HAL Transcript: Thomas Livers on AI in College Athletics

This is the full transcript of the HAL: Who Signed Off On This? episode with Thomas Livers, Associate Athletic Director for IT Services at the University of North Carolina at Chapel Hill. Listen on Libsyn. The show is also on Spotify and YouTube. For the highlights, read Does Banning AI Work? A University IT Leader on Athlete Data, FERPA and AI Governance.

The transcript follows the published episode. It is lightly edited for readability: repeated words, false starts and short acknowledgements are removed, and names that the transcription software misheard are corrected. Timestamps mark where each turn begins in the episode.

Transcript

Chris Sowa (00:00): This is HAL: Who Signed Off on This? Short conversations with the people who have to sign for AI, how they decide to approve, where they insist a human still looks, and what they would need to see before they trusted an AI with those decisions. HAL is the computer from 2001 Space Odyssey, for us old older guys that can remember stuff like that. This show is about the people who make sure that never happens on their watch. And I am Chris Sowa, I’m co-founder of RedactSure, an AI controls and governance and privacy company, and I’m your host. Today we’re lucky to have today my guest is Tom Livers, the Associate Athletic Director for IT Services at the university of North Carolina, Chapel Hill. The screens is people work on carry student records, athletic medical information, eligibility information, recruiting files, all the most important stuff that we could imagine flowing through a university’s athletic department. And I want to welcome Thomas, you to the show. Before we get into some of the hard questions that you and I are gonna try to brainstorm and figure out answers to, I just like you to tell us a little bit about yourself for a few minutes. Some of the challenges and opportunities you’re seeing from newer technologies, including AI. So with that, the floor is open to you, Tom.

Thomas Livers (01:23): I’ve started in, IT in 93. So I’ve watched, the industry grow. I’ve watched capabilities, expand over the years. And, on a daily basis, my group does everything from the typical IT break fix, of the computer support windows for printers. But then, we, do roughly 280 events a year. Everything from basketball, football games. Mean, when we have a basketball game, we have 21,000 guests. And then during those events, over and above our typical IT job, we will be there to support the police for fire and safety with their command networks, supporting the ticket office for both sales and for entrance. Supporting the vendors for concessions, supporting our vendors for merchandising, as well as supporting everything from even the replay systems in all the sports. Department, about 10 years ago, my department, we built our own replay system, which is what we use for all of our Olympic sports, Replace. So like we built the infrastructure, we built the technology. And we provide that to our contests or events and matches as well as we also, you know, basically are there to help support the TV trucks, making sure that they have internet, making sure that they’re pushing the content to the providers. And, you know, basically make sure, try to make sure our guests have a good time and then we do… Surveys at the end and try to improve.

Chris Sowa (03:14): Just give me a sense. How many, you know, people are at some of these events? Like just, you know, the largest what largest like a sold out event.

Thomas Livers (03:23): About 50,500. Our basketball games are 21,600. And then our.

Chris Sowa (03:31): My home my hometown here that I’m living in, Tom, we’ve only got about fifty thousand people here, okay?

Thomas Livers (03:38): So can fit in there.

