EP 111: How AI is (and Will Be) Used in Healthcare

This is the Everyday AI show, the everyday podcast where we simplify AI and bring its power to your fingertips. Listen daily for practical advice to boost your career, business and everyday life.

What will the future of healthcare look like with AI? You know, we've talked about it once or twice on the show before, but I think our guest today is going to be shedding some light and giving us insights on not just how AI has been used in the medical field and in healthcare, but also how it will be used. So whether you're in the health field or not, I think today's episode is a must listen, a must watch if you're joining us live. So thank you. If you are joining us live. My name is Jordan Wilson and I'm the host. This is Everyday AI.

This is.

This is a daily live stream, podcast and free daily newsletter helping everyday people keep up with what's going on in the world of AI. Because there's so much. So this isn't just the news. This isn't just a conversation. It's what we can take away from it. It's how we can leverage it and how we can grow our companies and grow our careers. Does that sound good? I literally can't wait to bring on our guest, but before I do, we're going to go over the AI news and if you're joining us live, let me know. What is your one, what is your one biggest question on AI and healthcare? Get your biggest question in now. I want to make sure to see what, what your questions are because we're going to get to them. But first we've got to talk about Meta. Let's go over the AI news. Meta just landed a hard punch in the AI battle with their announcements at Meta Connect 23. So the Meta Platforms has released new AI products, including a chatbot and smart smart glasses. Also, Meta has their new with the smart glasses, they'll be able to identify places and objects as well as perform language translation. Very cool. That's not all. There's so much more Meta News. So they're also launching their AI driven Persona chatbots across Instagram, Facebook and WhatsApp. We've talked about those probably a dozen times on the show and in the newsletter before. And developers will also be able to build their own AIs inside to to use on Meta's messaging platforms. Oh, that's not even all. They also released their mixed reality Quest 3 headset which will have some AI applications for 500 bucks. So not even crazy when you think, I think Apple's is 3, 3000 or 3500. All right, some more AI news that's not meta. Adobe launches their new Photoshop online. So it is including some popular AI features from its Firefly AI platform like Generative. It seems like they're trying to with the online offering, try to compete a little bit with Canva, except not on price. There is no free version. There is no freemium version. You know Adobe Photoshop, even with this new online AI powered version they're launching, it is still a paid product. Last but not least, OpenAI makes a little bit of news bringing back Browse with Bing. So we've, we've talked about it here on the show before. They, they had a Mode inside chat GPT before Plus users. So you got to pay the $20 a month again. Pay the $20 a month. I'm not working for Open AI to say it, just do it. But they had to take the browse with Bing feature away because previously it was kind of accessing websites that were behind a paywall. Not good, tons of lawsuits. But Browse with Bing is back. And here's my hot take. Do not use it. Do not use is in my opinion it is a marketing angle. It is not very useful at all. Use an Internet connected plugin instead. Don't get me started on that. I'll accidentally be here for three hours and then our guests won't be able to come on. So speaking of that, I'm excited for this one. This one has been a while in the making and actually our first guest, our first repeat guest. So please help me in welcoming to the show again Dr. Hardy Castro. He is the author of multiple books on AI in healthcare, a physician, entrepreneur, all around AI enthusiast, Dr. Harvey Castro. Thank you for joining the show again. Oh wait, I think, I think we have a, a little technical issue.

I got it. No, I got it, buddy. Sorry. I just want to say thank you everybody. It's awesome to be back. I love this show. I love everything that you're doing. So let's, let's do it.

And Harvey, Harvey brings the party because you know, whenever he's on the show, people, people come, they want to watch, they want to listen. So welcome if, if you're in Harvey's network. Thank you for joining us. And as a reminder, if you're on the pot listening on the podcast later, always check the show notes and check the show notes in the comments today. A lot of other resources but you can also come in and join the conversation if you're listening on Spotify, apple, come ask Dr. Harvey, a question. So let's start at the top. One thing that I want to know, Harvey, just. Just for everyone who's. Who's unaware, what are you doing in the healthcare space? Because I feel everywhere I go, everywhere I turn, even a Fox News article this morning talking about, you know, some new AI advancements for paralyzed patients. You're everywhere on the scene when it comes to AI and healthcare. So give. Give everyone a little bit of background on who you are and how you just became this. This voice of kind of an AI leader in the health space.

