Ethical and Legal Ramifications of AI with Kara Franker

Intro: This is Brand USA Talks Travel, elevating the conversation about international travel to the United States. Here's your host, Mark Lapidus. 

Mark Lapidus [0:09]: I promise that this will not be a boring podcast. She's an attorney, a DMO leader, and a former NFL cheerleader. This is not a combination that you typically find in travel, folks. I hear that you spent three seasons as an NFL cheerleader for the Kansas City Chiefs. Any chance you remember the cheers, and also, tell us about the experience? 

Kara Franker [0:28]: Yeah, absolutely, I remember the cheers, but if I did that for you, I would have to kill you, so we can't go there. But, yeah, I cheered for the Kansas City Chiefs, and obviously they are Super Bowl champs now, but then it was a trying time. We're talking about 20 years ago. I was 19 my first season; one of the youngest, if not the youngest, actually, on the team. I was going to the University of Kansas at the time, so going to college and also living life as an NFL cheerleader. It was absolutely amazing. And I did three seasons. It was fantastic. 

Mark Lapidus [1:03]: Wow. Any highlights? 

Kara Franker [1:05]: Yeah, I'll tell you, it was a highlight and a lowlight, and that was my 21st birthday, where it was a playoff game in January at Arrowhead Stadium, which is the loudest stadium in the country, playing Peyton Manning and the Indianapolis Colts. And everyone had big plans for not only winning, but celebrating my 21st birthday afterwards. Well, guess what? We lost. So the only person that went out with me afterwards was my boyfriend at the time, husband now, because everyone was so upset. So a highlight and a lowlight. But I did get to go when the Chiefs played in Miami, obviously led by Patrick Mahomes a few years ago, I got to go to that game and be at the Super Bowl. So that was definitely kind of full circle. 

Mark Lapidus [1:51]: Wow, what great memories. It's my pleasure to welcome Kara Franker, the newly-appointed President and CEO of Visit Florida Keys. She was previously the Chief Executive Officer of Visit Estes Park. Welcome, Kara. 

Kara Franker [2:03]: Glad to be here. 

Mark Lapidus [2:04]: First off, congratulations on the new role. How long have you been in it? Like a week, two weeks, something like that? 

Kara Franker [2:09]: Yeah, so I don't know when this is gonna air, but my first day is September 9th, so... 

Mark Lapidus [2:14]: Oh, okay. 

Kara Franker [2:15]: We're right in the middle of that transition. 

Mark Lapidus [2:17]: That's fantastic. When you look backward a little bit at your time at Visit Estes Park, what are a couple highlights from that? We should hear about that before we move into the present. 

Kara Franker [2:26]: Yeah. Three great years in a destination where it's the gateway to Rocky Mountain National Park, so just beautiful mountains, beautiful weather, a great community. And one of the big highlights is we figured out a way to get funding for workforce housing and childcare without dipping into the tourism budget, which is like an impossible thing that could be a podcast on its own. And then also extending the season, you know, trying to get folks to come in the winter, because it's a summer destination. So getting festivals like a really weird, quirky one called Frozen Dead Guy Days. So those are some of the fun things we did. 

Mark Lapidus [2:58]: What a great experience. Okay, so let's jump into the main reason why we're here today, and you're going to have to put your legal hat back on for some of this. We have this ongoing conversation on the podcast, Kara, about AI. It's never going to end, really, in the time I'm doing this thing anyway, because we're going to be talking about AI for the next decade. As somebody with both legal expertise and tourism experience, how do you see the intersection of AI and legal considerations evolving in the travel industry? 

Kara Franker [3:24]: Well, let me tell you, it's a minefield that all of us are going to have to take on whether we like it or not. Because as AI evolves - and we're talking about owned content and copyrights and all of those different elements and liabilities - you can't get away from the legal piece of it, and you also can't get away from the ethical side of it, too. It's something that everybody's going to have to deal with. 

Mark Lapidus [3:47]: So part of my hesitation with AI so far has been, actually, the hallucinations. We've been toying around with using a AI trip planner for Visit The USA, and the hallucinations are way down, like below 5%, but I can't have someone planning a trip to the United States and then all of a sudden wind up with an itinerary to Canada. Like, that would be a job-ending circumstance. 

Kara Franker [4:12]: It would be over at that point. 