Chris Sowa (03:40): Our whole town can come to you and see. Can you guys arrange that so we can have a special a special moment there we’re all gather together? My point to the viewers in the audience is that Tom is basically managing his team at the UNC is basically managing a small city’s worth of people for a few hours given. But there’s a lot of stuff you hear heard him talk about. Everything from the vendors, from the people coming in, the security. And we’re gonna dig into some of that stuff here today on the call. And we’re super lucky to have Tom on the show to kind of give me his perspective. Now this being Hal who signed for this, we always have the first three questions kind of grouped together. Just to challenge our guests. This is a this to see how they’re thinking about the future possibilities. And, you know, we in the AI industry have hurt have heard about, you know, this company Hugging Face that was hacked into you know unbeknownst to anybody for about three days until open AI finally they came and they said hey those are our agents that was oops right so we know agents do things that are unexpected right even to the labs and these guys how much money do they have Tom compared to the budgets we have they’ve put they we they’ve got a few dollars to spend so the questions are you know if an AI agent transferred a hundred thousand dollars out of your organization so let’s say there was some money flowing in from somewhere, probably ticket sales is the only place we’re gonna find cash like that, just hanging around, maybe. And you know, who would ultimately be responsible? It is somebody in accounting, you know, who potentially would be responsible for something crazy like that happening? And do you know what people can read off of their screens today? And if an AI agent did something, acted on a record, whether it’s an account or something else, you know, how would you go about figuring out on the back end who authorized that and who was really accountable? Because I like to say, you know, all kinds of people, students, teachers, everybody, a lot of times they’re happy to use AI to help themselves, right? So we’re they’re we’re happy to take credit for AI’s work when it suits us. But when the stuff doesn’t go our way, who’s on the hook and who’s responsible? And that’s what we mean by who signed off on this. So Tom, help yourself to any of that questions and any of the stuff that you guys are still working on. That’s perfectly fine. Because even the big labs haven’t got it all figured out yet.

Thomas Livers (06:21): As far as AI, you’re right. It probably would end up being something in the ticket office, right, for a hundred thousand dollars. But the one thing I can say is we don’t have it at all anywhere near our ticket office. Our ticket office, actually is one of probably the most secure locations we have. I mean, we have actual, they don’t even talk to DNS servers. They talk to a set up DNS server that only has the entries that they need to talk to. So you couldn’t even go open a browser and do anything other than talk back to our ticket company and do a sale and everything is done, you know, obviously securely. So that one, but if it happened like, whoops, let’s say something like that happened. Mean, you know, ultimate responsibility would come down on to whoever installed the set agent because again, our users can’t install software just on their computers, but obviously you could probably get a browser plugin without having elevated access. It would probably go back to the user, in the case of the way the universities are set up, that $100,000 loss would end up probably coming out of our budget. We would have to take a loss on that and then have stricter guidelines put in place. But I feel pretty confident about that at this point that one shouldn’t happen. But, know, who knows?

Chris Sowa (07:48): Well that’s a great answer. I mean th there’s clearly, you know, you’ve got controls in place and for at least for now it sounds like, you know, th there’s not even the ability to open a browser and it and that kinda shuts out my only chance to get really good seats at a UN C game. So between us guys that I don’t have any in connections. I think you need an in connection to get the hot seats. That said that said, you know, there’s I’d like to dig in and understand a little bit more about, you know, some of the stuff that you guys are doing, not just during the games, but maybe at stuff after as well, whether it’s related to film, tracking, recruiting, analytics, sports science. What’s happening in i with within those areas? How does your team your department, if you do support those, do that? And are any of the coaches that you’re aware of using AI or are they talking to you about how they could do use that to help in any of those areas?

Thomas Livers (08:51): Yes, so we have several initiatives that are going on. Currently in a proof of value phase with Lenovo about basically we’re taking all of our wearables and all the wearable data from all the different manufacturers and pulling it into one centralized source on campus in the secured location so that we can build reports and heat maps and do everything that we need to do to make sure that the student athletes are in the best shape that they can be and making sure that they’re hitting their marks. And also, hopefully, in the future, we’ll be able to prevent injuries because you’ll see, well, this person’s not symmetrical today. What’s going on last time we saw this with somebody? They pulled a hamstring or they did this. So we are working on that. The one place that I really see AI being helpful and not being too dangerous or nefarious as it can be would be in their video analysis. I’ve also worked with another company that we’re doing a proof of concept on right now with their storage and their storage can take any of the video that we have. Look at it. Look through the audio, clip it out, you can actually sit there in real time and go, show me the score at the 35 minute mark, and it will tell you the score at the 35 minute mark. You can say, hey, show me the drop pass in the fourth quarter on this play, and it will bring that play up. Before, know, again, you’d have video analysts going back and scrubbing it and cutting the videos for the coaches. And now literally with a microphone, a coach can then go. Hey, I wanna see this play and that play is there and they’re watching that play. Yeah.

