Episode #93. The Need for AI Regulation in Healthcare, with Sandy Dunn, CISO at the Brand Engagement Network Risk Never Sleeps ยท 15 Aug 2024 Source: https://www.podtrac.com/pts/redirect.mp3/pdst.fm/e/traffic.megaphone.fm/NSSMO3611193664.mp3?updated=1723551892 Transcript: https://podcasts.opennash.com/shows/risk-never-sleeps/episode-93-the-need-for-ai-regulation-in-healthcare-with-sandy-dunn-ciso-at-the-cc2ee3/ (via OpenNash Podcasts, machine-generated) Welcome to Risk Never Sleeps, where we meet and get to know the people delivering patient care and protecting patient safety. I'm your host, Ed Gaudet. Welcome to the Risk Never Sleeps podcast in which we learn about the people that are on the front lines delivering and protecting patient care. I'm Ed Gaudet, the host of our program, and today I am pleased to be joined by Sandy Dunn, the CISO at Brand Engagement Network. Did I get that right? You did, Ben. We call it Ben for short. I did see that, but I didn't know if folks would. So, all right. Welcome, Sandy, and welcome, Ben. Yeah, we're excited. And I'm so excited to talk to you. We have so much in common. We both have experience doing security within the healthcare space. We like to talk about AI, and we're both It's risk nerds. Yeah, this is going to be fun. You're not a Grateful Dead fan, are you? I do. I wouldn't say fan, but there is a couple of them that I have on the whole playlist. Awesome. Awesome. I just got back from Vegas. I was at the Sphere. I saw three shows with the Dead and Company. It was pretty cool. If you have a chance to get up to the Sphere, check it out. So first question for listeners, Sandy, tell our listeners a little bit about your organization and your role. So Brand Engagement Network, we actually, we create chatbots. That's our mission. Chat service, helping people understand how to engage with this AI stuff and actually not just make cool pictures. What business problem are we trying to solve and how do we enable humans? And you have one of the hardest jobs in the world, understanding and communicating the risks associated with AI. You've done a lot in this role. And so I'd love to explore, maybe start with your AI threat map, which I did see. And I thought, wow, that's really cool. I'm a CISO Chief Information Security Officer. I got started in cybersecurity before we called it cybersecurity, when it was just like three or four of us in a basement saying, hey, we shouldn't be able to access this stuff this way. And I was at HP at the time. And if you think about it, printers were really the first IoT devices. And so just dug in and studying. And because of that experience, I call people, like I said, we're seasoned professionals. We remember what it was like to not have a web page. We remember when Google came out and improved search because it was so cool. And how exciting that we've experienced this in our careers. We've been able on the front lines going through all of this exciting stuff. But we're also a little bit immune to the hype cycle. Like when things flow up, we're like, it's cool, but it's also dangerous. And when I first saw ChatGPT, I saw the news. Yeah, something came out. And I think even ChatGPT was surprised at how popular it became, how fast it took off. I just watched people playing with it. Yeah, whatever. Like AI, it's over promises, under delivers. and the first time I prompted, I got my chat to BT account and started prompting. I was immediately, I had the two extremes. I was like, oh my God, this is the most amazing thing that I'm a nerd. I love technology. And I was like, this is different. This is nothing like this has ever happened to me before. This is not Windows versus Mac OS. This is very different. And then, so I was amazed because you say, oh, wow, all of those problems that we can't solve that make our jobs a grind. Wow, this is going to be awesome. But also on the other side of that, the terror. Like this, I describe it as a nitroglycerin filled cobra. And I think that's, as we go out trying to communicate with our communities about, should we use it? Absolutely, yes. Is it dangerous? Absolutely. And we get back to what we really like to talk about, which is risk. And as I started to try and figure it out, I came up with my risk map and it started out with four areas and now I think I'm up to eight. But one of the newest ones I added is AI dependency. So I'll use myself as an example. And I'm embarrassed to say this, but I've always been directionally challenged. People come to me and say, I go, how do I get there? And they go south, north, east. I'm like, wait, I need rights, left, street names. Even imagine, like Google Maps to me was a life changer. I just was like, oh my gosh, I'm not lost anymore. This is so awesome. But even before when I wasn't great at it, I at least could remember. I had the skills I needed to go. And I notice now even simple stuff, I'm like, yeah, I'm just not. My brain is full. I'm just not going to remember that. I don't need to. I've got Google Maps. As soon as that shuts down, what do we do? So we have this, and I'm not just, I'm saying we have a technology dependency that is very concerning. Yeah. Yeah, no, that's a really, that's a great point. Look