AI and Industry Acceleration with Lisa Thee on The Cybersecurity Vault podcast

Welcome to Cybersecurity Vault. I'm your host, Matthew Rosenquist, cybersecurity strategist and CISO at Eclipse. Today we're going to talk about artificial intelligence and industry acceleration, and I'm going to be talking with Lisa Thi. Now, Lisa is a top 50 global thought leader for AI, privacy, and safety. She's a consultant to some of the world's most innovative healthcare and global technology companies, including Microsoft, and UCSF's Center for Digital Healthcare Innovation, which works to accelerate FDA approval for AI use in clinical settings. Lisa's also the CEO and co-founder of MineGuard, an artificial intelligence software company focused on making people safer online and in real life. She's also, uh, an author. We're going to be talking about her upcoming book, but we're going to save that towards the end here. And Today's podcast is created in part by Eclipse, securing data in transit through any cloud, network, or device. Welcome, Lisa, and thank you again for coming and talking with us.

Matt, I've been looking forward to, to it. Thank you for having me.

Well, we're going to have some fun. We're— because AI is such a juicy topic, it could be used in so many different ways, and healthcare is one of those areas. I mean, AI is emerging to become just an incredibly powerful tool in so many different areas. And it can be used for amazing good, like in healthcare, but it can also be wielded for malicious purposes. So there's always that two sides of the coin we gotta talk about. So.

Yeah, and if there's anybody that I love to banter about that balance between safety and privacy, You are my favorite thought leader to collaborate with and learn from because frankly, we're on different sides of that coin, right?

We are, but we always seem to come to a balance, right? There's got to be tension in the system because we have to represent both sides or multiple sides of discussions when we talk to— talk about that intersection between the good of technology when it comes to security and privacy and the potential downsides of it. So I always enjoy sitting down and talking with you as well.

So, well, I, I've never walked away without learning something from our conversation, so hopefully our audience has the same experience.

I'm sure they will, and if not, no worries. I'm here just to learn from you, so it's all good. So how did you first become interested in the possibilities, um, around AI? I mean, what really got your teeth sunk into it, dragged you down that rabbit hole?

Great question. So what really got me energized around artificial intelligence was seeing the improvements that were being made in the applications of personalized medicine and precision medicine in healthcare. And back in around 2015, 2016 timeframe, there was some really interesting work that was happening in the academic research space in partnership with the public— the private sector, including Intel, where we were both working at the time. Where they were able to take the human genome and start looking at specific cancer treatments based on the genetics of the person that was receiving them. And I remember one particularly inspirational story where someone had a very short amount of time predicted to live, a father of multiple children, beloved engineer at our company, and they were able to not only extend that time but get him into full remission with more precision medicine treatments based on using the Collaborative Cancer Cloud. And that just got me fascinated around the concept of what happens when you can access regulated datasets to solve really big hairy problems in a privacy-preserving way. So, I got fascinated with federated learning. I got fascinated with artificial intelligence and being able to crank through through just huge amounts of information, really for the human experience of what's possible when you can apply that towards something that benefits the individual or society as a whole.

So when you're saying precision medicine, right, I mean, you know, we've got doctors all over the world, we've got standard, you know, standards of care, you know, in the US, so on and so. What are we talking about with the intersection of AI and precision medicine? Give me some examples here on how is that different, you know, tomorrow with AI versus today with my normal doctor?

So, uh, I think the vision that I am excited to see go forward is this, uh, you know, precision medicine is being able to take your genes, your environment, your, uh, socioeconomic, uh, experiences around you as a person and apply that to your healthcare treatment. Um, I've heard a statistic, I don't have the citation for it, but I've learned it from credible people within healthcare that 80% of outcomes for patients are much more deterministic about where and how they live versus what they're diagnosed with. And so I want to make sure that we are make— we are looking at the whole person in healthcare and that we are creating a frictionless customer experience for people that are already in a very difficult position to be able to recover and have the right to thrive both mentally and physically. The reason I think this is an interesting inflection point today for something like that in the future leveraging AI is You know, when I got involved with this back in 2019 with UCSF, I had learned that there were only 6 novel models that were approved by the FDA for use in clinical settings globally. That's a staggeringly low number.