Chris Sowa (10:48): And that probably settles some arguments. If somebody said, yeah, this is what happened to no. Well, let’s just call it up. Boom.

Thomas Livers (10:54): And it is just, it is so slick the way that basically the AI can sit there and actually track the whole game and you can go, how many times did number 32 touch the ball? And it will tell you the stats for 32 by just watching the video. It is taking that digitally enhancing their number and then keeping their stats and then pulling from our stats and then just. Mirroring, merging the two together so that you have a bigger picture of what’s going on.

Chris Sowa (11:24): Now from a data perspective, you’re literally getting into, how is the health of that athlete on that day? I’m not saying coach could weigh, hey, do I poll you know, do I play him or don’t I play him? You know, so that gets that’s gets pretty personal. When you start talking about athlete data or student data, you know, one of the challenges you guys face within education is the FERPA compliance stuff. How d how do you think about that type of data and you know, guidance as coaches are starting to think about new ways of doing things that you would, you know, need to think through with them before you would kinda green light some stuff.

Thomas Livers (12:11): Well, mean, anything that really touches any of this student records or student information needs to run that through. We have a data governance group that’s on campus that we would run anything through them to make sure. And then we have the privacy office that also We’ll go through and look at everything and sign off on anything before we can even use it or try to implement anything.

Chris Sowa (12:36): So well it’s I wanna talk a little bit more about, some of the stuff that, because we’re all watching the game that the fans don’t normally get to see. But you got your team is really focused on it. And that’s, making sure that, broadcast, ticketing, medical related stuff, security, it’s all happening at the same time on the network. The whole net the whole university shares. But you’re in a critical infrastructure perspective. I don’t wanna say that, if the internet went down for a few minutes and somebody was res you know, researching a term paper, that wouldn’t be a bad moment. But, you know, if broadcast were to go out for a minute, people would raise hell, I think, in the UNC. So you’ve got you’ve gotta have some pretty resilient rails that you’re running on. Do you man how do you manage that resilient piece? Are there any specific tools? How do you think about that? It’s probably what, automatic for you guys.

Thomas Livers (13:40): Have a great relationship with, the university. You know, cause obviously I don’t think I told you this, but we have a staff of five full-time, IT members that, and, that’s even against the regular industry standard of like Gardner, think, one for 30 or right. So we should have five more just for the IT, regular IT work, not all the games. And all the replay work and all the other work that my team does. So we rely heavily on the university and the university’s network and their networking group. You know, whenever there’s a group of 20, 30 people that do what they do and that’s all they do all day, we will take advantage of that. Like, just like all of our server and server infrastructures, when I first got there, my predecessor had built their own server. Data closet and i moved everything to central campus because they have generators with dual power your dual generators to a power they have contracts for sixty plus days of fuel plus they keep thirty days of fuel on site so you know it’s just like why would i want to the headache of trying to manage that so every anytime we can find a resource on campus or another resource we’re that’s reliable, we are quick to work with them and build a partnership.

Chris Sowa (15:07): So I it’s in it’s incredible. The stuff the list of stuff that your team is doing, it keeps getting longer and you guys are doing it doing these gigantic events with five guys with all this stuff going on. So yeah, I can imagine you guys need to join arms, with a larger UNC IT department to pull off stuff like that, for sure. And I guess, that leads us into my next, question is, where does the athletics IT end and I in campus IT begin? And I understand you guys probably work so close as a team, so maybe that line is, you know, sometimes a l a little bit of a dotted line. But y if you could help explain how that works, you know, in practice.