what happened with Change Health breach where most of healthcare shut down. They could not collect revenue for weeks. And this was a several 100 million dollar impact to the industry. And it happened absolutely as multi-billion dollar. And what was amazing is that it's for even the most advanced systems and the most advanced CISOs snuck up on everybody. And it snuck up on everybody because... Oh, no. Again, the people I spoke to, they said we had no idea the concentration that we had with change because change had acquired a bunch of these over time. I got it. And also they were required subsequently. And so it rolled up to United. So people weren't necessarily connecting the dots on that vendor concentration and the potential systemic risk of that. Like you said, the dependency of that one vendor. And so I think that was the big I kept hearing that repeatedly. The surprise folks had for that that that level of concentration. on that one vendor. Log J4. I think the reality, and it really gets back to the kind of the HIPAA conversation and regulation. Like when a lot of a CISO and a cybersecurity team at a healthcare facility, everyone comes in and says, oh, how come healthcare professionals can't get your act together? You're always getting breached. If you look at what drives the needle, the most costly type of incident at a healthcare is any kind of a HIPAA breach. But that data, it could be my name, Sandy Dunn, which is public information, if I lose over 5 ,000 records of people's first and last names, I have to declare that a HIPAA breach. And I think that one of my hypotheses is, it's certainly around this AI regulation, which is if we look at the evolution of technology over the last 40 years, historically, we used laws and regulation and suing people and doing all that kind of stuff to manage, to keep people online, make sure that people weren't negatively impacted. There was no harm. First, then I went through this a long time ago, but I found that a vendor, Lenovo, was actually preventing how they had implemented their certificates actually flawed so that all of the certificate signing didn't work within the Lenovo system. The man in the middle did because they were trying to put all sorts of spyware. And at the time, there was no, you couldn't show harm from that. Yeah, it's bad and they shouldn't do it. But there was no way to really stop them from doing that. And I said, yeah, there was no remedy. And so we're, but I will tell you that I think we're at this tipping point where we're going, wait a minute, how we manage, how we define harm, how we help people and protect them from malicious systems, the system we have with regulation, it can't keep up. The technology moves too fast. It's too complicated. I see everyone talk about regulating bias and fairness, and I posted a little bit about it yesterday with a picture. I saw that, you know. Yeah, and I changed one word, cybersecurity, and the woman was ugly. I'm like, wait a minute, that's not very nice. How do you define bias and how do you measure it? And I'm emotional. In Australia, it's against the law to actually run TV ads of drugs because they think it's unethical. They say, hey, wait a minute, you shouldn't be pushing drugs to people. What am I supposed to do with these? Go in and ask my doctor. And here in the United States, we had the opioid crisis, which was horrendous. And the people were in. So I think there's, for me, I'm very enthusiastic about the future. I think that things have been getting bad for a really long time. Education. We all complain about millennials. And my students were in school. They're bored to death. We make them, we teach the middle. They're not inspired. It's just, and the only thing the teachers know, and again, these are the levers we've put into place as a society, is I'm judged by how well students do. Like, everyone has to pass the test, so I'm going to teach the test. Like, the system is broken. And so I think AI, we get to take a step back, look at technology, look at our relationships with it. What do we want from this stuff? How does this make us better, or how do we move that forward? So not an AI problem. This is a technology and kind of community problem with like, how do we want to grow as a society? Do we want Mark Zuckerberg tracking us on Facebook everywhere? If you watch crime, if you don't watch any real crime shows, it's all online. It's all every vehicle sold after 2022 has Wi-Fi in it and it connects. It always tries to connect. So even if you haven't enabled the Wi-Fi, yes. Do I want criminals caught? Absolutely. Do I want my data out there? Anyone stalking me? Look at air tanks, air tanks, cause it's a great solution. But look at how it introduced stalking through air tanks and stuff. So, yeah, that lots of exciting stuff. Yeah. And it's amazing. I was, like I said, I was at this dead show recently and I was standing next to somebody and this person gave me two bits of information, basically, where they, their first name and where they worked and where they worked was a large, it's like saying Google or Amazon. And so with those two bits of data, I was able to find the person on LinkedIn and find the person on Facebook. And