Responsible AI in Healthcare: Ethical Leadership, AI Governance & the Ways of Working

Asha Mahesh: The passion that drives me is anything that I do is getting one step closer to bringing the right treatments to the patients.

Chris Hutchins: The tagline for my company is humanizing AI for care.

Asha Mahesh: For me, ethics and responsibility is not something you write a paper and forget about that. It's not about a paper, it's not about some writing or anything. It's about building the ways of working.

Chris Hutchins: Super excited to come to the event. I mean, for people who are watching, we're at Flight Hollywood in Las Vegas at an AI conference called Put Data First. And I'm talking to Asha Mahesh, who's an expert in ethics, privacy, all the things that make people really nervous, particularly in healthcare, when it comes to the innovations that we're coming out with and AI in particular. There's a bunch of different facets I'd love to chat with you about. But maybe tell me a little bit about, you know, what's your passion? What kind of led you to get into ethics and privacy? And, you know, obviously in technology, that's a little bit more um technical, more so than even philosophical.

Asha Mahesh: True. Yeah, yeah. And no, that's true. I mean, my passion is I'm a uh technologist at heart. At the same time, I'm also very practical. Right. In the sense, uh, in the sense like, yeah, technology is it's if if you don't apply the technology the right way, it can end up being really bad, especially in in an industry like ours, like life science. It's it's all the more important to apply the technology the right way, with the right uh k giving the giving the giving it to the right people at the right time is it's very evident i that's really critical to get to the success of any technology that we use. The the passion that drives me uh is anything that I do is one getting one step closer to bringing the right treatment to the patients. As you see, I mean, we're all we're not immune to health problems. I mean, it's it's across the board. We all have our families, the clo friends, all you deal with them on a day-to-day basis. When you think of your work is actually making someone feel better, it it's all the more important. Like whatever we do, like that that that's that's that's what drives me, that's the passion. And sometimes like I end up going above and beyond and doing things because ultimately my goal is like it may not really look like like, oh, I'm I'm actually treating the patient or anything, but uh ultimately, but how it's you're doing something that is getting closer and closer to that that big majority.

Chris Hutchins: That's one of the coolest things about the you know, do dealing with data and AI. I grew up uh in a household where my mom worked in a in a hospital um and in the radiology department. Um, you know, my dad did mailing systems and database types of things. Um but net at that point in time when I was young, I never imagined that I could do work that could have any kind of impact in direct or indirect with data on healthcare. But I grew up around healthcare people. Um I was always inspired by it. But I find it remarkable that fast forward a couple of decades and all of a sudden, jobs that didn't exist, no one even heard of them. People like us can build careers and we can actually have direct impact. I think that's one of my favorite things about the kind of work that we're doing. Um, I I think what really gets me excited is is to hear people with your passion that are really responsible and looking at the ethics component of it. Um, and you know, my the the tagline for my company is humanizing AI for care. And it's may sound cliche to people, but the passion that uh is driving you is familiar to me because it really is about human beings. It's about relationship. We already mentioned, you know, the the people in your life. Um I I think there's probably an I would say it's probably about the most critical thing we could do right now is to get people focused and coming to places like this, not only to learn, but to actually lean in and influence the direction that technology is being taken in to make sure it doesn't lose that humanity component of it. Talk to a little about you know how you think about influencing um people as they're starting to think about design. And sometimes they probably are not starting where you think that they should start. I mean, what are some of the things that you see that we should be doing and we can kind of challenge people to lean in on to make sure that we're not losing track of why we're doing what we're doing? It's really about making people's lives better. It's not about technology just because it's cool.