Yes, it is. You'd think we'd have a little more diversity in the types of tools that we could use and the frameworks to apply them.

Which is accurate in a lot of industries, but in healthcare, being that the data is so sensitive, the industry is fairly risk-averse, they're highly regulated, there's a, there's a lot of concern about how do you safely share data in a privacy-preserving way that allows us to innovate without violating the privacy of the people that are being treated. And so I think with the advances that we've had in cloud computing and confidential computing and federated learning and in the opportunity for AI models to be trained on real data and not synthetic data, we have the ability to reduce a lot of the bias in healthcare. We have the ability to get better treatments for a broader population of people. And we have the ability to accelerate diagnoses for things that are difficult to identify. I don't think that AI is going to solve all the challenges of healthcare, but I do think it could be a catalyst and a strategic advantage for people on the front lines in clinical settings to reduce medical error, to reduce the cost of delivering service, and to be able to let the experts do what they do uniquely well, which is connecting with the patient on a human-to-human level and be able to help those people understand what their options are clearly and what the risks and rewards are. And I think AI can be a much more effective tool in the future of helping patients really understand the choices that they have at the places that, that they're at at that moment and what they can do about it. I think today doctors are saying, you know, well, based on this generalized group of people, this is what we kind of know, and you can make your own decision versus— that's a very different conversation than I've looked at your information based on your genetics, based on the disease that you have. You have a 92% chance of recovering from this if you change your diet and your exercise and you take this medicine. It's very, very different than, you know, you should eat better.

So, you know, it, it opens up an entirely new kind of aspect because again, you know, we go to our doctors and they say eat healthy, exercise, and that's that generic— yes, it's great advice, but it isn't tailored just to me. And with AI being able to start to grab this data that is out there, right, um, uh, genome sequencing for an individual patient, understanding their life choices and their environments and what they do in day in and day out, I mean, that's a tremendous tremendous amount of data that has to then be processed and analyzed and, and so forth. But it's also very sensitive personal data. It can be about the relationships you have, what you do day in and day out, your fam— your his medical history, your family's medical history, um, and maybe even the intersections of people that you meet and ailments that they have or their predispositions. I mean, it's a massive dataset and AI.

The beauty of it is though, Matt, that once you train a model on real information, you can do one-off scenario planning at a clinical setting without exposing— moving any of that data, exposing any of that data. This is really a tool for doctors to be able to provide better recommendations real time. And that's why federated learning is so special is when you can take the information that is sitting at leading academic research hospitals and train models by having the models visit the data versus the data traveling, you can reduce so much of those risks of any kind of data breaches. And you can accelerate all of the research that's happening in the pharmaceutical industry, the original equipment manufacturers for medical devices. You can accelerate other academic research hospitals and startups in healthcare that are thinking about how to improve outcomes for patients in a more frictionless way. In the most cost-effective way. And so when you can see a whole industry get a catalyst for improving outcomes, that's the place that I love to be at that intersection of influential leaders and innovation that are doing things in a safe and secure way. And the reason I know they're safe and secure is because they're building off a lot of the technology that has been developed for things like national security and things that are very highly regulated, sensitive dataset industries. Which is where my origins, roots come from. It's working with federal agencies around disrupting sensitive global topics like human trafficking.

And what's at stake is, right, those individual doctors aren't giving generic information. They're able to then— and a single human wouldn't be able to gather all that information and understand it and apply it, right? You have that global ability for very specific understandings for patient care. And it then becomes scalable with those models and the AI behind it, and very personalized for better outcomes. And I think in my mind, the consistency then also becomes a tremendous benefit because you may have a great doctor that knows you. In fact, I had a doctor for 20 years, fantastic, um, and then he retired and I had to start all over again with somebody who just read a chart, you know, a page and a half on me instead of understanding 20 years of who I am and what I do and what I shouldn't do and everything else and finding those balances, you know. So in my mind, the consistency is also there.

