Episode Transcript
[00:00:00] Speaker A: Welcome to Real Medicine, real lives. I'm Dr. Yasser Sambal, and together we're showing medical expertise in a human way.
Hello, everybody, and welcome to Real Medicine, Real Lives. I'm your host, Dr. Yasser Sambal, and on this show, we're going to explore how breakthroughs in science and technology are transforming patient care in real and human ways. Today, we're diving into something that on every healthcare leader's radar, but still feels mysterious to many.
Artificial intelligence.
Joining us today is one of the country's leading voices in emerging technology, Dr. Alan Badot. He's the CEO of Alan Bideau LLC and founder of Harmonic AI.
20 years of years of advanced experience in technology, including AI, cybersecurity and quantum systems. Dr. Bourdot has led massive transformations for commercial government clients alike. Today, he's helping us understand how AI really works and what it means. For your Next Doctor, visit. Dr. Bideau, welcome to the show.
[00:00:55] Speaker B: Thanks. It's great to be here.
[00:00:57] Speaker A: So this is a very, very interesting topic. Obviously, AI is very up and coming and, you know, it's just progressing day by day and clearly it's really been involved in healthcare a lot. People are using it to research, people are using it to transcribe notes, all sorts of things to try to simplify. I think it's really simplified our life in a way, although I think it's kind of dumbed us a little bit, but at the same time it simplified us. So maybe we should start by just, you know, talking about AI in healthcare and maybe you can give people a kind of an idea, in simple words, how, how AI is involved in healthcare nowadays.
[00:01:34] Speaker B: Yeah. So the challenge with, with AI and if you think about the, the involvement in healthcare, is that, you know, you've got folks that see things on ChatGPT and, you know, these other mediums and they treat it like Google and they are asking it for medical diagnoses. Right. It's really, if you think about AI just in general, it's. It's like a super smart, helpful, you know, assistant. It's read thousands of medical books. It has complex mathematical equations and it's got data, it's got the ability to recognize certain patterns and features. And, you know, that's honestly what makes it a good tool for the medical profession, if it's used the right way.
Gotcha.
[00:02:26] Speaker A: And so can you give us a story, for example? That's something that you witnessed that AI has helped the patient get better or helped with their career.
[00:02:36] Speaker B: Yeah. The biggest thing is really around, you know, the image scans, you Know, whether it's an X ray or whether it's, you know, CT scans or things like that, you know, its ability to take millions and millions of those data points and put them all together and see patterns that, you know, sometimes are more difficult for, you know, physicians. Or I would just say, you know, it can be borderline based on experience. Right. That, you know, AI is able to capture and to pick up. I mean, there's so many different cases of potential, oh, mysterious illnesses or maybe a child is presenting, you know, symptoms that are not typical of a child. For example, that AI may be able to find use cases. It may be able to find things that are very similar that allow it to more quickly, you know, diagnose, you know, an illness that may take a little bit longer for a physician to do.
Gotcha.
[00:03:38] Speaker A: Okay. And so, you know, AI is obviously evolving, and although it's a computer and it's pretty smart, but, you know, how does AI, for example, learn not to make mistakes on people? Because it can make a lot of mistakes and you kind of have to, you know, just in my experience when using it, you got to have to kind of some stepwise fashion to get to where you need to be with AI.
[00:04:00] Speaker B: Yeah, that's exactly right, because it is super smart, but it's also super dumb. And a lot of people, you know, forget about that, that aspect of it. Especially, you know, especially, you know, some patients may believe everything that the AI says. Right. And, you know, a good example is, you know, it's. It's training that's fundamentally what it's going to, how it's going to perform.
You know, as AI is learning, as it's taking in more information and more data, the biggest impact to those are going to be those, those edge cases, meaning, oh, somebody. We're not quite sure if they're this or if they have this. We're not sure if we should treat it with this or we should treat it with that. You know, AI is going to get the easy stuff. More, more than likely it can make mistakes. But those edge cases where it's not quite sure, I mean, you know, you guys would call them the hard calls or, you know, maybe a, you know, a patient, maybe a train wreck or something like that. Those are the cases that AI really needs to learn in order to really, you know, be. Be of greater assistance. And in some cases, it's just not there. And that's why you have to have a physician in, in the loop making those, those right calls. Because, you know, in some cases, AI just may get it, get it wrong or they may get lucky.
[00:05:19] Speaker A: It's interesting you said that. You know, I've read, I've read tweets from Bill Gates, for example, that says, you know, in about 20 years, AI is going to replace physicians in general and we'll no longer need physicians, but artificial intelligence is going to take care of patients. Kind of like the theory that robots are going to replace human workers. How do you feel about that and what's your thoughts about that as somebody with an expert in artificial intelligence? And, and the, the, the notion that artificial intelligence and what we can do with it is actually going to replace humans?