Thomas Livers (15:52): It’s obviously a team effort. We are one big team. We’re all the university whether or not we’re in the athletic department or whether they’re in the wireless or the networking group or their server group, but basically everything you know falls on us in the sense of if it doesn’t work there everybody’s gonna come to us first and then we’re gonna have to figure out how to go well, that’s a networking issue or that’s a server related issue or that’s this issue. We still have to do some of the first time or what do call it, triage, right? Like the, whose responsibility is this? And then get with that group. A lot of times at our events, like I said, there was like a football, which we had a football game this weekend. I don’t know what the actual, how many 30, 40,000 people were there. I know it wasn’t capacity, but it still was a great turnout until the rain came. During an event like that, we had one or two network outages, but that ended up, we had to troubleshoot it because two network outages basically happened at the same time, which makes people go, it’s the network. But it wasn’t the network. One of the providers on the back end was having issues on their own thing, which was why their equipment didn’t work here. And then the other provider. Just didn’t have everything set correctly and then it just happened to go offline at the same time but you know would make you think it’s networking so we had to go troubleshoot figure out what was wrong and then you know in both those cases it was on the vendors end so it was not us.

Chris Sowa (17:39): And in ultimately, you guys are the ones that have to explain if something to the dean and or whatever else, if something weird happens. You sound like you ha I know there’s a lot of young guys that are gonna be listening to this and they’re gonna say, Tom has the coolest job ever, right? These are guys who are IT guys, but they love they would love to be out watching the game. But Even you have bad days. I gotta believe this. Tell me it’s true. So what breaks on a bad day or what is a situation where you, something broke on a bad day and because of that pain that you felt this is these are the things you did to make sure that didn’t happen. So, share a little bit of pain because everybody wants to jump into your chair after listening to you talk about this.

Thomas Livers (18:32): Well, I must be a great salesman then, but no. Like last year during basketball season, the replay system, that’s not the one that we developed. It’s the one that the conference mandates us to use. It was not coming online. Mean, it’s in the middle of the game. We had it online, we got it online right before start of the game. The computer… That it runs on and everything crashed one minute to coming out of halftime. So then I had to sit there and get everything back up, get it all back up and working. We got it back up and working, I think at the 14 minute mark of the, it was the Louisville game. And not even a minute later was called for a review, but luckily it was back up and working. You know, those are the things that, you know, cause a little heartburn and make you lose sleep is because. You end up troubleshooting devices in the heat of the moment. Is it a network problem? Is it a device failure? Is it a power failure? What is it? And you didn’t even design that said device. Then you’re in the middle of calling the vendors, talking to the vendors while you’re on the phone. I mean, I had the vendor support. Obviously, that was great. They were on the phone. They’re like, try this, pull this fiber, do this. So you’re sitting there sticking your hands in cases and trying to get it back up. So yeah, that is quite stressful when things like that happen. Mean, when it’s a great day and everything works and you can sit back and see a game, I don’t think I watched that much of the, I saw probably some of the fourth quarter because it had rained and we had an hour and a half rain delay basically during the football game. So. Everybody had pretty much left, any problems would have been handled pretty quickly. But yeah, it’s not as glamorous as it may sound.

Chris Sowa (20:33): It’s not like sitting at home with a beer in your hand, is what you’re saying.

Thomas Livers (20:36): Definitely not. Definitely not like if the TV goes out, well, yeah, no. If something goes wrong, have people that let me know from our staff to our fans. That’s what we want to try to do is have the best fan experience you can have. Because God knows tickets aren’t cheap and taking a family there and everything and doing everything that you need to do for an event, you want to have a great experience. Otherwise you’d be sitting at home on the couch.

Chris Sowa (21:08): And so now hop into something that is maybe a little more mundane than sweating it out in the middle of the game. You know, so we talked a little bit about this before, student athlete records and educational records, you know, some of them also have medical and eligibility information. You know, how does FERPA shape what AI can see or touch? Or if they’re not touching it yet, you know, if a coach is was to kind of come to you and say, hey Tom, here’s what I’m thinking about. I’d like to use this chat bot. You know, what kind of guidance would you give them, right? Knowing what you do about compliance and you know, also being sympathetic to the to what they’re trying to do, ’cause you know these guys.