Mark Lapidus [4:14]: That's when the CEO is calling me and going, "What did we just do? Like, I'm reading about something in the Wall Street Journal today, and it says we're taking tourists and sending them to Canada. That's not cool." 

Kara Franker [4:24]: No, would be bad. 

Mark Lapidus [4:25]: So how do we deal with that? How, like, how do we move forward? Does everybody have to get used to the idea that there's risk, or do we just not do it? Do we wait till it's perfect? 

Kara Franker [4:35]: Yeah, we can't wait till it's perfect, because the second that we do that, the industry is behind, and we lose market share, and we lose our, you know, technical abilities to be above and beyond and savvy in this piece of it. So yes, there is risk, but there's ways to mitigate it. And some of that is going through those legal and ethical pieces, which are different because, you know, something may be legal, but it might not be ethical. And then how do you kind of look at those boundaries and make the best decisions for the organization? But the hallucinations are undoubtedly a big problem. So then you can't just turn on the technology and say, "Woohoo, go out there, tell all the visitors to come and where to go!" Because that's when we get into problems. There has to be the human element and the partnership between the two, and that's how you get the checks and balances. It's not going to be perfect, because humans aren't perfect either, but at least if there is a way to double-check. So there are some tools out there where, let's say you're going to ask, you know, "Plan my trip to Colorado, and I want to go to Rocky Mountain National Park." 

Mark Lapidus [5:38]: "Let's go to the Florida Keys." 

Kara Franker [5:39]: Let's go to the Florida Keys. "Plan my trip to the Florida Keys and tell me what to do." There are tools out there where at the same time that someone is asking that tool, whether it's GuideGeek or whatever, you can have a person assigned on the back end to basically fact-check as it's happening. So that's time-intensive and person-intensive. 

Mark Lapidus [5:59]: Boy, it sure is. But with thousands of searches going on at one time, how do you do that? 

Kara Franker [6:04]: Right? How do you scale? How do you scale in a way so that you analyze the risk? So at what point do we have enough people going through those conversations, and at what point is the technology good enough to be the frontrunner for you? And those are questions I don't know the answer. 

Mark Lapidus [6:19]: Don't we also have to get used to just the idea that it's going to be wrong occasionally? We just have to get over it? 

Kara Franker [6:25]: Yeah. And are we to that point yet? 

Mark Lapidus [6:27]: No, I don't think we are. 

Kara Franker [6:28]: There's so many bad things that have happened in that realm that I wonder. I don't know if we're there yet or not either, I agree with you. 

Mark Lapidus [6:35]: Okay, so let's talk about mitigation of legal risks. How do we best mitigate those? 

Kara Franker [6:40]: Well, the first thing is that everyone has to realize - pick your poison, whatever tool is, let's take ChatGPT. Anything that you tell ChatGPT can and will be used against you. Like, imagine, like, that's the moment that you had too much to drink on the night going home, and an officer pulls you over: anything you say or do can be used against you. Same with ChatGPT or any of the other tools. It's inputs and outputs. If we're going to talk technology and legal implications. So whatever you put into it, there aren't those boundaries. It can go into the ether. So if you're telling it information about yourself that's personal, it now knows it. If you tell it information about your visitors, it now knows it. And then there are laws that are out there that you could get in trouble with if you're sharing information. 

Mark Lapidus [7:27]: I think what you're saying here is if you inputted a bunch of data about visitors to your destination, that AI might use it in some way that you're unclear about. 

Kara Franker [7:35]: Right? The second that that data leaves your fingertips and goes into the ether, you don't have control over it. You don't know how it's going to be used. 

Mark Lapidus [7:42]: Betcha a lot of people don't think of that. 

Kara Franker [7:44]: It's a big liability for the organization. Because there are laws in place in, like, California and Europe - we're going to see more kind of in those lines. But if you think of a DMO, it's not just wherever you're located. So if we're located in the Florida Keys, it's not just Florida law that governs, because we're talking to visitors all over the world. So you can't just think of your own jurisdiction; you have to think of all these jurisdictions. And then if you just take that back a piece, okay, there's privacy laws in place, and some of them are more strict than others, especially in California and Europe. So you might think, oh, you know, wherever I am, it's okay if I take this information and this data about a visitor or my employees or whoever it's about, and put it in and tell it to put it in a spreadsheet. But the second that you do that, you're opening up a can of worms. So what do you do? You don't put it into open networks. There are providers that you can have a contract with, where that there's like basically a net around that data, and it stays with your organization, and then that way it's contained and you don't take the risk of leaking. Now, could people hack it? Yeah. There's always going to be somebody out there that's going to try to hack all the information. But at least you as an organization have put in those boundaries to try to make sure that that information doesn't get out. 