[00:05:50] Speaker B: I, with all due respect to Mr. Gates, I disagree from that perspective because the reality is AI is not anywhere close to that in 20 years. There's a lot that can happen in 20 years, but the realities are that having a trained physician in the loop who makes those hard judgment calls is going to not go away for a very long time, because, you know, AI is only going to be as good as its data. And we have seen that these large language models are trained on millions and billions of pieces of information, and those pieces of information can still be correlated. Wrong. They can put, to be put together and give you the wrong answer. And so I don't see that changing for a very long time. It's always going to be fundamental from my perspective. And if I was a patient, that if I have some sort of disease or some sort of cancer or whatever that is, I want a physician working with the AI, that's fine, but I want a physician to make that call. I don't want the AI to only make that call.
[00:06:59] Speaker A: Yeah, I mean, I kind of, I completely agree with you, and I feel the same way. I always tell people when they ask me that because I'm a cardiologist and, you know, people say that to me and I say, I just think there's a human aspect to medicine that artificial intelligence can't compensate for.
[00:07:16] Speaker B: Exactly.
[00:07:16] Speaker A: You know, a patient is crying in your office. Just the notion of reaching for a tissue for them or holding their hand or providing comforting words just really doesn't, doesn't fit with, you know, artificial intelligence. They may be able to give that to you, but there's an emotional aspect that you can never teach. AI.
[00:07:36] Speaker B: Yeah, yeah, I agree. And we build what we call cognitive agents, and they have human traits. They behave the same. Similar, similar. They can answer it from different perspectives. But you're exactly right. At the, at the, at the end of the day, the reality is, is that you Want to have a human to human interaction when you're in a very, well, you know, scary environment in some cases. Right. Just because you don't know what's going to happen. And having an AI tell you is going to be very different than having a human tell you. And I don't think, even if it could, I don't think we would still want to, want to have that done.
[00:08:17] Speaker A: Right.
So, so what can everyday families, when they're going to a doctor's visit, how is AI helpful or how they can use it to be prepared when they go see a physician?
[00:08:29] Speaker B: What I tell people is use it to ask questions. What sort of questions can I ask the physician to respond to? Because you've seen it probably as I know we have as a cardiologist, people will come in, they're intimidated, they're not quite sure what to ask because they don't want to ask a dumb question.
Use AI to ask questions. Don't use AI to diagnose. I'm, I'm probably the. Well, I know I am. I'm probably one of the worst patients because I see commercials on TV that tell you what the side effects are for these, these new drugs that are coming out and I have every single one of them, every one. And I'm like, okay, I'm dying. There, there it is.
Don't use AI to make those decisions. Don't use it to, you know, to say, oh yeah, you've got this, this and this. And it's some complex thing that nobody know, maybe two people in the world have. Right. Because AI is going to be that way. It's going to go to the dramatic. Use it to say, I have something, what questions can I ask my provider and you know, ask them to explain it in a, in a way that's going to be easy for me to understand.
That's what they should be using AI for.
Gotcha.
[00:09:37] Speaker A: Yeah, it makes sense. I mean, I tell people all the time when they say I googled this. And so I have all these questions and I'm like, you know, I went to school for so long, you know, there's so many websites and so much information. Yeah, it's, it's overwhelming. And I think that's probably the disadvantage of all this technology that we're creating is the, the overfeed of information that people just don't know how to discern because we don't have the experience, you know.
[00:10:01] Speaker B: That's right. That's right.
[00:10:03] Speaker A: I mean, I can't discern engineering off of, you know, stuff Because I'm not an engineer, so.
[00:10:09] Speaker B: Yes.
So, yeah, people forget that all the time.
[00:10:12] Speaker A: Right, right. So I tell my family this all the time. I'm like, it's interesting. Everybody's a physician nowadays. You know, that's the reality of it is it's. We don't know anything. But only and true patients know their own bodies and there are things that they know about themselves. But the science and the. In the, you know, know, about. About the pathophysiology of it is really unclear to people and it's hard to explain because we spend so much time learning it. And so. So, yeah, that's right. I completely agree with you.
[00:10:41] Speaker B: And it's great at recognizing patterns. That's what it's good at. It's good at recognizing patterns because it's taking all this data and it's putting it together. But that doesn't mean it's good at correlating, you know, different illnesses that all come together and making a decision. AI is just not ready for that.
[00:10:59] Speaker A: Right.
So. So somebody is, you know, the claim of robots in medicine, you know, so somebody's nervous about it. What might be a fact that you might tell them to calm their fears about robots in medicine?
[00:11:12] Speaker B: Well, you know, I would say that, you know, there are. There's been different robots, if you kind of think about it for a while.
Medical devices, you know, scans, reading some scans, you know, X rays. I thought, you know, really was the start of. Of AI participating in those kind of things. You know, what I would say is if you're nervous about it, you should ask your physician, is it used in certain parts of the process? Process, and are they involved in the final decision that is made on whatever my treatment plan is going to look like? You know, it's good to be, you know, to ask those questions to make sure you understand where AI is supposed to be in the process, but then know for sure what sort of manner it's used for and how that treatment plan came into place. Because otherwise, you know, you're not going to know. You may see something on your insurance that may surprise you and, you know, just being open and asking those questions is going to be Good.
[00:12:13] Speaker A: Great. Well, Dr. Badeau, thank you so much for breaking it down in such a human and clear way.