I appreciate it. You know, honestly, it starts with this thing called the iPhone. And sorry, I got actually Covid, but I'm not contagious now, so I'm good now. So no one will get it over the air from this air right here. Rather. Anyway, it starts with the iPhone. Basically, I was coding a patient, and with iPhone One, and at the time, the nurse got this big textbook. I said, hey, we need to start this EPI drip. We need to do this other drip. And she's like, let me verify these doses. And literally she got a textbook, started thumbing through it, and then got the medicine. And I thought back then, I was like, gosh, there's got to be a better way. So I taught myself how to program. I created the first IV meds app. It went viral. And as a result, I was able to grow a company into 30 different apps. And fast forward, this thing called chat GPT comes out, and I'm playing with it, and I had this aha moment that I did with the iPhone, and I thought, oh, my God, this is going to change healthcare. This is going to help doctors, patients. And my brain just exploded. And I just started writing book on it. And I'll never Forget, it was December 20th, 22nd. My wife comes into the office and says, like, what are you doing all day? Like, you're like obsessed. And. And what is this Chat GPT thing? And I was like, oh, honey, this is going to change the world. She's like, no one's going to read your book. No one knows what Chat GPT is. Doctors are going to laugh, like, don't waste your time. And I'm like, no, no. Even if I sell this book and it's one person, I don't care. At least get the word out. Well, fast forward, all these people have been calling me to do world stage keynote speeches around the world. Just different opportunities to just help people. And honestly, it all started in a good place. It just started not as a business, didn't start as anything else other than a public service announcement. And since then, I've been getting calls from Fox News, different news media around the world, asking me to do quotes for different articles. But my mission hasn't changed. I just want to help doctors and patients. I haven't officially created a company in the sense that I'm like, this is my company. No, it's more like, let me help out. And from it, other opportunities have arise that I'm actually being able to consult and do other things. But overall, it's just been just such a fun ride.

Yeah, it's, it's been great. And, you know, I'm even learning so much, you know, new information about AI in the health, health field. Dr. Castro always keeps me up to speed, you know, if there's new, new studies or anything. So if you've heard anything on the everyday AI show before, there's a good chance that, that Dr. Castro sent it to me. And yeah, even, you know, I mentioned this. We'll, we'll shoot this out in the newsletter.

Don't worry.

But Dr. Castro was on pretty interviewed for a pretty great breakthrough, you know, that was just, uh, reported on this morning. But I want to talk, I want to talk about this because we can all. We can go read that in the newsletter. Right. But I want to take advantage of every minute we have you on the show here, Dr. Castro, because one thing I want to know is in general. Well, actually, let's even start it further back. Let's hit rewind. Where are we at right now in the medical community with AI? Bring us up to speed because I know there's a lot. I know AI has been used in healthcare settings for decades, but give us it, especially with generative AI. Give us the, the 62nd. Here's where we are right now.

Yeah, so the biggest news was early in the year, Microsoft OpenAI and Epic, so let's call it a trifecta, got together. And EPIC is an electronic medical records that a lot of hospitals around the world use. And they basically said, we're going to use the summaries of OpenAI and plug it into the EMR and big picture. What I've heard is that they're supposed to roll that out this year. I've seen some minor updates that they're working on, the summary part of it that they're working on, other features. I personally haven't seen it because I'm not working at an EPIC place. But I. This article right here will show it from April. But then there was, I think, an Update in August, which will be good, but big picture. My point is this is a huge milestone for us as a society. I think once this comes out, EPIC is going to kind of pave the way of how this was used around the fda, what parts they did, and I think like what's literally an API or this way of connecting with OpenAI and other companies, it's going to help startups create other products in healthcare. And I'm really stoked because once that comes out, I really think the leader is going to be literally leading how others are going to be able to do this.