we connected obviously because we had good conversations, but it's scary. Those two bits of information, you can find out and connect. And then you get more bits of information and more bits of information. So you're spot on with that. And it's certainly a scary world. Who's not having night terrors around Terminator? He stays as it relates to AI. I get that. Everyone wants to. And I think that's part of the human condition. Yes. For many years, our minds went to worst case scenarios because it was a safer desire. Like we're we are we're somewhat wired to look for danger, look to be fearful. But when I hear people tell me, oh, the world's terrible, everything. I go, let me give you some statistics. It wasn't until 1970 that women could even have a checking account for me. I think about YouTube today. You can learn anything. And we remember going to the library. Oh, I do. I do. the card or I love anyone. You can do anything. Community. You can find your community. There's so much awesome stuff in the world today. I would tell you, we live in the best time. I wish I could live forever because I think in 20 years, it's going to be amazing. We keep doing things really amazing. But do we have to, where we, there's an opportunity is to educate. If you look at a computer, people can lose their lives. They can lose their homes. We basically gave them a loaded handgun and said, good luck. How do end users protect themselves? I take care of a very large extended family trying to dodge all of the crazy stuff that might happen to them. But think of all those poor people who don't have someone who's looking out for them. It's so difficult. And I I had a family member that just yesterday was scammed out of $500 and in an attack that was pretty well coordinated. And I was blown away because this is not a this is a pretty, pretty intellectual person who should have known better, but they didn't. And it all looked good. It all looked real at all. When I went and I checked it out, I'm like, wow, this is what an unbelievable attack this was. And it's just they're getting worse and worse. And AI is helping us out, but it's also helping out the bad folk as well. It absolutely is. And so part of the groups, I have a lot of different communities I'm part of. And many of them are, we're a lot of us with that kind of red team mentality. I don't even consider my black hat very black. But I wrote about this on LinkedIn. If you think about the outcome, all I want to do is make money. And I don't want to go to jail as a criminal. That's what I want. I want to make money and I don't want to go to jail. So why wouldn't I create some very realistic scenario right before earnings calls, short the stock? Here they can disprove it. All I'm looking for is a change in the market. So I can, and there's no law against deepfakes. That was, there hasn't been. I know people are working on that. So yeah, we definitely have a lot of work to do, but I think my, I'm very optimistic about the future, which I tend to be Sandy Sunshine, but I think we're finally going to have tools or we're going to be have solutions that really help people be better humans and they don't have to be technology experts. Yeah. Let's talk about that because you did mention and I'll take liberty here, but let's talk about deepening human connections through AI. Okay. Yeah. Because there was a post on LinkedIn about this. And so how do we deepen human connections to AI? I think it's just, I think it's like this. I, the right now, the communities that I belong to, we meet typically weekly or bi-weekly. And it's because we're excited about something. And it really is about people sharing stuff. And my view on it is I think there's too much, not even just too much bad stuff, but there's too much good stuff. Who are the AI I follow experts I should follow. There's so many out there. There's so much noise. How do I find signal? And I think what in my experience has been how I do that is I find the people I trust. I look at, I follow them. They get recommended. I have my trust. Oh, I like, oh, you know that. Okay. Okay. That makes sense. That makes sense. Then all of a sudden I'm following him. We're going. So I think that there, that's been my experience and that's what I hope for other people. I think that just, I see people talk about email, you know, it's our favorite thing to hate. Okay, my chat bot is going to search, write an email to your email. I'm like, hey, wait a minute, maybe email's the problem. There's a lot of people proposing solutions out there that you're going, that doesn't even make sense. Like maybe we just stop sending email. Maybe that's the problem. And I would tell you it's frontier. I would actually argue that all of technology, I would call it frontier. Like none of it's fully baked yet. We keep moving too fast. Yeah, we're only, what, 60 years into this journey. And truly, the last 40 years have really been where all the innovations happen. You mentioned community. I know you're part of the OWASP and the top 10 for large language models. So tell listeners that may not be familiar with OWASP what it is, how to get involved, and then specifically, what's the top 10 group for LLM? I don't even