Yeah, and that's a future state. I don't want to imply we're there yet. I, you know, that's the vision of where we could be. I think today what's really exciting is for things like COVID or things like, um, detecting diabetes, we have developed some novel models that are now going through the process of getting FDA approval to be able to You know, have more effective screenings at scale to understand long-term implications of things based on radiological charts, not requiring a doctor to fully know you perfectly. It's really about reducing the barriers for people to have better outcomes. Let me give you an example of how something like that could work, because I think it helped me kind of solidify it. If I know that I'm a woman, middle-aged, and Bangalore, India, with this type of cancer, with this type of genome, AI can help tell me how many other people with my genetics and my gender and my type of cancer that have tried this treatment that have had positive outcomes. I don't need to know their names. I don't need to know where they are. I don't need to know— I need to know from an FDA perspective for approving treatments that this has been tried 10 times and it's succeeded 10 times. And so it may be a really good path forward. You don't need to break privacy and share people's raw information in order to do that, but you can't train models unless you have robust, real information to reflect the patterns of what's happening in the world.

Well, it all comes down to data, right? I mean, you have to have that massive influx of data And de-identified data obviously can help in some areas, but there's also other areas that we do need to know some more personal information, right? I mean, you need to know the patient's genome. You may need to know parents' genomes and behaviors and associate those with the patient. And as we get bigger and better, right? I mean, you had mentioned COVID. It's now, it's not only my data, but as I'm walking around and our phone apps are tracking whoever is next to me, we may need to understand them as well. So there becomes this gray area now where there is de-identified data and there's pseudo-de-identified data and there's de-identified data that you could actually use AI to identify people, you know, and so you start getting in these areas.

That's the point, is de-identified isn't really de-identified anymore. We all know you can reverse engineer most things. And so The excitement that I've seen in this space and that disruption is let's not commingle the data, let's not move the data, let's process the data on chip for very sensitive datasets to train models and then delete everything that happened and just look at the validation report. Just be able to prove that this is working properly for transparency's sake, but never move people's information around. And today that is managed by lawyers, it's managed by research hospitals, it's managed by, you know, protecting IP. And how can we start to leverage technology to make this information more accessible without increasing the risk profile of, you know, having exposures of data leaks? And that's what's exciting to me. We're taking very high-end technology to apply it to a very specific problem. This isn't a generalized solution where it's going to be everything everywhere. So, you know, I'm, I'm always excited to be at the intersection point where benefits to society can be achieved at scale and we can do it in a thoughtful way that balances the privacy and the safety of the people that we're looking to help.

And privacy is always huge in, in healthcare. I mean, that has to be part of it. Um, so—

which is why the industry is a little slower to adopt, right?

We need to get it right because we've gotten it wrong for so long, unfortunately. Um, and that's why we've seen so many data breaches. But again, you know, using the right technologies, uh, you know, applying the right behaviors, instituting the right processes can help mitigate and manage those risks. Now, tell me more about this project with, uh, UCSF. And their digital healthcare innovation. Because I just, you know, I've been a part of UCSF and some activities there for a while. What's going on in this space and how are they using and how are you helping them use AI for better outcomes?