Yeah, that's, that's so important. You know, one thing, one thing as well that I wanted to ask you about, Dr. Castro, is it seems like, and I've, I've said this on the show before, I think that the medical field in general has actually been ahead of the curve, you know, and they have for, for a long time when it comes to advancements in AI. So, so using, you know, deep learning, machine learning to discover new drugs, you know, new proteins. Right. New, you know, potentially early detection for diseases. So that piece is, is not new in the medical space. But where do you see kind of, you know, when we see those positive developments over the last couple of years, where do you see it going from here? It seems like in the medical field there is a lot of positive momentum around the use of AI, but where do you see it going in the next couple of years? Are you still running in circles trying to figure out how to actually grow your business with AI? Maybe your company has been tinkering with large language models for a year or more, but can't really get traction to find ROI on Genai. Hey, this is Jordan Wilson, host of this very podcast. Companies like Adobe, Microsoft and Nvidia have partnered with us because they trust our expertise in educating the masses around generative AI to get ahead. And some of the most innovative companies in the country hire us to help with their AI strategy and to train hundreds of their employees on how to use Gen AI. So whether you're looking for ChatGPT training for thousands or just need help building your front end AI strategy, you can partner with us too. Just like some of the biggest companies in the world do. Go to your everydayai.com partner to get in contact with our team or you can just click on the partner section of our website will help you stop running in those AI circles and help get your team ahead and build a straight path to ROI on gen AI.

And that's a good one. And I Pause and think. Because, you know, healthcare in general is very traditional. People don't want to like, they don't like change. They don't especially. I'm going to call out. All the doctors like myself tend to be set in their ways. And so now you have the new kid on the block and it's AI and there's some turf wars in the sense that some doctors are like, no, I'm not going to let this replace me, or no, this is not smarter than me. I've been doing this for 30 years. And no. And so there is like this paradigm shift and I think this. I'll bring in my first prop based on your news from Meta. These are my Ray Bans and these are the ones.

Let's go.

These are the ones from the last generation. So I haven't got the next generation, but they look exactly like it. So the reason I'm wearing them is it looks exactly like the next one. And I am going to order the next ones. But it has a little button here. I can video tape. The future will have augmented reality. And so I know this look. I look silly wearing them in the doctor's office. But if I was your doctor and wearing them and I explained now I could possibly see in front of me your medical record, I could look at your X rays. Where the future is and where I'm wearing these is to explain. Imagine I'm discharging you. Imagine you have some disease that you've never heard of. And I'm going over it. How cool would it be if I gave you a pair of glasses, obviously sanitize and all that, put them on you and then together we're seeing the same thing with. We're walking through your disease process. I'm pausing, I'm turning it, I'm explaining it to you. And it's almost like this video where you're like, oh, now I see why I need to take my medicines for my heart med. Now I see why my diabetes is this way. Oh, my gosh, Now I realize why he gets upset when my sugar is so high. And so I think this is going to be the feature now, the other prop, and I'm joking here, it's going to be the iPhone. Not just the iPhone, just the mobile device. I have it set up where I have OpenAI in the background already listening to our conversation. These are the kind of things if we had a question, I could ask ChatGPT and I won't do it on live, but I was gonna. But my point is these kind of Tools. It's gonna augment me, it's gonna augment you. Imagine in the world you're my patient in the future. Imagine we're two different languages or cultures. And imagine I'm in, I grew up in New York and East coast and I may use an east coast example. You may have grown in different part of the states or another part of the world. How cool. If I could speak an example, it could convert it and use AI and put it into your language, your culture even. It may not be the same example, but it parallels the example to your culture. And now I'm really communicating now I'm giving you better healthcare. So the future I think is going to be using AI. The other cool example that I talk about in the future is going to be, it's my favorite case in England. This hospital has discharge instructions and imagine talking to a five year old and how much time you get with a five year old, it's hard to get across them. Imagine if they had a disease and you had to explain the disease. I won't go into the details, but basically they took the discharge instructions, fed it into ChatGPT and said, Talk to me as if you were a five year old. But then they use Mid Journey to create a coloring book for all the big concepts and put it together and said, here you go child. And now the doctor, the parents can sit with the child. And so I thought this was just an amazing way. So I think the future will be multimodality, things like our phone, our glasses. And I think that's going to change healthcare and it's going to blow my mind of all the possibilities. And then I was watching the keynote speech yesterday for Meta and all I could think of is all these different medical applications, just using these as an aside. Next month I'm going to be with a company that works with NASA and they've asked me to see if I can help just give some suggestions on the Mars mission, how we can use AI in Mars. And so I was just honored to be able to be able to like just contribute to such a big mission. You know, this is just kind of cool for me.