Can you give us an idea if any of the, you know, followers want to figure out how to talk to you or reach out to you where you're available?
Just kind of give us an idea how they can contact you?
[00:12:26] Speaker B: Yeah, sure. So they can go to my LinkedIn page. Very easy to find there. It's just LinkedIn Dash AllenBadeau. Or they can go to my website or they can watch my TV show on no Media Wednesdays at 6pm Central called AI Today with Dr. Bideau.
[00:12:42] Speaker A: Great. All right, everybody hang tight. We're going to come back with Dr. Bideau for some more information about artificial intelligence.
We'll be back after the few minutes. Stick with us. We'll be right back with more real stories, real breakthroughs, and real lives transformed.
Welcome to Real Medicine real lives. I'm Dr. Yasser Sombal, and together we're showing medical expertise in a human way.
Welcome back to Real Medicine Real Lives. You love what you're watching. Don't miss a moment of Real Medicine, Real Lives or any of your favorite NOW Media TV shows. Live or on demand, anytime, anywhere. Download the free Now Media TV app on Roku or iOS and enjoy instant access to our full lineup of bilingual programming in both English and Spanish. Prefer to listen on the go catch the podcast version of the show on the Now Media TV website, www.nowmedia.tv. from business and breaking news to lifestyle, culture and everything in between, now media TV is streaming 247 ready whenever you are.
We're back right now with Dr. Alan Badot, AI strategist, technologist and CEO of Alan Badot LLC. In the first segment, we looked at how AI and why it matters in medicine. Now let's talk about access. There's a lot of people that live in rural communities. They have to leave town or the state to get advanced care. But the truth is today's technology is bringing high quality medicine to places it never reached before. So let's figure out and explore how this is happening. Dr. Badot, welcome back.
[00:14:10] Speaker B: All right, thank you. It's great to be here.
[00:14:11] Speaker A: So, you know, clearly that's something. You know, I live in Sugar Land, Texas, which is technically a big suburb in Fort Bend County. I get people that come from an hour, hour and a half away to see me. They feel like the care out here is better. They don't want to see their local cardiologist because they feel it's a small town.
So how can AI or cloud tools for small hospitals offer big city testing, for example?
[00:14:35] Speaker B: Yeah, that's a huge problem. I went to school in West Virginia and went to West Virginia University and that was a topic that I would hear just about every day. You know, this small hospital is closed and now a person has to drive two hours to go to WVU or some other major medical center to get treatment. And that's really hard on folks. And, you know, what it allows you to do, though, is if you've got access to those resources, then you can almost borrow, I call it like borrow. A brain is a good way to think about it where you can, you know, have an MRI done in rural West Virginia, but then have either AI or, you know, that's trained on it or something else, you know, assess it, make some sort of, you know, diagnosis at least to get things started. And then maybe a local doctor could, you know, have some sort of chat, video chat, using whatever the tools are that the hospital would approve to, you know, talk with specialists at bigger cities or other, you know, statewide, you know, facilities.
And that really is the best treatment that they can get. One, they don't have to go very far, you know, two, they're, they're still getting top notch, you know, support from a variety of medical resources. And three, you know, it's just, you know, it's faster and faster treatment usually is a better, you know, it has a better end product to it.
[00:16:08] Speaker A: Gotcha. And so, you know, if some rural area wanted to find some way to help doctors, you know, read scans and minutes, are there low cost devices that you're aware of that use artificial intelligence to kind of make this happen?
[00:16:21] Speaker B: There are some, you know, there are, you know, I know, like portable ultrasounds that they have that allow, you know, they're about the size of a smartphone almost. And I think they're, they're like seventeen hundred dollars or something like that. And for a hospital that's not, that's not too bad for a cost that, but they can connect to the cloud, they can, you know, take that data in and then they can get a, you know, a reading back from that or an assessment of that. They've got digital stethoscopes, they've got, you know, Johns Hopkins just started a, an AI based MRI capability or MRI and X ray capability that you submit those into the cloud, the AI reads it, we'll give you a diagnosis. And it takes like minutes as opposed to, you know, a little bit longer in, you know, some other cases. And those kind of things where high tech is starting to really impact that, that care. But it doesn't solve some of the other financial issues, you know, because if there's no hospital there in the first place, there's, there's no way to bring in those other technologies. So it's really a, a speed to deployment issue because if they're, you know, they can get it, but there's no, nobody there to use it. It's not going to help.
[00:17:42] Speaker A: And, and can you have any stories about, you know, real success stories where remote technology may have saved a patient in rural area that you've been involved in maybe developing?
[00:17:52] Speaker B: Well, not so much that I've been developing, but that I've been following for sure. There was a, you know, there's a tool called Telestroke and you know, it's a kind of a robot type tool and it allows you to identify symptoms of early on stroke and really get treatment. What is it? The golden hour I believe is what they try to say. And you know, using this tool they were able to identify it, get information down to Seattle and then Seattle was able to examine him remotely. And then, you know, they, they were able to send that information back to Alaska and where a local nurse was able to give him, you know, the clot busting medicine and, and some of those other treatments that, that really, really saved him. And he, you know, my understanding is, is that he is fully recovered and you know, has no permanent disabilities and it's that, that kind of treatment, if you think about it, that is huge. That's really where you can have these, these breakthroughs. And I think AI, if it's used in those, those means that's a great place to, great place to use it.