remember what it stands for Open Web Application. It really started out as Web Application. But from my perspective, whatever you're doing, it's got a web interface. It really is pretty. If you have a question, there's a cheat sheet up there on OWASP for it. So it's a nonprofit worldwide. Worldwide, I call it non-denominational. And it's the only organization that really is like that. We get to be that layer between every standards body, every type of different organization that's out there. We talk to NIST. We talk to the people in the UK. We talk to the people in Japan. And everyone contributes their time to it. And just jump in, pick a project. Anyone can stand up a project. I think they're starting to put a little bit more rigor around that, but just jump in and start contributing. So the OWASP top 10 really was started by Steve Wilson, and it was right after ChatGBT, so early 2023. He stood up the community, and it was about the same time that I was like, oh, this is scary, interesting stuff. And so I started lurking in the Slack channel and everyone kept talking about vulnerabilities. And I put on my CISO hat and I went through my OODA loop and I went, okay, I'm a CISO. The business is trying to bring this in. What should I be scared about? What's the risk? What's the threat here? And so I started out with my mind map. I started trying to map it out. Then I came up with a cheat sheet, basically a checklist where I was like, OK, this is for my brothers and sisters in the cybersecurity, the CISO community. Here, start here. Here's signal in the noise. And I just jumped in and it's been an amazingly positive experience. Excellent. Excellent. So how did you get into the industry? What's your origin story? Oh, that's really funny. The truth is, I was at a yard sale way back in the 90s and found some computer books and got intrigued. And then, yeah, so basically self-taught. It was right. I have an old picture of me at Micron PC where I'm selling. They used it at the back of my head for an ad. And it's for a Netframe 9200. It's like 1995, I think. We're bragging because it's got 256 megabytes of RAM and an eight gig hard drive. And you see the back of my head. It's just been so fun. Like, it's been such a sense of wonder. And I've always, as most of us who make a career in cybersecurity, just driven by curiosity. Now, and I just, I'm not afraid to ask a question or say, wait a minute, I don't get that. That doesn't make sense. That's fantastic. I love that. And when you think about the last couple of years with the pandemic and everything, what are you most personally or professionally proud of? Oh, I'm not really. I'm such a I have a ton of I've had a lot of fun and I'm proud that I'm happy. I won't say I'm proud, but I'm really happy about the communities and the contributions and things that I've been able to give back and engage with people who sharpen my skills and challenge me. Yeah, I love that. I love that part about working in technology and working in healthcare in particular, that shared mission that what has when working in healthcare is unlike any other industry. Outside of this job that you're currently in, what are you most passionate about? What would you be doing if you weren't doing this work? I love the AI stuff. I've just had a ton of fun. I was talking to my husband about it because as a cybersecurity professional, as a CISO, no one's ever happy when you walk in a room. Like you never come in with good news. That's weird. Great. You know, just try to get everyone to obviously passionate and try to get help everyone understand and be business focused and everything. But you're not the fun person at the party. And the one thing about AI and the way that I compare, I said, I've been selling hospital visits. And this is the fun part. Like I was at the, my daughter was at the college rodeo finals and I was just walking around the booth and there was a lady there that had awesome stuff. And I said, where's your website? And she goes, I don't have one. And so I pulled that up and I black magicked an image for her future website. And the sense of wonder on her face was just incredible. And it's so fun to bring that to people, to talk to it. people who I don't consider technology positive, when they get with ChatGPT and show them what's amazing, it's really fun to get them excited about it. Yeah, it's amazing. I had the same experience of wonder when I started to engage in it about a year or so ago. And I've seen a bunch of things over the course of my career, including the birth of the internet and the web and all, into the pc and i carried around they called it the sewing machine the compact 386 machine if you remember those things and yeah and when it happened i was just like whoa this is bigger than the internet i think this might be bigger than the internet yeah it's hard to because the internet was big yes i think that it was slower like much slower yes but and i first we talk about it and Let's also put in perspective, like AI has been a concept since 1940. From the very beginning, people were saying, how do we make a machine that thinks like a human? This has really been part of it the whole time. What's fascinating when you dig into