Thomas Livers (21:55): I mean, that’s the danger, right? Like the dangerous thing is if you have it on, if you have the, which you shouldn’t even have the FERPA data on your laptop, you shouldn’t have any of that. But I mean, if you do, and it, and you know, goes off the rails and starts harvesting that data, I mean, you know, that to me is the biggest scary thing that I can imagine happening would be getting somebody’s, you know, FERPA records out there. Luckily, I have not had a coach come up to me and ask directly about that. Mean, AI right now for me is part, you know, I’m, I’d be too old, right? Like at this point, too old school to go. It’s great for like making my emails, you know, sound good and everything else, but I’m not quite, I guess ready to let it loosen in a lot of places.

Chris Sowa (22:50): Just as you said that, just to kinda like a for instance, and I don’t even I don’t I’m not the guy I’m not I’m never been as close to coaches and guys as you are, so maybe you can you I’m sure you can come up with a better use case than me. But you know, if a coach was looking at the roster, right, and he’s looking at the other team coming in, that they’re you’re gonna play Louisville or whoever else is coming in, and he’s trying to figure out, you know, matchups and things like that, and he said, you what? I’d like to dump all these stats and stuff in about these players, the compet competition and also my team into AI and just do a few different, what ifs, you know, what if I played this guy? What if I’d put this guy over here in this position kind of a situation? What would you know, need to see in place before you’d let AI near that roster?

Thomas Livers (23:44): Actually, for that example, there’s no real FERPA data. There’s nothing there. The only thing that I would say is I would want to make sure that the LLM, whatever they were using, was not saving that data, was not going to use that data for its own training, because that would be helpful to the competition, right? But the rosters are public information. So you can see somebody’s roster. I could pull anybody’s roster right now and sit there and match my roster against, Louisville or Clemson or whoever. And who should I play against whom? For that kind of use, would have no heartburn with something like that.

Chris Sowa (24:29): Y you’ve got maybe it’s more in the know, I happen to know that, Billy’s, got a got an injured ankle or it’s a little bit sore. But there’s some potential, medical information or other stats that are maybe more personal that they’re tracking on that particular athlete during the week if they’re wearing these wearables and stuff. So while it’s not, it’s not super sensitive yet, you know, what you can see on you can potentially whether it’s in the head of a coach and he types it in, or if it’s kind of built in with an agent looking at different systems, you know, you’ve got more stuff that I could use if I was, you know, gonna run out to polymarkets and make some bets or something like that, right? So I guess that’s we could talk about that a little bit more, but that’s the future.

Thomas Livers (25:26): Yes, so it is the future, right? That’s, unfortunately, the one thing is I don’t know a lot about polymarkets or any of that because we’re not allowed to bet on anything really. So, any of the sports or anything. So like, know nothing about all this, but it was, cause it was like the game. Yeah. It was like the game, the game this week weekend, against East Tennessee state. It was 35 nothing and the rain happened right at the end of the third quarter. And we’re all like, why don’t they just not do the play the fourth quarter? Just call it right now. Mean, just handshake it. And everybody’s like, well, because the betting markets in Vegas and everything, you can’t do that. You can’t change. I’m like, I don’t care. I can’t bet. So it doesn’t bother me one bit if they have to give the money back. That is one of the things that happens. I can’t remember the name of the company, but basically we had this company wanting to put these devices in a few years ago. And when we looked up who owned the companies, they were basically owned by the vetting sites because they wanted to see, Course they want to see, yeah. They want to see who’s basically, it would have given them more information than like we would want anybody to have, but they’re doing that because of what they’re trying to get done, right? Like their marketplace. But yeah, I.