Well, I had the luxury of working alongside Dr. Michael Blum and his team that were focused in on spinning up some IP in this area. And they actually have successfully launched their company Beekeeper AI in 2022. So they are— he has officially taken a sabbatical from his cardiology rounds and teaching at the medical school to be the CEO of this company, as he was the chief digital transformation officer at UCSF going into this. But with that said, I'm really excited to see how do we start blocking and tackling and making the vision real? They've just released a product in Microsoft Azure called Escrow AI that's really exciting around doing exactly what Escrow does. How do you create a place where research institutions and people that want to train their models can come together in a safe and secure way to get validation and training information and learn while reducing the cost and the time to do that? When we got involved, and it takes anywhere between 3 and 5 years per model to get FDA approval for use in a clinical setting. And a lot of that time is literally just lawyers going back and forth trying to figure out how do we collaborate. And what's really tricky about that is until you have the legal agreements in place, you can't even have a sense of the shape and the size of the data to even know if it's useful or not. So people are burning millions of dollars and many years just to get access to see if something might work. And so that's where in the Azure Marketplace, Beekeeper now has a product that allows you to do some of that shaping and sizing upfront and accelerate that timeframe so that we can start to make more high-quality research information from research institutions globally, like the UC system, available to startups, available to original equipment manufacturers to help them to start to think about how do we do this at scale and make better outcomes for patients. Because medical mistakes are third leading in the world for cause of death. Nobody shows up as a doctor, I think, to say, I want to miss the one thing that could have saved a life. I haven't met that doctor yet. And so it's not realistic to expect humans to be able to have the ability to ingest this much information. So having that digital assistant there, alongside the human, 'cause you'll never fully replace the humans, to say, hey, we're in a trauma situation and this particular patient has these factors that tell me that if you don't give a blood transfusion to them in the next 7 minutes, they're gonna bleed out on the table. That was our first use case. There are 2 or 3 different things that you can compute on that tell you that you need to have a blood transfusion ready in a trauma situation. And today it relies on somebody knowing those three things and having it top of mind while they're trying to put you back together after a car accident. Tomorrow at scale, I can see that every person that comes in for trauma surgery automatically is scanned and we don't miss it because you only have a certain number of minutes to react and it takes time for the blood bank to bring the blood to the room. And if you miss a signal, right out the gate, you can't recover. It's those kinds of moments that matter that I think is the promise of AI.

That's amazing. And you know, and you're talking about situations where doctors are still there, they're still providing care, but this becomes an enhancement, right, to make sure there is that consistency and comprehensiveness in what they're doing. And we're all human, we're all fallible, especially, you know, working double shifts and you're in the emergency room and and adrenaline is flowing and people's lives are on the line, having that extra tool that might be able to identify something that's missed or something that's really relevant. Now, you had mentioned all these models coming out of UCSF or the ability to create models in a more rapid fashion. Where is all that test data, right? Where is it all gathered? Who owns it? And how do they make sure that yes, it still is going to fit with this accelerated intention of creating new models that are going to have better outcomes? You know, again, it all comes down to the data. How is, you know, the UCSF team dealing with that? In regards to ownership and privacy and security. Does anybody that join get access to that? I mean, what protections do we have?

Absolutely not. So, uh, the data stewards, as they would like to be called, are typically the research hospitals. They currently have all the protocols for holding that patient data and securing it in their HIPAA-compliant clouds and on-prem infrastructure. And that's the promise of what Beekeeper is doing, is it's not going to move any of those pieces of information. It's going to bring the relevant information in, in a privacy-protected way from both the hardware and the software side of encryption to make sure that none of that is ever exposed to even Beekeeper itself. So I can't see your data, you can't see my model, and I can't, as the person processing that information, I can't even see either one. So it truly is that black box compute environment. That is destroyed upon the end of the computation, but the model that was in there got smarter. The model, you learned whether that model worked or not. That's where we can continue to evolve from. It really isn't just about UCSF innovating, it's about all the research institutions like UCSF being able to make their high-quality data accessible, in a way that does not increase risk of breaches so that the world's innovators can help to solve things like how do we diagnose autoimmune diseases faster and get more effective treatments? How do we identify that someone is gonna bleed out on the table? How do we identify, you know, the one-off cancer patient that if you use chemo on this person, it's going to be, significantly more detrimental than if you try a different treatment. It's those moments. This isn't going to replace healthcare. This is going to enhance the ability for the decision makers and the subject matter experts to have the right information at the right time to guide us.

So is it safe to say AI is not about creating a new tool? It's about being transformative in how we approach healthcare in using a potential series of tools and capabilities for better outcomes.