[00:19:05] Speaker A: And so, you know, you were running to, we were talking about rural areas and et cetera. And so obviously in rural areas certain things like slow Internet, you know, difficult ways to communicate things that may have blocked the progress or block the ability to use AI. So how do people fix those issues and overcome them in order to be able to use AI in healthcare?
[00:19:28] Speaker B: Yeah, that's been going on for years.
Like I said, I went to school in West Virginia and the last miles just seemed to never be built. And broadband was always a challenge. It still is a challenge. And rural West Virginia and same thing with rural communities throughout the, you know, throughout the country.
And these initiatives just, they seem to get started and then stopped. But until we can fix those, and there was supposed to be funding for some of those, for whatever reason, it never gets spent.
You know, there's, there's a laundry list of issues where poor Internet now is the driving factor. It's not the doctors, it's not the technology.
It is a fundamental lack of having good Internet. And that's, that's really sad. And unfortunately, you know, we've, we've got to fix it, but we've got to have some folks that are in certain government Positions to really start to own it and, you know, and stomp for it, because otherwise we'll be here five years from now and we'll have the same problem.
[00:20:40] Speaker A: Yeah, no, I agree. I mean, I think, you know, I laugh all the time when I'm in an elevator and my cell phone reception drops and I say, we can talk to a man in space, but you can't seem to get a cell phone reception inside an elevator or.
[00:20:53] Speaker B: That's right.
[00:20:54] Speaker A: And it makes absolutely no sense to me. You know, like, I just don't get it, why the technology somewhere is really robust in the technology where everybody's living and using it. Just can't seem to get it together. I don't.
[00:21:05] Speaker B: Yeah, it's unfortunate. It's unfortunate. Even cell phone signals and. And stuff, they try to do that and it's still, like you said, it still drops. It's still slow.
You know, it's. It is really sad that we. You're right. You can't have. Maintain your signal in an elevator.
[00:21:22] Speaker A: So how does technology, for example, help people, you know, loved ones stay closer to home during treatments they may undergo? And what experience have you had with that and been part of, you know, implementing or et cetera?
[00:21:36] Speaker B: Yeah, yeah, the telemedicine stuff. I mean, Covid really drove that. You know, the people didn't stop getting sick. Right. They still. They still had Covid, but they still had all these other things that had to still be treated. And I think for the first time, it was a little wonky to begin with, but, you know, bandwidth was an issue. And what kind of software are you going to use, how you're going to transmit data, all that stuff, approvals.
That's the first time, though, we didn't have a choice. We had to use something in order to try to, you know, maintain your health or maintain, you know, an ability to support folks that get other types of, you know, diseases and stuff. And, you know, that really tested a lot of different technologies that folks weren't expecting. And those solutions now are getting better, they're getting faster, they're getting AI support now. And that's great.
But, you know, I'm a little concerned though, too, that maybe it's becoming a little too impersonal. Sometimes you're just going through the motions, either the patient or the doctor. So we've just got to really make sure we have balance between those.
[00:22:45] Speaker A: Yeah, no, I agree. I mean, we, you know, during COVID I remember offering telemedicine. Interestingly enough, nobody wanted to do telemedicine. To come to my office. They all wanted to actually come in.
[00:22:55] Speaker B: They wanted to get out.
[00:22:57] Speaker A: They wanted to get out. You're right. You're absolutely right. Although, I mean, Covid was a really, really tough time. It was sad to see, you know, some of the people that had to say goodbye to their family members over an iPad. And, I mean, it was great that we had the technology to be able to do that, but it was also very, very sad at that time that, you know, those are the kind of things that happen.
[00:23:14] Speaker B: Yeah.
[00:23:16] Speaker A: So, yeah, I mean, I think artificial intelligence has really come a long way, obviously, since the development of it, and I think it's going to continue to progress in health. Right.
You know, I'm actually just personally looking into these AI services that offer, you know, transcription services, because I'm tired of dictating all my own notes all the time. So.
So, yeah, I think this is. This is really, really interesting information.
So one of the. One of the things I might want to ask you last is, you know, somebody wants to learn how to have their local clinic benefit of technology.
What would be the first step for them to do?
[00:23:51] Speaker B: Well, I would say I'm more than happy to help people for free if they have an issue and they want to learn something about AI and how they can use it or any kind of technology, you know, reach out to me because I love doing that kind of stuff. It's important for me to give back, and any way I can help them in that manner is great. I would also say, though, make sure that they contact their local hospital system or, you know, somebody like that where they can get an idea on what sort of things they're doing and how they can. How they can help them so it stays local.
[00:24:23] Speaker A: Great.
Well, Dr. Bideau, thank you so much for this insight. You know, we're going to take a quick commercial and we're going to come back and discuss more artificial intelligence with you in healthcare and how we can see how it can improve our lives. Hang on, everybody. We'll be right back. Stick with us. We'll be right back with more real stories, real breakthroughs, and real lives transformed.