Oh wow, that's, that's amazing. Hey, if you're joining us live, let me just recap everything that's going on. Right, so Dr. Harvey Castro is joining us, wrote the book on chat, GPT and healthcare. And not only that, now he's gonna, you know, be consulting on, on how we can use AI to get to Mars. Right. Amazing, amazing. Right. So we have a lot of Great questions. But I wanna, I wanna first, you know, because Dr. Castro's, he's, he's not new to this show. He's, he's always contributing in the comments, you know, connecting with people, putting out great information. But I'm gonna kind of put you on the hot seat here because I, I know you have hot takes, but I want to ask. So, you know, this kind of popped in my head after you, you know, we're giving the story of, you know, writing your book on chat GPT and your wife is like, will anyone read it? And, you know, people are obviously reading it now.

Now.

So what is maybe one other thing that specifically in the AI space that right now that you're seeing that you're saying, hey, I, I have an extensive background here in AI and healthcare. I see this happening. And maybe some of your colleagues are like, I don't know about that, Dr. Castro. That doesn't seem like, like it's something that's, that's going to happen. Maybe. What is, what is that one thing that, that you're really advocating, or maybe not even advocating for, but you're seeing this happening that everyone else in the medical community that is maybe not as in touch with AI is like, are you sure about this?

I got it for you. This, this is, again, imagination. But I really think this is going to happen. So let me give a little backstory. So literally, PPP here. I'm going to do it, but in medicine and talking to you. So imagine a world where the average nurse, and this is true, the average age is 54. So if your average age of your nurse is 54 and 20% of those nurses in the United States are 65 and up, then imagine if we kept that rate the same, then in 10 years the average nurse would be 64, with 20% of that being 74. And so we have a huge problem. And imagine more and more patients like Covid and other things where it's putting a stress on our system to the point where we're going to have to either lower our standard and accept more people into nursing, or nurses are so burned out that it's going to make the problem worse because they're like, this is too much work. I'm out. And now we're really hurting as a nation for healthcare. My idea that I don't think has caught on, that people are like, that's not gonna happen, is the following. I put things together. I love playing chess. So, number one, we got OpenAI coming out with its robots at the end of the year number two, chat GPT4. They're really pushing this multimodality. And chatGPT5 will be here by June of next year, in my opinion. With video fast forward, I'm saying to Doctors, imagine putting ChatGPT 4 and 5 in the eyes of the robot. Now that robot can be a nurse assistant, can bring vitals, can read vitals, can, with the vision of this camera, can tell what the hemoglobin A1c is, what the blood pressure of the patient is, heart rate. And now imagine it becoming a virtual assistant where I walk in with the robot and I'm having a conversation and by the time I'm done, the robot's either summarizing, giving, looking up all your medical records, throwing your images up in the air to show you, I mean, it's just going to change health care. And I don't really think health care people are ready for this because they're like, oh, that's going to replace me. But I really think this is coming.

Why? That's wild. It's wild. But I see it because, I mean, you point out an obvious fact, right? If there's an ongoing nursing shortage, right? And if that is only projected to get worse and there's only, you know, there's going to be fewer nurses in 10 years. I actually now, now that you say it, hey, maybe, maybe you're great at opening our, our eyes. Like I have the, I have the meta Ray bans on now, seeing the AI future, but it seems to make perfect sense and, and especially here, I'll, I'll even share a personal, a little personal story on this. So many times I feel, you know, especially if I'm seeing a new specialist, a new physician, whatever, so many times I feel they just go to Google something or like they're on their computers and they go, you know, refer to WebMD, right? So, so I, I think a lot of people are going to hear that, that hot take there, Dr. Castro, and they're gonna be skeptical and they're gonna be like, no, not a chance. But I don't, I don't think it's that far off, right? Like if you can put an encyclopedic knowledge that is constantly updated, it seems that, that actually it's maybe not crazy. It's maybe the responsible and ethical thing to do as a first line of defense for people that need more accessible and affordable health care. And if there's not enough trained humans on the back end. Wow. Wow. My, my brain is, is, is pretty shook this, this early all right, let's, let's take it another direction. So, on the contrary, so, so I just, you know, put you on the hot seat there. I said, hey, what's, what's one thing that only you believe? What's. What? Maybe let's look at it in a different, different lens. What's a common theme that's going on right now in the medical field, in the healthcare community, in regards to AI and everyone's, you know, kind of, you know, on this same train and everyone's traveling down this same lane, and maybe something that you're like, ah, I'm not sure about that. You know, I'm, I'm knees deep in this every day. I'm not sure if we're heading in the right direction with our, with our train of thought on this piece in medicine or healthcare and AI, Maybe. What's one of those things?

