go a layer deeper, what is something that you would want every student, every employee working on campus to understand about AI?
Keith: I guess I'll kick this one. I think I know I'm sounding very redundant, but it's knowing what the standardized tools are and the level of sensitivity of your data. ⁓ using non-supported, non-standardized tools puts you and the organization at risk. So we've picked the tools that we have for reasons. But most importantly is to protect
Elizabeth Riede: Hmm.
Keith: our organization and the individual users. So that's probably A number one. Number two is just the ethical use of it. You Tom's talked about it. And that's why literacy has been so important. And we run a lot of great workshops across campus and we have a great website that we stood up ⁓ that we're really proud of. So I think we're communicating well, but how do you get to the masses with this?
Elizabeth Riede: it was so interesting. We did ⁓ an all staff So it was ⁓ all layers of our operation from frontline first employees to managers to frontline employees that been here for 25 years. And ⁓ so you can imagine kind of the spectrum of ⁓ literacy that was in that room. ⁓ just a wide variety of perspectives. And ⁓ our trainer pulled up Notebook LM. And ⁓ gone through ⁓ series of points around AI already. And they pulled up Notebook LM. And you could see everyone in the room. It had been 25 minutes. And ⁓ were in various stages of engagement. ⁓ And then he put in our Surf Safe training. materials into Notebook LM. And he had it generate flashcards for this training and, you know, Surf Safe is meticulous and it's also tedious. ⁓ Some of staff would even say boring, pointless. ⁓ And so. But he, so he had to generate these flashcards and immediately everyone in the room sat up, eyes forward, they were engaged, they were interested because all of a sudden they could see themselves in the technology. And I think you're getting at the heart of, I think what makes AI literacy so important is that people can, they can only be what they can see. And I think a lot, a lot of what we struggle with in hospitality is that people can't see themselves in this technology. And so I really like that point about making it relevant and making it be a place where people can feel safe to engage.
Keith: Right. And we're trying to meet everybody where they are. Tom and I, ⁓ we've been on a campaign to visit with the leadership across campus and ⁓ those are demos largely. And I think to your point, it's more demonstration of the power and the functionality of these tools. ⁓ And it gives AI some context and some folks. think the part that gets missed in this, think too sometimes is the level of change, right? You talked about the variety of within your staff, the pace of change to one person is going to not equal the pace of change for somebody else. And I don't think it's ever been more evident than with AI. So we got to be mindful of it and respectful of it, but it is certainly a challenge.
Elizabeth Riede: Mm-hmm. I think looking more broadly, when we think about those populations and kind of the different places that people come to us from and the experiences they come to us with, to your point, it's not one size fits all. Are there populations that either benefit first or kind of gravitate to these tools more readily? And then are you seeing populations that are at risk of being, you talked a little bit, Tom, about ⁓ equity in AI, you see populations that are at risk of being disadvantaged through this process?
Tom Van Gilder: I do worry about the AI being the new digital divide kind of thing where I'm a student and I'm making, you know, poor grades and I said, but I can afford ChatGPT Pro at 20 bucks a month kind of thing. And I'm going to fund that. And suddenly I'm an A student because I'm starting to use this. And then the student that can't afford this same tool and they just, they're going to stay the C student because they are, they don't have the same tools or whatever.
Elizabeth Riede: Mm-hmm.
Tom Van Gilder: then that's just not really fair, is it? So I worry about this as a human. mean, we've got to figure out as an institution, how do we not make this the new digital divide?
Elizabeth Riede: Yeah. Yeah. I think in hospitality, already see we're managing multi-generational workforce. We see people come to the table with different skill sets, obviously, and different experiences. And our supervisors have to manage that. And so I think that's kind of validating to here, but it's also concerning, right? Because I think we see that in our own operation. You see
Keith: real.
Elizabeth Riede: the managers that can engage readily with these models and can adopt the technology faster, can get a handle on the pace of change a little faster. They tend to rise to the top even if the capability of another person might be just as high or even higher long-term.
Keith: Yeah.
Elizabeth Riede: When we think about that pace of change, what's a skill in either campus operations or technology in higher ed that mattered 10 years ago, ⁓ matters less now?
Keith: Okay. I don't know. keep going back to Tom's, know, human in the loop. think there's, I don't think the human is going to be ⁓ irrelevant in any of these processes. so I'm not thinking anything of anything else at the top of my head, Tom. I got a little bit.
Tom Van Gilder: But as we come back to like the, like the, we used to have registrar group, you know, would do manual processes of degree auditing kind of thing. that was 10 years ago. Now that we have the, some AI tools that can help them, you know, and help our students do the degree auditing, kind of AI driven degree pilot of sorts and perform 99 % of those kind of same tasks within minutes, instead of a human having to do that.