Welcome to Real Medicine, real lives. I'm Dr. Yasser Sambal, and together we're showing medical expertise in a human way.
Hello, everybody, and welcome back to Real Medicine, Real Lives. I'm your host, Dr. Yasser Sambal, and we're here with Dr. Alan Badot, still an AI evangelist and CEO of Alan Badot, LLC. We've covered how AI has helped explain, expand and equalize healthcare now we're going to turn to speed. So if you've ever waited for days or weeks for test results, you know how agonizing that waiting period can be. But AI is helping change that story. Dr. Badot, welcome back.
[00:25:18] Speaker B: All right, it's great to be here.
[00:25:19] Speaker A: So can you give us an idea? You know, obviously, you know, people wait for test results. I have patients that call me, like, you didn't read my echo yet, et cetera. Sometimes I'm a little behind. But how can machine learning models, for example, read X rays and labs in minutes and give people the results that they need right away?
[00:25:36] Speaker B: Well, they can, they can do that as long as they're trained on it. You know, X rays are great examples. They're, you know, the, the system that they just released at Johns Hopkins and, you know, it's been trained on millions and millions and millions of different X rays. And so if you have one and it gets submitted, it'll give you a diagnosis in seconds. Right, where maybe a radiologist that, you know, may take a little bit longer, if they're remote, then it may take a little bit longer. Right. And so, you know, that's, that's a really good example because again, it's just looking at patterns and it's good at patterns. Now it could be pathology, again, something that is very good at, you know, patterns, being able to look at those and make some of those, you know, tiny judgment calls just based on, you know, the amount of training that it has had in there when maybe, maybe a human might, might miss that. But, you know, in that specialty area, if you've got a lot of echoes and you've got, you get behind and those kind of things, having AI as a sidekick would be, would be pretty good.
[00:26:46] Speaker A: So what kind of diseases that we see nowadays that, you know, are the biggest, you know, can. AI can help save time and diagnosing and interpretation of results, etc.
[00:26:59] Speaker B: I would say from a disease perspective, what I have seen and even in some of the literature, it's, it's kind of, you know, 50, 50. I would say in some cases I've seen the AI get it right and they, they, you know, they got the, the human presentation right, the symptoms right, and they got the, the diagnosis right. In other cases, I've seen where the AI got the diagnosis right, but the symptoms were not correlated. And in some cases it just completely missed it. And so what I would say is that you've got to be very careful at using it for certain things. There's a reason why chatgpt says at the bottom that large language models are, you know, can make errors and they can make mistakes. And those models that a lot of folks are using online are not trained on medical data. They're not trained on the medical, you know, literature that's out there. So I really would not feel good about just saying, oh, it's going to excel at certain diseases, quite honestly, certain patterns, certain scans. Yeah, I think it can do pretty well with that. Other things, they're specially trained and I, they're not just really ready for prime time yet.
[00:28:26] Speaker A: Yeah, that's true. And so in that aspect, what, what is out there that is helping? For example, how, how do you, as an AI specialist, how does, how do you help AI get better at this so people can understand what the model is that we use to improve the system so they can continue to trust it?
[00:28:43] Speaker B: That, that is a, that's a great question, because it's all about data and it's getting a doctor to participate in different types of studies.
For instance, from a cardiology perspective, it may be looking at different blood flow simulations in and out of the heart. It may be obstructions and its impact on those and trying to get some sort of additional test data into the AI to help train it on those edge cases that we, that we see. It could be, you know, a, a dermatologist who is participating in an AI study for melanoma, so it can identify melanoma or types of melanoma faster. Right. It's those kind of things where, where AI will excel.
Just using a general model is not going to work because they don't understand everything. And it's, it's, it's, you know, people think that these AI models, because they're trained on billions of data points, that means they're trained on everything. It's not even close. And so specialty models, specialty studies, specialty data around the medical practice and how to treat patients is going to really dictate how quickly AI can really be a better assistant to the medical provider.
That's also one of the reasons why I don't think it's going to get replaced in 20 years.
[00:30:09] Speaker A: You mean, you mean the physician won't be replaced?
[00:30:11] Speaker B: That's right. The physician won't get replaced in 20 years.
[00:30:13] Speaker A: Right.
I'm sure a lot of technology and maybe ancillary services could probably potentially be replaced with artificial intelligence or even robots or, you know, that, that makes sense because it's not requiring really much in the way of human interaction per se. But yeah, I could, I could totally agree I mean, just like I don't think general contractors are going to be replaced by, you know, robots because I just don't think it makes sense.
[00:30:37] Speaker B: Right, that's right now. And if, and if you think about it, where you could or other physicians use it is like having to write letters to justify, I don't know, you know, some sort of special sticker that they get on their car or something that, you know, that they can use that an insurance company may not pay for that you have to write 10 letters for. Right. AI can help you with that.
The treatment though, you know, gives you more time to be with the patient and really, really personalize, you know, your treatment and spend more time with them. That's where the doctors can see the biggest improvement.