Yeah, I think I'm gonna pivot a little. I think the big picture is people are so scared and hesitant, and hospital systems are like, we got hipaa, we got fda. There's things we can't do. And I think they're, I think in a good way, they're focused on things that are outside the fda. They're trying to do things that will help with patient with physician burnout and healthcare burnout. And that's like epic. And creating these summaries. You know, imagine seeing I'm an ER doc. Imagine you coming to see me and you've had like eight MRI reports and. And you come home, come to me with a big folder saying, here, doc, here's all my stuff. And I'm having to go through all of it. How nice. If I could even just feed it into some kind of GPT that's HIPAA compliant, privacy secure. But then it looks at it, OCRS all of it, and then gives me the summary of all of it. That way I'm not spending all this time trying to look at all the detail where it's there. That is becoming accepted now. The big controversy moving forward is, do I give OpenAI access? How do I secure this information? And I really think the future is going to be in creating kind of like llama at your hospital. And that's actually what I'm going to be doing next year, is working on llama models, working with hospitals and helping them create their own large language model. That way they don't have to worry about privacy and these kind of things. I feel like it's a dichotomy, but I really am excited to see that Llama and Meta and all these people are coming out with all these products because big picture, it's going to help us and it's going to start adapting. The more people start using AI, the easier the administrators are going to say, okay, let's use this tool now it's being adopted.

Yeah. And just, and just for those that, you know, might not be like me and Harvey and reading the daily large language model news, you know, llama 2 is a model from Meta and one of the big benefits is that you can run it locally. Right. So you, you don't have those same type of concerns about, you know, private, you know, patient information. Hipaa, they're, they're a little alleviated when you can run those models locally and you're not necessarily uploading or sharing, you know, sharing sensitive patient data, which, which is actually a great transition to, to Woozy's. So woozy. Thank you for joining us. Thanks for the question. So he's saying what is the biggest challenge in getting AI into healthcare in terms of HIPAA? Because I'm also thinking here, Dr. Castro, because, you know, we have all of these trusted platforms, you know, epic, like we have all these, you know, electronic medical record systems that we upload all of the information into without second thought. So I guess why or what are the hurdles still until we have that same mentality with, with updating it to, you know, an open AI or a Meta or a Google with, with their new model. So what's, what's the biggest challenge?

The biggest challenge, and I know someone just comment that chat GPT Enterprise can address some of this and then I'm going to butcher, butcher this. It's Azure Server. The problem with, with this is the information, you know, HIPAA is so important that if, if your name, your address gets literally hacked, the defiance can be over $100,000. And there some of these things, depending how you do it, there's actually jail time if you mess this up. And then on top of that, your insurance carrier won't insure you if you make a HIPAA violation. So a lot of companies are very hesitant to do this. Incorrect, obviously. But I think what they're doing is they're picking parts. There's 18 identifiers, meaning if you do one of these things, that's hipaa that, and they're looking at those identifiers and saying, okay, where in the medical history, like discharge instructions, I could do that and not fall into HIPAA because it doesn't have any of these 18 identifiers. And so I Think the future will be these kind of things, but again, I think it's going to be scrubbing the data slash, having your own large language model at your own hospital system. That way the cloud is within your infrastructure and it never leaves. And I think that is the future. And I think that's smart of EPIC to start integrating inside of them, because then now they're trying to create it where it's not really going into a cloud. Going back to open AI, they're kind of bringing back, they're only using certain functions of ChatGPT, if that makes sense. Hopefully that did.