Elizabeth Riede: Mm-hmm.
Tom Van Gilder: laborious amount of hours to do the advising part. So I do think that's high up there. ⁓ Same with financial aid and kind of looking at a student's financial aid ⁓ request. And we used to have data entry people that were speed typist kind of thing. that was a data entry position. And those are becoming maybe ⁓ somewhat less important of speed typing kind of thing. ⁓ But again, I think my message is AI should be co-pilot, not the pilot kind of thing. let's just make sure that we are, that's framing our thinking moving forward. And Ollie, I saw some kind of video reel that someone sent me the other day that was like, AI is still not getting 96 % of the jobs correct anyway. let's. fearful about AI taking all the jobs away, but let's figure out how AI can augment the jobs that we do have. Maybe that's a better way to frame it.
Elizabeth Riede: Yeah. Yeah, no, I think we talk a lot about, again, there are just moments that shouldn't be given away. There are moments that should always be human. But then there are some that I think everyone would be happy to give away, like folding laundry or cleaning the bathtub, right? ⁓ are some that it's like, yes, please take that from me.
Keith: Yeah.
Tom Van Gilder: Yeah, exactly.
Elizabeth Riede: my Christmas tree of laundry, I will gladly give that away. So ⁓ your perspective ⁓ people who eat and people who experience hospitality and service out in the world, ⁓ also as people who ⁓ thinking about the ⁓ possibilities, is there a place where you see AI can improve either hospitality service, customer service, the food experience. ⁓
Keith: Tom, I know you've been working closely with Elizabeth. What comes to mind?
Tom Van Gilder: Yeah, I'm excited about the possibilities of like food waste because that really gets on my nerves. I do worry about the amount of food waste, especially when we've got folks in our own community, have food insecurity and ⁓ how can we balance that? ⁓ Can we use AI for ⁓ the good kind of things and thinking all right, ⁓ instead of
Elizabeth Riede: Mm-hmm.
Tom Van Gilder: this stuff, all this stuff going to waste. And I know you've had a lot of, you know, I've had some conversation around this, but yeah, AI for good is what I would like to see happen in that front.
Keith: Yeah, think labor waste comes to mind too, right Tom? I know it wasn't, but you're thinking about that. think inventory management and inventory management systems with all of that data, the logistics I think are going to be impacted by AI. ⁓ I just never want to see the customer service component taken out, right? think, yeah, that's, ⁓ yeah, I think people still do expect to have an experience. And if they're talking to a machine,
Elizabeth Riede: Mm-hmm. Yeah, yeah.
Keith: A lot of that is lost.
Tom Van Gilder: Yeah, just because we can doesn't mean we should kind of thing. And I'll say not to not Bojangles, but I will. Yeah. I go from the drive through in five minutes, arguing with the dang food ordering system. That doesn't seem like good AI to me.
Keith: Right.
Elizabeth Riede: Yeah. ⁓ Bo-Linda, yeah. She's great, isn't she?
Keith: Yeah, Tom, I got to tell you a funny one. I keep saying thank you to the Automated Assistant.
Tom Van Gilder: You
Elizabeth Riede: Well, when there are robot overlords, they'll remember Keith was really nice to me.
Keith: Hahaha!
Tom Van Gilder: And Tom not so much.
Elizabeth Riede: Yes, I always say please and always say thank you, just in case. ⁓
Keith: We're way too southern, I think.
Tom Van Gilder: think maybe I'm the old man of the group of curmudgeon and just like I'm the tap tapping. I was like, human, give me the human.
Elizabeth Riede: Stay off my lawn. When you think about your hospitality experiences, whether they're on campus or elsewhere, you just mentioned Belinda. So that is a red flag for sure. But is there an example where you can think of that AI has added value already?
Keith: Mm, mm, mm.
Tom Van Gilder: Yeah, I mean, went to, I guess I'm trying to think of it was like one of the Hilton change or something. But anyway, it kind of remembered that I liked this, the same temperature at night and then turned down to the nice 68 degrees for me and not me coming into the room and trying to click through the wall thermostat there. So I kind of liked that. I didn't remember that I'm a 68 degree kind of guy. So I thought that was kind of cool.
Elizabeth Riede: Yeah, ⁓ that's a how did it know kind of moment.
Tom Van Gilder: Yeah, this is a little creepy, but I kind of like that it remembered that I'm a 68 degree kind of guy.
Keith: I think the personalization of the experience, think, you know, kind of that intuitive piece of it, I think is going to be helpful.
Elizabeth Riede: awesome. What excites you most about AI's potential in higher education?