That's what I hope. That's the vision I want to make a reality. And that's why I feel really inspired to double down on embracing AI for digital healthcare transformation. Just full disclosure, I'm a long COVID survivor. And so I have a lot of things that I have to manage around today that 2 years ago were not relevant in my life. And there's a lot we don't know, including the doctors. And I think that the more that we can leverage AI to bring the right answers at the right time to the front lines and democratize that information, the better patient outcomes are going to be. And I, unfortunately, I'm not alone. There's a lot of people that are living with a new way of living. And, um, I think that the same way that AI has disrupted things like the transportation industry or the fashion industry and made things much more frictionless and fluid and allowed for a lot of innovation, healthcare is ripe for that. But we have to do it in a much more mindful way because it does affect so many people. And so that's why I think having people that can come in from different industries like national security, where you and I have both having done some work in that space from prior places is really relevant because we are very accustomed to bringing together sensitive public and private data to make important decisions. And so we know how to do that in a way that is not irresponsible.

So, the stakes are high, the data is sensitive, but we also have experience, you know, in the world starting to get our arms around this. And, you know, it takes, innovative groups to push industries forward, especially highly regulated industries like healthcare. So, I mean, that's just— it's amazing. And it's fortunate that we've got people like you that are out there, you know, pushing or pulling or doing whatever you need to do to, again, achieve those transformational changes because it's not going to happen by luck or happenstance. It's a matter of a lot of work and dedication and there'll be missteps. But that's how we learn. That's how we get better. And if at the end of the day, we're able to save more children and adults and transform people's lives for better everyday living, it makes a lot of sense that we're investing this brainpower to make it work.

Yeah, for me, Matt, I'm sure you see this as well. Like, this is a team sport. This is not an individual sport. And so I think, you know, bringing thought leaders like yourself that can be like, whoa, whoa, whoa, did you think about this?

Time out, time out. Let's address the problem, not sweep it under the rug. Yes.

The policy folks that are like, have you really thought about how this could be used in other parts of the world that maybe don't have the same human rights standards? Or, you know, the folks in, you know, the insurance industry that have to codify all of this stuff and make sure that we're not creating bigger problems. For providers to be able to do what they need to do. And so, you know, I see myself as a very small piece of a much bigger play. And I think the, the reason I'm energized by it is who doesn't want to spend time with the most diverse set of smartest people in the room where you've earned your seat at the table because you've seen it done somewhere else? That's kind of where I see myself. I've I've cross-pollinated a lot of industries. I've kind of seen what works and what doesn't work. And I can be there to be the, the nudge, keeping everybody moving forward and saying, we're not just going to stop here and say good enough is good enough because lives are on the line and healthcare shouldn't be only for people that can spend their time and money and support system to keep them afloat. Like we need to make it more accessible for everyone. We are at a place where we have the technology to do that.

And you have a tremendous background, right? You know, you've worked in all these different things. How did that base of experience help you as you started moving in and understand the possibilities for AI?

So I think having a background from Intel is a luxury in that, you know, for 18 years of my career, I could pick up a phone and talk to an expert in whatever I was interested in. And, um, usually people are pretty generous when they are inspired by the mission that you want to make real in the world. So I think my corporate days gave me access to people like data scientists and engineers in the cloud, and, you know, just frankly core areas that aren't my area of expertise. So I could bring together the right people, including people like yourself, the cybersecurity experts, right? To say, hey, if we took this, if we look at this societal problem or business problem, how can we all come together to overcome these hurdles? And I think, you know, AI is a good unifier because it's a tool. It's not a solution. There is no one person in the room that can fix the situation. Everybody needs to work together to create a new future. I think through that process, I learned that it's a little bit easier, at least for my temperament and the speed with which I like to work, to innovate in more of a startup environment than in a large corporate culture. And so that led me into entrepreneurship in my 40s. It was an accidental move. That was not my lifelong aspiration. But I saw some problems in the world that I wanted to be solved, and I recognized that there wasn't anyone that was going to solve that from only the nonprofit sector or only the tech companies. It had to be a broader view. And so I could convene those people, which is what led me into entrepreneurship. And then last but not least, I would say that, you know, being an entrepreneur for the last 5 years and being able to be an advisor for many CEOs now, both CEOs of small companies and, you know, VPs and up of large companies, I recognize 99% of the time the teams that need to come together don't understand each other's languages at all. And they're saying the same things in different terminology and totally missing each other. So, if I'm really honest with myself at this point, the role I typically play is translator between the business people, the technical people, the, you know, the different areas of expertise to get them energized and focused on the mission of what's possible and help them get out of their own ways when they get stuck.