[00:31:14] Speaker A: It's really interesting that you said that. I never really thought about it that way. You know, I mean, just, you know, amount of bureaucracy to get testing done. Nowadays you're required to call the insurance company for everything and get authorization and do all this stuff.
I never actually thought about the notion of how do you even in your own note, when you're trying to order a test, how can you use AI to make that note fit with the insurance company's needs that meets the criteria for approving the test so you don't have to waste all this time on the back end with authorizations and phone calls, etc. That is a very interesting concept. I actually never even thought about that. I might have to see how I can implement that in my practice.
[00:31:56] Speaker B: Yeah. And I've worked with some centers and physicians where we have tailored different solutions for them specifically for that. Oh, here's one insurance company we have to justify again, here's an AI model. You put in, you know, the basic information into your medical note. The AI reads that note and if it needs to justify something, it uses that, writes the letter, sends it away.
[00:32:24] Speaker A: So how does a faster diagnosis, for example, and using technology, you know, help a family's emotional journey?
[00:32:33] Speaker B: Yeah, you know, nobody likes to go to the hospital and you know, the faster that you can treat something, the quicker you can get them home, quicker you can make them comfortable and you know, usually you're going to have a better outcome. Right. With those kinds of things. And if a person, you know, can really, you know, benefit from those technologies that are being used, then there's no reason that they have to go to the hospital to have certain things done or to have follow ups or those kind of things because they can just stay home. You don't have to travel, you don't have anxiety from sleep because you're nervous, because you've got to go in or, you know, it's depending on what the, the diagnosis is, of course, but the quicker you can get them out of the hospital, the better.
[00:33:26] Speaker A: That's true. I agree. I personally, I don't. Even when patients tell me I don't want to be here, I'm like, neither do I. I don't want to be here either, you know, and I just, I get to go home at least at the end of the day so far. So.
[00:33:36] Speaker B: That's right, you know.
[00:33:38] Speaker A: So how, how do you recommend, you know, if AI results are accurate or even better than human reads, let's say, I mean, and are they, are they accurate and better than human reads? Because I think, you know, it's easy to look at a number within a lab value and say, well, this doesn't fit in the range, you know, and people tend to treat numbers, whether it's blood pressure numbers or electrolytes or etc. But it doesn't take it into the context of what you're actually looking at because there's an overall picture when you order a lab or a study.
So how does AI look at that information? How can we train it, for example, to be able to interpret it within the context of the patient?
[00:34:20] Speaker B: Personalized medicine, when you start looking at those kind of things, it really becomes important to get as much data as you can. But the realities are you can never have enough data. And that's what I stress to my engineers. That's what I stress to, you know, folks that are trying to release these kind of products into, into medicine. There are so many different body types, makeups, you know, I don't have to explain that stuff to you, but correlating all of that information and centering it on the patient themselves is very different than looking at them from probability perspective. Perspective. And AI looks at it from a probability perspective and picks the best outcome. That's very different. And it's a very different type of treatment. And I just don't think we're, of course, there yet. And I would not say it's going to be more accurate because that's not always going to be the case. In some cases it will be accurate, more accurate, and others it won't. The one thing that it will be is always consistent.
It will always, you know, so if you see a diagnosis that is all, it's always getting wrong, you can correct that because you're changing some of the data and you're changing some of the math with humans, it's harder to always do that because you're really trying to focus on the patient. And so that's where that relationship between the AI and the physician is going to really become key. And that's where the biggest acceleration is going to be in the biggest improvement. Using them both together.
[00:35:57] Speaker A: That's great.
So this was really powerful. I'm really enjoying this conversation. We're going to have to take a commercial. But anybody that joined us late, can you let them know how they're able to get in contact with you to be able to Discuss all this AI information and diagnostics?
[00:36:11] Speaker B: Sure. They can go to my LinkedIn page or they can go to my homepage. It's AlanBedot AI. Or they can watch me on AI Today on Now Media at 6:00pm on Wednesdays Central Time.
[00:36:27] Speaker A: Great. Everybody hang tight. We're going to go to a quick commercial and come back and discuss some more artificial intuitives with Dr. Bideau. Stick with us. We'll be right back with more real stories, real breakthroughs, and real lives transformed.
Welcome to Real Medicine real lives. I'm Dr. Yasser Sombal, and together we're showing medical expertise in a human way.
Hello everybody, and welcome back to Real Medicine Leaves Real Lives. Do you love what you're watching? Don't miss a moment of Real Medicine, Real Lives, or any other your favorite NOW Media TV shows like On Demand, Live or On Demand, anytime, anywhere. Download the free Now Media TV app on your Roku or iOS device and enjoy instant access to our full lineup of bilingual programming in both English and Spanish. Prefer to listen on the Go catch the podcast version of the show on the Now Media TV website at www.nowmedia.tv.
from business and breaking news to lifestyle, culture and everything in between, now media TV is streaming 247 ready whenever you are.
So we're going to close out this segment today and continue our discussion with Dr. Bideau, a leader at the intersection of AI, cybersecurity and innovation. We've been discussing how technology can improve care access and speed. But with great technology comes great responsibility. In this final segment, we're going to talk about something that keeps many patients up at night, keeping your personal medical data safe.