Keith: I think it's going to transform teaching and learning. think the ⁓ employers are going to expect for you to be AI literate. So I think it's incumbent on us to make sure that our students are ready to hit the workforce. ⁓ But I'm excited for the efficiencies. Tom's mentioned several opportunities that we've already created very significant efficiencies that's saving the organization money. And I think
Elizabeth Riede: Hmm.
Keith: you know, that's not getting rid of people. It's just making them more efficient in the jobs that they're in. And I think that's, that's kind of the lens that Tom and I look at this from is, you know, how can we enable, enhance, improve their operations and make sure that they can focus on the right things instead of the mundane things, right?
Tom Van Gilder: Yeah, I kind of like where AI can be used as strategic kind of ⁓ more of an AI as a liberator kind of thing. So liberates me to do higher level kind of things over here. So if it's faculty and ⁓ spending four hours a night doing all this grading and it's all the same kind of grading that I know it, throwing it into AI assistant that could help me do that grading and get it down to ⁓ minutes instead of four hours, then let's...
Elizabeth Riede: Mm.
Tom Van Gilder: think through that. But on the same coin is like, then how do we teach our students to use AI effectively and not just as a way to kind of get through all the content and stop doing the higher level thinking themselves. There's a lot to this coin here.
Elizabeth Riede: Yeah, it's all a very double-edged sword. Is there anything else that you would want listeners to walk away with? I it's ⁓ such a good conversation because I think ⁓ much of our time we focus in hospitality at least ⁓
Tom Van Gilder: very much so.
Elizabeth Riede: the tactical, like, how do I make this process work? How do I do this thing? so think it's really important that we pull ⁓ other folks in to kind of challenge, what we be thinking about? What should we be considering that maybe we're not considering? What are some ⁓ potential risks that we run to help us future-proof our strategies and our operations. So is there anything that you would want our listeners to walk away with or be thinking about?
Keith: Tommy, you wanna kick it out?
Tom Van Gilder: Yeah, I mean, I guess I'm the IT guy who's maybe the very humanistic IT guy here. But I do think that, like I said, where are these areas that AI could be used to release the kind of from the routine work, but ⁓ which there's then double downing on how we act and as what are the things that AI definitely still cannot do, like moral discernment, deciding what's fair and what's not fair kind of thing. empathy, things like that, inspiration and motivation. ⁓ the human side of this, I ⁓ the AI could be critically important to help us get more efficient with the work at hand, but how do we double down then on things like accountability and efficiencies ⁓ doing it the right human way?
Keith: I would say do your research. ⁓ think there's folks that we talked about ⁓ terms of pace of change that are resistant to any new technology, whether it's ⁓ or anything else. And there's a lot of noise ⁓ the ether about ⁓ bad it is or how good it is. ⁓ so I would really encourage everybody to do their own research. ⁓ advantage of, ⁓ our organization, ⁓ any professional development opportunities to learn about it. and be open-minded about it. And be open-minded even to the point of, I might be able to utilize this in this particular area, and it's not as scary as I thought it was. Yeah, think Tom and I deal with both ends of the spectrum, right? The ones that are highly resistant and never will want to introduce it to their organization to how fast can we go and how much can we get done in the next 20 minutes. you know, that, but we're really thinking about the folks in the middle. and how can we really support them? So my big ask would be do your own personal research, come to your own conclusions about the use of AI, ethical use of AI, the sustainability ⁓ components of AI. I think all of those are ⁓ well-researched areas, but if you're just getting your notes from social media, it's probably biased, right? every organization's gonna have their own kind of tenor around AI. So I think understanding your organizational culture and willingness to change, think, is ⁓ something to be understood as well.
Elizabeth Riede: Yeah, I think we talk a lot about the, I guess all the balls that you have in the air ⁓ when you're running any business, but especially in hospitality, you you have the employees opinions, you have the guests opinions, you have your clients opinions or your organization's opinions. And it's juggling all of those through the lens of whatever decision you're trying to make. So I can definitely appreciate that not every organization is going to be Let's see what we can innovate. Let's see what we can do today. And I think it's important that even, but your customers might be. ⁓ Why haven't you done the AI ⁓ voice at the drive-through already? ⁓ And then Tom's here to tell them why.
Keith: Yeah, exactly. And he's not going to tell them thank you either.
Tom Van Gilder: It would just work.
Elizabeth Riede: Well, Keith, Tom, thank you so much for joining me in this conversation today. I really appreciate your perspective and I appreciate your willingness to come and share your role, your responsibility and a really valuable perspective for us in hospitality. So thank you.
Keith: Thanks, Elizabeth. Thanks for having us.
Tom Van Gilder: Thanks for having us.