Asha Mahesh: Yeah, no, that's that's a great point. Uh ultimately, that that's where uh you look at what is the intended use of the technology, right? Yeah, what is your intentions? Is it purposeful? So yet in terms of influencing to your point, uh what you mentioned about humanizing, yeah, that that that's it, that's it really a concept we all have to, I would say, embed in ourselves. Sometimes we're so passionate about technology, the data, and all those things. So there is no North Star, right? We're all we're all running towards like it's it's pretty cool. I want to do that. But however, when you have that North Star, and then to your point, uh where where is that ultimately going? Who who what is the impact? Who is gonna impact? So uh he he need you need to figure out a way to inspire people on that one. Yeah, you're right. How do you influence? Influence is showing that that North Star. Right. And and that that's something that we do uh in a company like ours, is we have something called cradle. Uh, what that is, is it's it's it's something writing on the wall, but at the same time, they've done a great job in terms of uh, I would say, embedding that cradle values in everybody and every employee, saying, like, yeah, our first, I would say there are there are a lot of things on that one, but basically, in a nutshell, it says we are here to serve our patients, our uh uh the providers, and who are the whole the our customers. That's what that's what we are here for. And one other thing also we do well is in terms of bringing heart and science. Yes. The purely the science by itself will not really solve all the problems. You need to also bring your heart to that whatever you do. That way you when you when you combine science and heart, that's when things actually happen. That's when it becomes more meaningful and uh impactful to your point. How do you influence uh people is also show the, I would say, success and the value. What has done. I mean, if you really look at uh a lot of things that we have done, we have applied AI in in terms of um in terms of uh for I'll I'll give you an example, right? During the pandemic, right, uh we had a vaccination program, which we are we are working on, we are all working on that vaccination, COVID vaccination. So we actually uh used a lot of data and also uh AI, machine learning, whatnot, all those things. We built a lot of models in terms of uh predicting where where we want to run the clinical trials in. So we we our predictions were so good, so accurate. We were able to finish the clinical trial ahead of time, it's also less patients enrolled into the track. So that that that's that's but by itself is amazing in in terms of like, I mean, getting getting even one day closer, like one day earlier, it's it makes a difference. In especially when I when you're running against the time. That's right. Those are some of the I would say examples. There are some grad great examples. I mean, when you look at that and say, yeah, we were able to do it successfully. Why? Because we were all so passionate. Yes. We we felt like we gotta do something. I mean, this is not the way to live for people, right? You know you need to get there. I mean, that that that's that's when you look at those examples and use them to influence people, right? Uh, in terms of that. And also you mentioned about the how do you do responsibly, ethically, and all those things. That's not something like you can forget about. For me, ethics and responsibility is not something you write a paper and forget about that. It's not about a paper, it's not about some writing or anything. It's about building the the ways of working using the ethics and responsibility. So if you if you build that culture, it's all about culture again. Building that culture into your uh organization and how how and how to do that, and also inspiring and also I would say recognizing people who do that, right? Who actually actually apply that on a day-to-day basis. That's one way of the question.

Chris Hutchins: I I love the example you gave between dur during a pandemic where I I was actually working on the health side health system side of that, uh, actually in New York, and I have never seen such good in people that I saw during the pandemic to your your point. It was a passion. I mean, is that it was a shared one because we're all we all felt an immense responsibility because it was a crisis of like we'd never seen before. And I don't think people probably have an understanding of how little we knew at the outset, which makes what you'd accomplished developing a vaccine even that more remarkable. Um we didn't we thought it looked like pneumonia, we thought it looked like flu, we thought a lot of things. Um, if we had applied any models that we had at that point, we would have misdiagnosed. Exactly. Right. Uh but the fact that people rallied around it, I mean, in even in my own health system, the thing that was really remarkable to me is I saw executives responsible for marketing that were showing up to go help set up the tents and administer the COVID tests. I mean, it was an all-hands-on-deck type of thing. And I really saw the good of humanity, the humanity on a whole different level. Um, but I really wanted to bring attention to the fact that you you mentioned how much was done and so quickly because people came together with a with a mission, it was passion, and it was about people. It was not about technology.

Asha Mahesh: No, it's not about technology. We were all using technology and applying. At the same time, it's all about our goal was to get the treatment fast, get that vaccine to the people, that way they can lead a normal life. So to your point, it was the people came together, and there wasn't a single day anyone complained that we were working rightly, working late, working long hours, and not a complaint. The people were like l literally happy to do that. It's like, to be honest with you, people who weren't even savvy technologically, and they cut they come to us and say, like, how can I help? Can I can I do the the data curation? Can I do this? Can I do annotation? Can I do anything? Like uh th that was really truly truly inspiring. I wish we could bring that same culture without a pandemic.