Chris Sowa (26:53): So kinda looking down the road a little bit, you know, you ha you know, AI is the latest player on the field, right? But you’ve s you’ve seen this game before, you know, you’re in the course of your career, the internet kinda grew up, then you had all this cloud buzz and you were taking stuff off your servers and putting it in the cloud. Y n if any of us had a crystal ball, we’d be on these polymarkets, not on the game, but on the AIs themselves. So I d I’m not asking for the perfect answer, but just in your world, w are there some things you might expect might happen in athletic organizations, let’s say over the next three years? Whether it’s stuff that’s good or you know, challenging, let’s call it.

Thomas Livers (27:37): I really think that, you know, again, like as I said earlier, the AI in video analytics will probably be huge because like now when I can go and have a cut a game in seconds, I would have taken a human to cut the game. Probably hours right because they have to watch it cut it the way the coach wants it you tell the AI hey this is how my format this is how I want my clips this is what I want and it just does it in real time basically that to me would be is a huge Benefit of AI I think in the athletic world I can see it, you know, like were talking about rosters possibly a possibly. I mean that’s probably public data that you could get and the coaches would be able to tweak it with, I know Billy’s hurt or this or that, right? But as far as others, it may be predictable eventually, maybe five or six, maybe sooner where they can predict injuries possibly, because with the video analytics, they can actually tell how you’re landing and everything else to where some of these wearables may be even replaced by just video analytics with AI.

Chris Sowa (28:52): All right. Well, I think that, we’re all gonna be w I think sitting on the sidelines as that unfolds and you’re gonna be there. So it’ll be interesting to see how that stuff comes together. You know, I just wanted to ask, you know, one more thing, you know, from your perspective, if you’re s you’re at a meeting with a bunch of other guys that are, you know, similar roles from another university. And he was he just kinda looked over and y the CIO says, we just don’t allow AI. We blocked AI from everything in the university. So, you know, we’re fine. We don’t have to worry about any of this stuff in the future. You know, w what would you kind of gently tell that guy, if you felt comfortable enough to do?

Thomas Livers (29:40): Well, if I felt comfortable enough, I’d be like, well, guess what they’re doing at home with your data? Because that’s what they would be doing. Mean, again, you find all the time, everybody goes gets a chat GPT, right? And then you don’t know what data of yours they’re leaking out into it, which is one of the things where you can’t have a total, basically no AI. We’re never going to touch the tool. But you can have a phased approach and have put in the governance and put in rules and then pick tools. It’s like the tools that are allowed for campus because all the data, everything else stays with inside our environment, our secure environment that’s provided by Microsoft. So it does not go out and… Anthropic is not building on it and Google’s not building on that data. That data is inside our tenant. So from that aspect, you know, there are ways to do it to secure yourself. But I would say that you’d probably be pretty foolish if you didn’t think any of your data is getting used in any of the chat GPT, Anthropic, or any of the AI learning models, because somebody in your organization or somebody who’s working for your organization probably is cutting pasting it somewhere.

Chris Sowa (31:06): Tom Livers here with us on the show. Thank you very much for all of your insights. I still know there’s a bunch of young guys out there who are gonna be looking up your job description and figuring out how someday they might be you. But that’s the show. So I just wanna say, as we always do, if you sign for AI decisions in your organization and you have an answer to who signed off on this, or if you’re still working that out, we’d love to hear from you. And thanks to all of our viewers. Thank you.

About the show

HAL: Who Signed Off On This? is a show of short conversations with the people who have to sign for AI: how they decide what to approve, where they insist a human still looks, and what they would need to see before they trusted a machine with more. HAL is the computer from 2001: A Space Odyssey. The show is about the people who make sure that never happens on their watch.

About the host

Chris Sowa is a founder of RedactSure and a former CEO of an AI company; he started it years before ChatGPT existed. He previously led AI at Accenture, served as Global VP of Strategy & Innovation at Schneider Electric, was CCO of Sovos, and spent more than a decade at Oracle, with earlier roles at SAP and IBM.