That's amazing. There's no one person that knows it all. So bringing all those people to the table and having that translator. Where's the AI for that? We need AI for translation.

Yeah, I think Meta's working on that. I interviewed the head of research for Meta at the World AI Conference 3 months ago in France, and he indicated 3 years from now, their goal is to have a model that would speech-to-speech auto-translate. But I think that's probably only going to be codified languages. I think we have a little while to go before we get into VC jargon. Yeah.

The nuances.

Data science language that is, even though I have an engineering degree, I still don't fully understand. And, you know, you cybersecurity folks that, you know, specialize in million-dollar words that make us all want to invest in things. So, there you go.

So, you know, AI is, you know, it's got an interesting reputation and many people have certain amounts of resistance, right? They're concerned for AI having too much power. Yeah, same hear sometimes. How do you, um, how are you approaching getting over those barriers, that friction, that resistance that people may have, that apprehension when we start talking about AI and now their healthcare, um, and their data, right? And the treatments that are going to be recommended or applied that will be relying upon this magical black box AI system. How do you address that, and how do we need to move forward as a culture?

So, first and foremost, I think a lot of it comes back to design thinking and looking at the user experience and the empathy mapping and the journey mapping of what would it take to create something that makes people feel comfortable and not have an increased risk. And surprisingly enough, as a technology CEO, I'm actually quite risk-averse. My husband loves to make fun of me for being a technology laggard. I got an Apple Watch like 3 weeks ago, and I'm really excited about it. I'm a little late to most parties in terms of the technology. So although I love to be at the beginning of the innovator curve and the early adopters, I'm not one myself. Myself. I like the tried and true. So I think that there is something to be said for looking at the, the problem as a business problem and AI as one of the tools in the tool belt, but not as the solution. That allows you to get over some of those hurdles because we already have real-world controls in place to prevent some of the worst-case scenarios that we're all being vigilant about. And it's how do you apply the technology in a way that doesn't remove those real-world barriers. We are not ripping and replacing and making all decisions based on AI. We're not moving and commingling data that's regulated. We are not— we are operating in a zero trust environment by design because that is the only way that we see going forward to be responsible doing something like this, which means it's not a broad brush stroke. This isn't going to be one piece of software that's developed and sent all over the world and go be the billion-dollar unicorn. This is the way for the absolute experts in their field from multiple different disciplines, including legal and IT and engineering and medicine and all those things, to be able to collaborate across silos in a responsible way that benefits the patient. Uh, this will not be used everywhere for everything. This is the Porsche of a car. This is, you know, this is for Formula 1 driving. This isn't for getting your groceries.

So, you know, we're talking about how privacy is so important, how compartmentalization of data and access is so important to this process. On the flip side, though, how important is transparency and where do we find that optimal balance between privacy and transparency?

I think there's a lot of people smarter than me working on AI explainability and transparency. I can say in this particular first step of an instance, really everything is focused on transparency and verification because that's what the FDA requires in order to approve a model for clinical settings. So the whole goal was to increase the transparency of training to prove that a model will work as well on a woman as a man, on a, you know, a person in Iran as it will in, you know, Columbus, Ohio, that it's not going to only be trained on one population that does not generalize well for the treatment of others in other places. That's the whole point of it, is to be able to demonstrate not on the— on real patient data that this model can be reproducible, predictable, and effective. And that's what the escrow AI capabilities are there to do. It's not to actually create the models. It's not there to start putting other people, people's private information into new places. It's to prove that I have built a model that when I train it on data that's beyond what I have access to, it still performs at the level of acceptable range that it benefits and doesn't deter. And that's what excites me about it. I don't think AI is going to ever take over healthcare and everything's going to be done, but I think it can be deployed strategically, just like it is in businesses, in the places that really can accelerate and, you know, drive innovation in a way that gets us results that we've wanted for decades, if not hundreds of years, but we haven't had the processing power to do.