Mark Lapidus [8:56]: I understand that you implemented an AI at Visit Estes Park. Tell us about that. 

Kara Franker [9:00]: Yeah, so we actually put two in place. One of them is an internal function, like how we're talking right now, of how can staff use AI to give them superpowers and to make them, essentially, more efficient and faster at their job, but not just into the open ether where that information goes everywhere. So we used one company for that, for staff to use internally. 

Mark Lapidus [9:22]: Who was it? 

Kara Franker [9:23]: Intentful. The product is called Happy Places Plus. We added the Plus because we wanted certain things with that product. And then basically that's what people probably already use ChatGPT and others for, which is, "help me write this great social media caption," "help me write a blog post," like some of the content generation things. And then that way it's protected so they can use it in a safe environment. And then we have an external version. And so that's through GuideGeek, which is owned by Matador Network, and we call it Rocky Mountain Roamer. We hit bumps in the road in the beginning when we were first testing our version. We were training it on our website, and then in other sources, like the National Park website and that kind of thing. And then we would give it to like our board members and certain community members and be like, "Okay, test it, throw everything you can at it and see what happens." And then we thought we were in like a great spot, and then one of my board members asked a question about his company. It didn't know a thing about it. 

Mark Lapidus [10:17]: Uh-oh. 

Kara Franker [10:17]: Right? So we hit, like, some road bumps, and so what we figured out is, we were going to use our website as like the first source. And then our website has been this content creation hub. So every stakeholder has to have like built out information. And then Rocky Mountain Roamer is scraping the website constantly and becoming smarter and smarter. And at the same time, staff can go in, see all the conversations, go back and send somebody like extra information if they need it, or even edit a conversation live. But that also goes back to this question of, how do you have enough staff? If enough people are finally using this travel tool, how do you have enough staff to be able to go in and check the technology? None of it's perfect, but it's the process of trying to figure it out and then learning as we go and making mistakes. Failing forward, is what we call it. 

Mark Lapidus [11:09]: So, time for another shameless plug, as I often do. Ross Borden was on this podcast, I don't know, maybe two or three months ago. So if people are interested in hearing about GuideGeek and what that's all about, they can listen to that podcast. And of course, I've had numerous conversations with Ross about a potential travel planner. It's a little easier, I think, to do it for a solo destination that is an entire country; we have a few more challenges than most people, but we're still interested in pursuing it. At some point, we'll probably go to RFP and see what company out there can really help us. Did you have an RFP process when you hired GuideGeek, or was it just the go-to? You just found it, liked it, and adopted it. 

Kara Franker [11:49]: We were like their first client. 

Mark Lapidus [11:51]: Okay, 

Kara Franker [11:51]: So nobody was doing this. There wasn't a lot of options. Like, we could have gone to an RFP, and then people would probably been like, "Oh, we can do it for you." But we were, if not first, like their second signed up. So we've been in it from the beginning. 

Mark Lapidus [12:06]: So back to the other AI that you were talking about, which is for office use, is it expensive? 

Kara Franker [12:11]: I'll be completely transparent with you, because we were also one of the first with Intentful and with Matador Network. And Visit Estes Park is a smaller DMO; the budget is under $10 million. So we were able to get in for very inexpensive rates for both organizations. Now that it's been a year and more and more folks are doing it, I don't know what the going rate these days is, but for Visit Estes Park, it was affordable. It's one of the advantages is if you go in first, because you can be the test case. 

Mark Lapidus [12:41]: Well, and fortunately, there is more competition coming on, so that always lowers prices. We'll see what happens in both spaces. We haven't touched on ethics yet, so talk to me about ethics. What are the ethical implications of AI and tourism? 