Dr. Boudo, welcome back.
[00:37:53] Speaker B: Hi. It's great to be here.
[00:37:54] Speaker A: This is actually a very, very big topic. And you know, there's, I mean, I was actually one of the hospital systems here, I think about a year or two ago, got, you know, roped into one of these cyber attacks and the whole epic system went down and it was a big, big mess. And it's interesting not just from a patient data perspective, but also from a hospital logistics perspective. I mean, I trained in an era where we had to write our notes and we had to write orders, and with all this technology, we no longer have to write orders. It's all templated. We don't have to really write notes anymore as much. It's all dictated or templated that we can just fill in the blanks. And it was amazing to me how hard it was when we actually had to sit back and do that again.
I mean, just writing admission orders or post op orders and go through the process of remembering how to do it after I had not done it for maybe over 10 years.
So what, what are the biggest cyber risks, you know, hospitals face with patient data today?
[00:38:58] Speaker B: Well, it felt like the, the early 2000s, didn't it? The dark ages came, came back. Right.
You know, we have become so dependent on these IT systems that are, that are out there today that when something like that happens, it really has a ripple effect across everything and it affects everything. You know, patient care, billing, scheduling, you name it. And really the biggest component of that is around these ransomware attacks. If a hacker can gain that kind of information and gain a patient's, well, the patient database of everything that they have. That's why hospitals are paying, you know, to, to, you know, to get the, the ability to get that, you know, unlocked because that really drives everything and it drives how they're doing and it drives, you know, the how you're going to treat folks. And just being able to get that, you are, in essence, crippling a hospital system. And a big hospital system can generally prepare a little bit more, can be, you know, a little bit more agile to, you know, have backup data and things like that. But these small hospitals, one, they can't afford it. Two, they don't have the technology to do it. And three, if something like that happens, they're done, they are done for and they cannot operate until they get that turned back on. So those ransomware attacks are huge.
[00:40:34] Speaker A: Yeah, they really are. And so how does the encryption act like a secret code for our health care records to protect it?
[00:40:43] Speaker B: If IT systems are not encrypted, and I don't know of any hospital system that is not using some sort of encrypted system, then, you know, it's easy for folks like, you know, myself or other, you know, folks to gain access to those encryption, allows you to scramble the, the information and it makes it unreadable for somebody that doesn't have the key to unlock that encryption. And think about your writing in your diary and you're using symbols and you're using some other, you know, maybe you're writing right to left instead of left to right or whatever that is.
It's the same sort of thing with encryption and medical data. It's just scrambling it so somebody else just can't look at it and say, oh, yeah, this is Alan Badot. His birthday is this. Yada, yada, yada. That is the, that's the big driver around your ability to protect that data.
That's true.
[00:41:44] Speaker A: And I think it's important people understand, I think these, these attacks and these cyber attacks are not so much about. They want to use your information to do anything, but it's more of, you know, an attack. That's the only way they can attack these institutions to get some sort of money out of them. So.
[00:41:57] Speaker B: That's right.
[00:41:58] Speaker A: You know, it's not really about, you know, who's got diabetes and what your birth date is. It's really just about a, you know, it's kind of like credit card fraud. Right. They just are going to gain some sort of money or potential out of it. So.
[00:42:09] Speaker B: Yeah, yeah. And one of the bigger, one of the bigger things that we've started to see on the rise is really using, you know, AI to take your healthcare data, modify it and submit different things to insurance companies. So insurance fraud from a patient perspective is becoming a little bit more, you know, dominant in that area. But a hacker. You're exactly right. They don't care that Joe has, you know, X, Y and Z disease. They want the hospital to have to pay to unlock that so they can do everything else around that patient. That's where the benefit to those guys are. They, you're. You're exactly right. They don't care about the other stuff.
[00:42:51] Speaker A: So what, what extra steps do you recommend to clinics or hospitals to take, you know, before AAI touches their private data?
[00:42:59] Speaker B: Well, one, they've got to make sure that they're prepared for these kind of ransomware attacks. There are steps that they can take to, to make sure that they are really prepared to, to respond. They're prepared to, you know, continue and have that continuity of operation. But then they're also prepared to, to pay a little bit more to store their data somewhere else, or at least a full backup copy of their data somewhere else. Because if that happens, then they can spin it up, it can continue.
But if they, if it happens and they're not Ready? They are at the mercy of those attackers. There's nothing that they can do except pay.
[00:43:38] Speaker A: Yeah, that's true. That is.
Are you aware of, you know, how many people this happens to that they end up actually having to pay? I mean, I know usually the FBI gets involved and. But common that they end up having to pay just to get this issue resolved.
[00:43:53] Speaker B: It usually.
So the smaller banks right now are the more popular ones.
Last year there were smaller hospitals. Now universities are starting to get attacked more for their, their student databases and stuff.
I would say probably it could be as high as 35% of hospital systems have been hit with it.
[00:44:13] Speaker A: Wow, that's, that's insane.
[00:44:15] Speaker B: And you don't hear about them very often either. That's the scary part, because they want to keep it quiet.