You know, it sounds like an impossible challenge, and yet we're moving forward. We're bringing the right people into the room. Yeah, no, but it's steps forward. It's progress. And, you know, I've known you for many years, and you're always making those steps forward, whether it was at Intel or CEO of your own company or or, you know, protecting children, privacy, AI, all these different areas. So I know you're writing a book. What's going in this book out of all the things and all the lessons you're learning and all the experiences and not only the successes, but the mistakes as well, because we can learn from those. What's going in this book and when is it coming out?

So just like AI, humans learn much more from their failures than their successes. And so, this book is me processing what I've learned over the last 6 or 7 years of deploying AI in the wild, stepping into the C-suite, leading from a board of directors perspective. And what I learned was that I have a unique ability to over-function and work myself to burnout. It's, really pronounced when I'm working on mission-based things where I have a hard time balancing the problems that I see in the world that I want solved and my own needs as a human. So, uh, the book is called The 90-Day Career Cleanse. Uh, I'm writing it with a co-author, uh, Vicki Mueller-Espinosa, who is an excellent career coach. And a lot of it is really being transparent about my journey towards is, uh, going from burnout to a more sustainable career. Uh, I want to be a marathon runner, not a sprinter when it comes to innovation. And I think that it's a topic that is very timely for a lot of the folks in the middle management and upper management generations that have been working through this future of work situation post-pandemic. I think a lot of teams are really feeling the pressure of burnout, and I wanted to bring some of the knowledge that I learned from transitioning from corporate America into entrepreneurship about how to manage that more effectively so that you can bring your best self to your life and your work. And, you know, a lot of it is learning what your core values are. I walk through how I define those for myself and some of those frameworks. And then having the chutzpah to set and hold boundaries, which frankly, I didn't learn till my 40s. I, I— it's still a work in progress. I'm writing this book to remind myself as I find myself getting sucked into more and more and getting more and more exhausted that only I can fix this problem.. And I know the frameworks with which to use to recharge the batteries and make sure that I bring my best self forward. Um, I've coached multiple executives in this space and been able to improve their health outcomes, their relationship outcomes, and their, by far, their career outcomes by being much more intentional about where they spend their time and what they're here to do and not overusing the muscle that got them there. And that's typically that Type A overachiever, just keep grinding. That grit serves you incredibly well in your teens and twenties and maybe into your thirties. But as the responsibilities of life pile on with family and friends and caregiving and all the things, um, it can be— your strength can be turned into a weakness very quickly.

Yeah, especially with A-type personalities, right? And, you know, in the technology field where things are moving so fast, that energy helps you in the front end, but if you can't manage it, it will burn you out on the back end and definitely shorten your career or maybe even your life, right? And unfortunately, we've seen that in the industry.

So I have been— I have had some low lows where my strength has been overused to the point of working on morphine drips in the hospital. And, you know, I don't really regret that one particular moment because I was prepping to go to South by Southwest to share with the world what we are doing to recover missing children. And in the audience that day was a human trafficking survivor that came up to me afterwards and privately pulled me aside and said, "This happens all the time in the village I grew up in, in southern South America, and nobody ever tells our story and nobody ever cares. And thank you so much for making people know that this is real and that it's affecting hundreds of thousands of people. And you know what, like my 10 minutes I worked on that morphine drip to make sure that my hotel was secured and I got there. I still think I made the right call, but all the calls that I made that led me into the hospital with a back injury to the point where I couldn't walk, maybe not so much. So it's about getting, getting the signal early, just like AI helps us with, right? It's those early signals when it's recoverable that you want to focus on, not the late signals that that ship has sailed. And frankly, I didn't walk comfortably for 12 months after that injury. And I share some of those stories in the book. I want to encourage people through this.