Chris Hutchins: Yes. I mean, I mean, for for once, could we actually hold on to something that was really, really good and and keep it because it is good? I I think you know, one of the challenges we're we deal with, you know, particularly where we're trying to introduce new technology is the trust factor. And what I hear more about in terms of trust is not really where I think our biggest challenge is, but there it's more on can we trust AI. But what I've come to understand that I hadn't really thought about this, but I was having recording an episode with with a gentleman who's a clinical psychologist, his name's Dr. Larry Kuhn. And he's been you know working with executives for a long, long time and coaching them and leadership and mentoring and and things. And he had just published a paper about the erosion of trust. And uh he put some stats behind things that I think we all kind of intuitively can feel and we know. But the example is he said about 20 years ago, if you uh surveyed 10 people, uh, eight of them probably would say that they trust the government. Or maybe the same number would say they trust uh clergy or they trust uh the law enforcement or they trust the CEO or whatever. Trust has eroded to a point now we're talking about the low 20s and in many of those cases. And so we have a much bigger challenge ahead of us because we people just don't trust each other very much anymore. And so the things that we're talking about that can have such life-changing impact, life-saving impact, how do we start to attack this trust issue that we know we're working from a deficit and to start to engage people in a way that they feel like they can lean in and start to trust? Because the the the biggest thing I'm sure you hear about all the time is fear. And if you wouldn't mind, I'd love to hear you know your your thoughts on how we can really address the fear. And more importantly, when someone like you is uh having an opportunity to to talk to people, I want them to hear what motivates you and what they can do. Because all of us get inspired at some point, which changes the trajectory of our life and what we really commit ourselves to do. Then I would love for people to hear from you what what you feel like is really important, what's meaningful, and what can they do, and how do we help them to get past the fear?

Asha Mahesh: I yeah, when you say fear, is it the fear of uh I would say using AI or is it building the trust? It's gonna take my job.

Chris Hutchins: My job.

Asha Mahesh: It's gonna take my job. Yes, yeah, yeah, yeah. True. True. That's true. That that fear is there. I think what we I mean, what at least I do when I go talk to, for example, when I go talk to a scientist, uh, saying, oh, yeah, we are bringing this AI. Correct. That's gonna do the next, I would say, ta uh what we call the they can design a molecule or whatever it is. I mean, there are scientists who that's their life and blood, right? I mean, they they do experiment, they do, they studied biology for so many years, even the clinicians and all those things. So uh I think you need to go with an attitude of saying how it is gonna help you. So, I mean, I I use I I use the the framework called what's in it for you, what's in it for me. That's the framework I usually apply.

Chris Hutchins: Most of us can relate to that.

Asha Mahesh: Right, right. It can relate to that, right? Uh in terms of like go with an uh go with uh the the messaging around like this is a this okay, this is an AI. You you it it is gonna do so-and-so for your job. Right. But but uh to focus on how it is helping them, right? Not not not something like, okay, this is gonna replace you. I mean, that's a different story. But when they say it is gonna help me do my job better, they will be receptive to that thing. That's something that has worked for me over the time. And also when I go to them, uh I go with the attitude is can I help you with something? Like what what can I do to help you? What is in it and also make sure that you're focusing on what is in it for them. It's it's that that that actually helps them get the fear out. So one thing I mean, I also this is an example I want to give you. I was working with a clinical development leader who have happens to be an MD, they have a lot of MD PhD, they have a lot of, I would say, knowledge in whatever this is, and we are bringing this system that would say, oh, it's gonna answer all these clinical questions and all those things, and we're gonna roll that out. Uh one thing I I I went and assured, like, there is no way this system is gonna do what you can do, because you you've gone to school. Like, how many years it is? Ten years of school with all this practice that you have. There is no way any anything this can do even like half of what you can do, or that that knowledge. But at the same time, it's gonna, I would say, eliminate the tedious work that someone would do. So go go, I would say I would go with that attitude of like how it is gonna helping you, what's what it's gonna do for you, versus like saying, yeah, it's gonna do like magical things. So people have that, uh, I would say, have that fear, right? I mean, uh, yeah, it's i it's it's true with all of us. Like that that's gonna replace our jobs, but how are the more and more that you look at that, it's it's yeah, but but but but the way I I I approach it is that that that I would say that that's the approach I take. Than saying, like, yeah, it's gonna do something like magically, it's gonna solve all your problems, which is not the reality to to begin with. I think people understand that.