But it's happened, right?
[00:44:22] Speaker A: Yeah, I mean, had I not experienced it just being part of a hospital system, I would have never assumed it was happening and et cetera.
So going on that note, how does a patient, for example, check or see what kind of questions can they ask to make sure that the hospital or clinic that they're part of is following, you know, strong security rules to make sure that their data is safe and protected?
[00:44:46] Speaker B: Yeah, the biggest thing is going to be around HIPAA compliance. Are there soft? You know, is the software HIPAA compliant? I mean, for big systems that are using these larger ehr, you know, software tools and stuff that are out there, they're, they're, they're HIPAA compliant. You can ask about training. Is there any cyber training that takes place? Is there any. Do you sell my patient data to anybody? Most of those, I would probably say 98% of those hospital systems are. The answer is going to be no.
But making sure that they understand that when they log on, the patient portal is secure, the emails that are sent are secure. Those kind of things are really what patients should, should ask and think about.
[00:45:33] Speaker A: And so if a breach happens to a hospital or clinic that a patient is part of, what, what immediate rights or supports should they be expecting from those institutions?
[00:45:45] Speaker B: Well, there are, there are fundamental things that you have to do as a system. You have to notify them within 60 days.
You have to tell them exactly what was stolen.
And in most states, probably, I think it's 48 states, you have to offer them some sort of free credit monitoring, you know, at least for a year. And I think in Some states it's 18 months actually. So you can get copies of your medical records. You can make sure that any errors that were put in there during the breach are corrected at no cost to you. So there's a lot of, a lot of ways to do that. The. You just have to file a complaint.
[00:46:24] Speaker A: And so when you, when you mentioned that when they do these attacks, is there only certain data they extract or is it kind of an overall thing they just take it all or.
[00:46:34] Speaker B: Because you mentioned they take it all.
[00:46:35] Speaker A: They take it all. So it's not.
[00:46:40] Speaker B: They want it all because then they can really crush it.
[00:46:44] Speaker A: Wow, that's. That is amazing.
[00:46:46] Speaker B: Yeah.
[00:46:47] Speaker A: The more technology we get, the, you know, the more dangerously sometimes the world becomes because it's gone from, you know, violent crime and, and so forth to really just this cyber crime. That's just crazy. Whether it's credit card fraud or mortgage fraud or, you know, artificial intelligence using, you know, I mean, I've heard stories of, you know, artificial intense people. You make voices of family members and, you know, call you and say they're desperate for money. And yep, it is, it has become really, really a dangerous thing. Is there. Do you. Are you aware of any government, you know, interventions or regulations coming to be more.
Make AI more regulated and more secure so that, you know, we can avoid these issues?
[00:47:34] Speaker B: So, unfortunately, I think we, as the United States are a little bit behind the regulation part of it. I think we're behind from a data perspective.
I don't believe people own their digital identities really anymore. And there's been a lot of liberty that have been taken with, oh, I'm going to use this data to train my model, or I'm going to use this data to do something else, or I'm going to sell this data to a third party that's going to use it for something else. And I think we are behind. I think the first thing that has to happen, there's got to be a digital identity Bill of Rights that has teeth to it where if somebody or some entity violates that, there are penalties associated with, with those. And I'm not talking about a slap on the wrist or something like that. There have to be financial penalties and criminal penalties that are part of that. And until we get the digital identity piece done, the AI piece, it, it can still accelerate even if we put some limits on those. But we're not. And so, you know, we're still behind. We've got some broad visions. You know, the President, President Biden had, you know, some initiatives. President Trump is, is, has taken a few steps farther. And I think at least, you know, they're the, you know, Congress and the administration are looking at a broader set of AI capabilities and how to, how to handle those. But unlike Europe, Europe has, you know, some more laws in place and they're, they're, that's where they're ahead of us, is really around, you know, solidifying some of those laws. I'd say in 18 months we'll have to do something because the outcry is just getting too loud down.
[00:49:22] Speaker A: Dr. Badot, thank you so much for helping us understand, you know, and stay safe in this new digital world. It's clearly just as crazy as the criminal world, you know, for viewers, just one last time, you know, can you just kind of give them a reminder how they can find you, reach out to you, talk to you about any of this stuff?
[00:49:38] Speaker B: Yeah, sure. They can go to my LinkedIn page or they can go to my website, AlanBedot AI or they can watch my TV show, AI Today at 6:00pm on Now Media TV.
[00:49:51] Speaker A: Great. Thank you so much, Dr. Bedot. Your insights have brought clarity to an often confusing and sometimes intimidating conversation. Thank you for sharing your expertise with us and reminding us that the future of healthcare isn't about machines. It's about trust, people and thoughtful progress. To everyone, watching technology is really not something to fear. It's something to shape.
With the right voices at the table, the right questions being asked, artificial intelligence can help us really save lives, reduce pain, and make health care more human, not less.
Let this be a reminder that every innovation still needs compassion, even breakthrough still needs responsibility.
And no matter how smart our system gets, real medicine will always be about real lives.
I'm Dr. Yasser Sombol. This is real medicine, real lives. See you next time.