That's scary. Don't do what I've done.

It's not a good plan. And to amplify their courage to grow into their own unique voice as a leader, whether that be inside a company, whether that be running their own company, wherever that looks like for you. Not spending all of your energy playing the game and getting— spend your energy doing something you can do uniquely well. I mean, when I started in this field 20 years back and I wanted to work on things that improve the lives of marginalized women and children, I can't tell you the number of early managers in my career that were like, just go work for a nonprofit. I have an engineering degree. I don't want to work for a nonprofit. I'm not trained to do that. I want to bring innovation to the people that are on the front lines at the nonprofit to be able to do what they do uniquely well, faster, and more effectively. And technology can do that if it chooses to focus on it. You know, sometimes that's the only role I play. I'm the neck. You know, I'm trying to turn the heads to focus on different areas that maybe aren't the first places where people innovate.

Well, AI is one of those technologies that can drive innovation, right? And, you know, hugely important, powerful tool, but there's still a step between, okay, yes, it's a powerful tool, but how do we actually implement it? For meaningful outcomes and that it can be a chasm. And, you know, we need people like you who are out there trying to build the bridges so lots of people can get across.

Yeah, I mean, English is a powerful tool to communicate, but just because you know English doesn't mean you know how to write a novel. Like, AI is the same thing. It's a language.

So what is your advice then? You know, what is the key takeaway for the audience that's listening to this? Who may be in that technology space or wanting to use AI or is concerned about AI, you know, what's that piece of advice that you would pass on?

Yeah. So I think these are my few takeaways. First and foremost is if you feel yourself on the teetering of burnout, recognize that it is okay to, to slow down, reassess, and redefine your career. Number 2, if you are concerned about AI, good, you should be. The question I ask myself as a leader and as an advisor every time of the people that I'm, I'm coaching is, if you build this technology and it gets acquired by someone else and completely misappropriated, Should this exist in the world? And I think that's a question that we all have to ask ourselves every time. Third is that a lot of innovation in, in technology, I mean, just full transparency's sake, gets funded by things like criminal enterprises in the pornography industry. They are early adopters. And so I just encourage people to think about What exists in the world that makes certain things work really, really well that could be applied to places that would benefit the public sector, whether that be your passion about animals or seniors or the environment. I don't care what you care about. I care that if you care about something, you're thinking about how to help people at scale. And that's what technology does. It is not positive or negative. It's a tool that amplifies and builds communities and makes things go faster. And so I would love to see more leaders in technology taking those amazing skill sets that they have for their day jobs that pay their mortgage and maybe using their nights and weekends or, you know, maybe stepping out and leading a company in a place that could really drive systemic change to benefit all of our human experience.

Well, you're definitely leading by example. You follow your passion for as long as I've known you. You follow your passion. You're using technology to build those bridges and get people and organizations and businesses to the place that they need to get to for the greater good. So thank you very much. I look forward to the book coming out soon. When is it coming out?

Yeah, we expect to publish in 2023. So we will keep you posted, and if you want to be kept up to date on when it's available, you can visit my website at lisathiee.com. That's L-I-S-A-T-H-E-E dot com, and that has my TED Talk, it has some of my keynote capabilities on there, and if you ever wanted to connect to talk about some consulting in the areas of digital safety and confidential computing, I lead Launch Consulting's Data for Good sector, and we have a full suite of engineering, project management, and management consultants that can come in and help you take your vision and help you determine where you are today and a map of where you want to go.

Outstanding. And thank you all for watching. Be sure to subscribe and, and go out to that website, LisaThee.com, and take a look and catch all the Cybersecurity Vault episodes. Where we chat with industry leaders to dive into the most relevant and interesting cybersecurity, AI technology perspectives and best practices. Thank you much.