Chris Hutchins: So yeah, I I I think you're right, but uh that there's definitely something to that in really making it personal for them. In fact, what what is in it for them? I love that because you're you're you're putting it back into the context of, hey, this really is about how we can help you, how well how this technology can come along and support what you need to do and make things easier for you. Um there's a a conversation I've been having uh a little bit recently, um, and it's about how much uh human like we should make AI. And that's initially I was really thought, well, it's actually good in some ways if we can kind of emulate certain things so that, you know, for example, if someone's having a really bad day and they're running your doctor's office and you call them and they're stressed, they might be short, right? So having someone who's even keeled, I mean, I think that's an interesting use of AI. But the other thing that concerns me even more is people have come to trust technology so much so that they're not aware of the risks that they take now. Um the things that they'll put online that they probably should never put online um because they're just comfortable. It's just become second nature. But it's only in what 19, the early 1990s that that was even a possibility. Um, but so for me, I think there's we've got to have some guardrails, I think. Not quite sure where to draw the where where do you draw the line, but um how much do we really want it to feel human? Because it people can get way too comfortable way too fast. They've already done that. And I think we've really not done a great job making it feel human.

Asha Mahesh: Yeah, yeah. I mean, I I know, I know what it yeah, how in terms of humanizing and all those things, like we also talk about uh like every solution or every product that we design, like a yeah, we we want to bring in like a human aspect, like human-centered design. Is it gonna resonate with a human at the end of the day, right? That that's what we do. Uh uh in terms of that, uh, I would say, yeah, but uh there still needs to be a human touch. I mean, you can humanize everything at the same time. Uh one example I would give is let's say a patient goes to a hospital and yeah, there are many times you can have in a mostly automated robot treating you, whatnot, and all those things. But ultimately, you know what a patient we all see the placebo effect, right? Placebo is effect is a is a it's a real thing. Like ultimately what really matters is someone is like liquid, holding their hand and saying, you're gonna be okay. Right? That carries a long way than anything else that you can do at that point in time. I I I think that going to that level, uh I I would say there I f I feel like that yeah, but you no matter what you do within a machine or a in in in in a way, but that that human touch still needs to be there. In in in a very strategic point, I'm not talking about like going in entertainment or something, you're playing a video game. Yeah, great, that's fine. But when it comes to the the critical aspects of humans, whether it's health, are are in a you're you're in a you're in a situation where uh you actually need a little bit of a support. Someone holding your hand and saying those are the aspects I don't think we can replace with an AI. I mean that that still will be there. So one other other also uh one thing um I want to mention, one example is when we design these AI-based diagnostics and also clinical decision support, all those things. So we always say human in the loop. There is a clinician in the loop, they're the ones looking at the output and making a decision. But how are the regulators? They come back and say, you know what? If you a human can get complacent at some point, they will start trusting that. They'll try they they if that makes a mistake, they will know if it's a mistake. They're gonna start following what kind of guard rights you're gonna put in place in order for humans to even to re do the critical thinking and see what what's right, what's not right. So those are the aspects that we have to consider when we do it that the level of clinical decision support and the diagnostics type of.

Chris Hutchins: I agree with you, I appreciate that so much. Well, Asha, I can't thank you enough for for joining me. It's been a pleasure chatting with you, and uh I hope that we can stay in touch. I'd love to dig into more of this. I think as things evolve, there's gonna be a lot more to talk about in this space.

Asha Mahesh: Thanks for having me.

Chris Hutchins: Thanks for the That's it for this episode of the Signal Room. If today's conversation sparks something in you, an idea, a challenge, or a perspective worth amplifying, I'd love to hear from you. Message me on LinkedIn or visit Signal Room Podcast dot com to explore being a guest on an upcoming episode. Until next time, stay tuned, stay curious, and stay human.