All right, I'm going to put you on the hot seat again here because my mind is opening up to different possibilities. So, you know, like we say, everyone just automatically trusts and will upload anything into Epic because it's, it's been the leader at least here in the US for, I don't know, probably decades. But, but what's to keep a company like OpenAI or like Google from maybe, I don't know, acquiring an EPIC or a company like them, you know, because right now it is the tech companies, you know, and someone out there, if you're listening, you're like, oh, there's no way, you know, Epic is a, is a giant company. Well, in terms of valuation, I don't know if you, if anyone heard this, but OpenAI right now has a valuation of just under 100 billion, which is more than three times of the valuation of an epic. So might we see. And maybe not an epic, but maybe, you know, kind of a second or third place kind of similar system. Might one of these big tech companies acquire one of these health companies so then instantly they become the future of healthcare.

That's funny you say that.

It's a hot seat question.

No, I love it. Again, I'm gonna put my horrible analogy of chess. I feel like Microsoft is playing their chess game here really smart. They got the secure server, they're getting in bed with Meta, they're doing a good job with OpenAI. They're acquiring different parts of healthcare. And you're right, what's to hold them back from acquiring someone else that, that takes them, that chess piece to another level. You know, acquiring EPIC would make sense. You know, I don't know what they're worth, but I know Microsoft has more cash than they do and I'm sure they can go 10 billion. Thank you. And so I'm sure that could be pulled off. And then the other thing, technology moves so fast that if Microsoft, I know they, they really wouldn't but in a way, they could support another startup that could be the next Epic that would be better in using this technology. Technology can trump Epic. Epic is not a nice software. I, I hate it personally. So I think it's taking over easily.

Yeah, there we go. And I think. Yeah, I just, I just did a little search. It looks like the last valuation most recently was about 32 billion, but still. Okay, so let's talk this. What are you working on? Let's. Let's go there because, you know, I always love, like, people who are very involved with everyday AI who come on a daily basis. They're contributing. I'd love to give them a platform. So tell people what are you working on? Because apparently you're. You're helping advise for a company related to NASA on, on AI and medical space. But what other projects are you just excited about right now or, or even, you know, other releases from, from other AI companies? Like, what are you looking forward to working on in the next couple of months? Being someone that is so ingrained in AI in the health space.

Yeah, there's. There's two things that are really kind of dear to my heart, and it's kind of weird, but I'll explain. One is I literally clone my own voice and it's called the GPT podcast. And all I do is I find fun articles like AI and healthcare. I feed it into ChatGPT. It creates a fake conversation between my business partner and I. We both use our clone voices and then we look at the manuscript and if it looks good, then we push it out and, and it sounds very natural. And I think it's cool. We're actually going to come with the SaaS model for anybody else to do the similar thing. But the part that excites me is being able to take any kind of information and just put it out there and help people. And it's just really cool because I'm using Midjourney for our pictures. I'm using ChatGPT helping to put these little summaries together. This came out this week. It's just really cool, my way of doing things. I'm so busy, but yet I can still have a voice. Maybe a clone voice, but I still have a voice. The other thing that is cool is I love just figuring out how can I take a AI and apply it to what I do. And so I made a Decision Maker plugin with, through ChatGPT, and it's called Decision Maker. And all it is is I wanted to help startups and companies with like, okay, I have a marketing problem or hey, I have a decision to make with my child or as a doctor, hey, do I give this liver to this person that's drinking every day versus this person that needs a liver that's 18 years old, that kid and has a healthy issue? So there's a lot of ethics involved. So I made this app where you ask the question and it walks you through the answer. And thanks for featuring me week, Jordan. That was cool on your YouTube. So that, that's. I know that was this week, so I appreciate it. So, yeah, so those are the kind of things at the end of the day that's what makes me excited. It's more about helping others get them to another level, but just using the tools that are out there to do something new.

Love it. Love it. Yeah. So many, so many fun and exciting projects that Dr. Castro is working on. And make sure to check out the newsletter today as well because we'll be sharing a lot of those because, yeah, oh gosh, we didn't even get into that. But you know, Dr. Castro has put together some literally fantastic resources, especially if you are in the medical field on, on different, even chat GPT resources. He put together a great cheat sheet. So go to your everydayai.com Sign up for the newsletter. We're going to be putting all that out. Another question here from, from Peter. Peter, thank you for joining us. So Peter's asking, do you think that an LLM can have an impact on other LLM in the healthcare response to the user? I think I fully understand the question there. I'll leave it up to you to interpret the answer because we could go in a lot of different directions with that one.