it is that we understand humans because we understand AI. There's always been this dual relationship with, as we try to understand our brains a little bit more. I'm reading a fascinating book right now just on the history of intelligence written by a neuroscientist. And he goes way down in the biology weeds. But it's so fascinating because he starts talking about the different ways our brain works and how our eyes. You think about an eye that we can see something, but then look at it from above or below or at a different angle. And our brain still understands. That's not even you. our animals, they can't do that. And bees have different brains and do different things and do others. I just think, I agree with you. I think that this is big. It's all the frontier, lots of danger, but lots of opportunity. Yeah. Yeah. Any hobbies that you have? Oh, I'm a horse girl. Ah, there you go. All right. Yeah. We rodeo. I'm a horseaholic. That's the, everyone said, tell me about your career path. And it all might, it always starts with, I needed more hay. So I had to get that job promotion. I haven't heard that answer yet. So that's good. And when I look at my curiosity and what I love, which is training and understanding what makes them work and how they think, and this bit does this and this does this, it all comes back to what I love about technology and what I love about AI. It's the system thinking, What makes it all the levers and wheels work together? Exactly. Yeah. If you could go back in time, what would you tell your 20-year-old self? Don't marry me that first guy. No, I'm just kidding. I'm kidding. I hope he's not listening. I know. Guaranteed he's not. Yeah. Just because he's handsome. No, that's not a good decision. These are dangerous conversations because I think it all of you, I'm sure this is a Black Mirror episode or it's certainly what is that the Christmas movie we all like so much at Christmas where he changes one thing I wish I'd never been born yeah yeah so I I don't no regrets I love I think we have the coolest jobs in the whole wide world I love doing what we do and it's exciting to be part of it and I really appreciate you inviting me to come talk No, that's terrific. A couple last questions, if you have the time. Sure. This is the Risk Never Sleeps podcast, so I have to ask you this question. Riskiest thing you've ever done. Oh, that's a tough one because I am a little bit of a danger addict. Oh, good. Most people don't even have a choice, so they're like, oh, I don't know. I'm not really risky, so I don't take any risks. Oh, yeah. I like that adrenaline taste in your mouth. Probably a horse I shouldn't have got on. My husband was rodeoed forever, and that's a lot of driving, and we'd always switch drivers going down the road doing 80 switch drivers with big horse trailer and that's a little bit risky. That's a little risky. I do that now. I was just driving through Vegas in a horse trailer probably thinking about too much. My 20-year-old leaned over and said mom over a hundred's bedroll. I'm like speed up. That's a good kid. You've done well parenting. yeah exactly so i i like playing the game and living a fun life so probably the riskiest thing is probably drive too fast or get on a bronchi horse i probably shouldn't have if you're on a desert island you can only bring five records or movies what would they be i'm not okay this is my husband is a music person i'm not i'm really a music person so for me it would probably be books. Oh, okay. Good. Yeah. But I don't have five at the top of mine. I favor it and I'm not going to remember the name, but I'll get back to you with my five favorite books. Fair enough. Fair enough. Last question. Any advice to someone coming into IT or cyber, maybe out of school or with a couple of years experience wants to break into the stream? Yes, use the tools. Be self-motivated. Don't wait for someone to explain to you where your career should go in 20 years or what you should be doing. Listen, watch. Who are the successful people that you know that you would like to do? They look like they're having fun. They're doing cool stuff. How did they do it? You don't necessarily, we always talk about mentors. Mentors doesn't mean they have to take you to lunch. You can watch someone and say, okay, they seem to have their act together. What are they doing? So that's what being very proactive. I teach at our college and that's one of the things that I found really frustrating was that they just, there was none of that hacker mentality. Like when we were early, like no one told us how to hack. We just broke stuff. Yeah. Interesting. Yeah. That's really great advice. Sandy, I appreciate your time. Thank you for joining the program. This is Ed Gaudet from the Risk Never Sleeps program. And if you are on the front lines delivering patient care and protecting patient safety, remember to stay vigilant because Risk never sleeps. Thanks for listening to Risk Never Sleeps. For the show notes, resources, and more information and how to transform the protection of patient safety, visit us at censinet.com. That's C-E -N-S -I-N -E-T dot com. I'm your host, Ed Gaudet. And until next time, stay vigilant because risk never sleeps.