Kara Franker [12:55]: I think what you have to think about is, just because something may or may not be legal, does that necessarily mean it's ethical? And so this is a question to ask yourself, not only, like, for life in general, but in any industry, too. So everyone deals with those same side of things, ethics versus policies versus the law. For example, is it illegal to go in and create an ad and have AI create an ad using a famous person? It may not be illegal; is that ethical? And would that be a good representation of your destination anyways? You know, like, is that a good work product? Those are the kind of questions when we're doing, when we're talking about tourism marketing, and we're talking about what we're getting out of AI. Okay, is what we're doing here and what we're putting in front of the public something that we need to tell them about? Did the ad that we create, or is the copy that we're creating, or the image that we're creating, such a distortion, or was it created so much by AI instead of a human, that we should tell them this was AI-generated? 

Mark Lapidus [14:00]: So you're talking about, for example, you create a video ad, and then the bottom, it would say "Generated by artificial intelligence" or "Generated by AI," however you wanted to phrase it. 

Kara Franker [14:08]: Right. At what point should you self-report that, as an organization, you used AI to create that? That's an ethical question. Like, there's not a law that talks about it right now. 

Mark Lapidus [14:19]: I'm thinking about this. It also drives down into imagery. So when you have an AI create images, the thing I worry about that is that it's using composites from various places. And how do you know for sure they actually have the copyright ability to transfer that to you? They probably don't, because third party pass-through is a very difficult thing unless you obtain it in advance. 

Kara Franker [14:39]: Mark, that is beyond intelligent. And for not being a lawyer, most people, I don't know that fully understand what you just said, and that is gold. Like, yes - anything that's coming out that AI used to create it could be a composite of things. Whether it is or it isn't, you don't know that that copyright isn't being violated. And there's actually a ton of lawsuits being thrown about right now, saying that AI is taking copyrighted work and using it. So creating an image is a great example. If you're asking AI to create an image for you to use outwardly facing for your tourism promotions, that is pretty dangerous territory. 

Mark Lapidus [15:22]: It's equally hot for music. 

Kara Franker [15:23]: Oh, so hot for music, too. That could also create a bunch of lawsuits. So you have to think that there are people that have, not knowingly, put their information out on the Internet, but then AI has scraped all of that and taken it in the training piece of it. AI has taken all of this information, and then the copyright still follows that piece of content, but that's not being communicated through the technology. And so you just have to be so careful. That's why when there's talk about jobs being eliminated, there's so much that humans still need to be at work doing, because we can't completely trust the technology. We can use it, we can harness it, but if we're using it completely or unleashing it without those legal and ethical boundaries, we got problems. 

Mark Lapidus [16:14]: Kara, before we go, I want to hear about your white papers; I hear you've written several of them. 

Kara Franker [16:18]: Yes. Shameless pug here is, VisitEstesPark.com/AI. And that's where you can find the two white papers. One of them is legal and ethical consideration, so a lot of the stuff that we talked about today, but in a deeper purview, and coauthored that with a great attorney in the industry. Her name is Roxanne Steinhoff. Roxanne is a rock star and she has her own law firm. So anybody who wants to dive deeper into that, I would suggest pick up that one. And then the other one is talking about the two white label AI's that we also talked about, with Intentful and then GuideGeek, and basically just like, okay, this was our playbook that we did in Visit Estes Park. Here's what worked, here's what we think that we could improve on; that way, if anybody kind of wants a case study to see if they want to try it or not, they could take a look at what we've done. 

Mark Lapidus [17:06]: Beautiful. What was the URL again? 

Kara Franker [17:09]: VisitEstesPark.com/AI. 

Mark Lapidus [17:11]: Thank you so much for joining me today on the podcast. I hope we can talk about AI further in the future. 

Kara Franker [17:16]: I would love that. Thank you so much for having me. 

Mark Lapidus [17:18]: And that's Brand USA Talks Travel. I'm Mark Lapidus, thanks for listening. 

Outro [17:22]: Your feedback is welcome. Email us at [email protected] or call 202-793-6256. Brand USA Talks Travel is produced by Asher Meerovich, who also composes music and sound. Engineering by Brian Watkins. With extra help from Bernie Lucas, Nthanze Kariuki, and Casey D'Ambra. Please share this podcast with your friends in the travel industry. You may also enjoy many of our archived episodes, which you can find on your favorite podcast platform. Safe travels!

Download Transcript