Yeah, I'm kind of confused, but I think it will open the door. Going back to the Meta example, that LLM is going to help it because now I can ask questions in the future of just healthcare question and it'll be right there, you know, or fun facts. And so that's one way that one LLM and then I know that when OpenAI sees what meta is doing, they're like, no, no, no, we're going to do the same thing and we're going to put an API so that if you have glasses, you can just start upgrading it and having those upgraded to do anything. So I am excited about that. My biggest, biggest worry is I don't want patients to do this and not realize there's a hallucination effect. I'm always saying, hey, if you do this, please make sure you check with your doctor or check with the right specialist because I'm very, very worried that especially outside the United States where someone can self diagnose, walk to the pharmacy and get some drug because they think they have X disease, in reality they don't.

That's a great, great transition here. Promise. I didn't tee that one up, but Mike, thank you for the question. Mike is asking what are the cautions for non medical people seeking self diagnosis on chat GPT. Fantastic question. Yeah. Should, should people be doing it? What, what are the cautions? Because everyone's using Google anyways, right?

Yeah. And I'm old enough to say I remember the days of Dr. Google when it first came out and people would show up with these papers and like, look, Dr. Google told me all this and I remember my attendings and my boss hating it. You know, here's the truth I'm going to caution you with check your references. The best reference is going to be a health care provider for liability and legal reasons. You know, I would say don't do it just because even if it's a 1% chance of hallucination, that's a lot given your health. It's not like we're picking a good stock or bad stock. And worse comes to worse, we lose money. Here is your health and that's worth gold. And so that's why I'm very, very cautious. I'm easy talking to doctors because I'm like, look, here's the tool, here's the good, the bad. But ultimately you're a doctor and you have medical training and you know what's right or wrong. My worry with ChatGPT with patients is that it can help, but it's so convincing that if you're not medically trained, it could say a half truth and half false. And then you're like, oh, this is good, I'm gonna go with it. And ends up not and may hurt you. And so I'm more worried about that.

Yeah, no, that's a, that's a great concern. Okay, I could, I could keep going on that one. But another great question and make sure to get your questions in now. I can't keep Dr. Castro for four hours, but I think we can get at least one or two more questions, so. So get them in now. Great question here from Dr. Bill Koslowski. Hopefully I got that right. So he's talking about Mark Cuban, his new Cost plus drugs program, saying that it's not about selling prest prescription drugs, but it's actually about selling consumer trust that's built through Transparency. So Bill is asking how can healthcare establish similar trust with AI?

Oh this, this one. I love this question. So here's my answer back. I think the future is going to be we need to have a large language models with food labels. Why not have a little label for ChatGPT saying it was trained on the following data. These are the parameters. Here's the biases in the large language model. Here's why it's good, bad, ugly, here's why you shouldn't use it in India or Africa. And literally it's a nice little white paper but very food label like very simple to the term. That way anyone can open this up and say, oh, okay, I see why this one's good, this one's bad, why I shouldn't do this and then do it in healthcare. The other part to add more trust is we need our own large language model, not ChatGPT. I'm an ER doctor, so we need a ERGPT or we need a healthcare GPT where it's only trained on healthcare data, medical literature, patient information in the sense that it's private information, but it's factual information are doctors with their clinical gestalts. I have 20 years of ER experience. So someone like me working with ER LLM or ER Chat GPT equivalent and just, just reinforcing and then now you would trust it because you're like, okay, this has been vetted by you know, 20,000er doctors.

You know, one thing that I think has changed so much and even since our last conversation, right, because we've, we've done this once before and you know, I'll, I'll make sure to, to link in the, in the newsletter to our, our previous conversation from a couple of months ago. But one thing that you brought up there, Dr. Castro, is, is the kind of the personalized, you know, GPTs, right? Like ER GPT. But it does seem like there is such a, a lean now. Big companies are trying to bring personal personalized health data to, to, to, to the masses, right? The biggest one I would say is Apple, right? So Apple with, with some of their new product, you know, some of their new products, you know, they have patents in to, to give consumers an actually insane amount of health data that goes straight to their phone from their watch other devices, right? What do you see being the future with, with personal data like this? And you know, Obviously, you know, OpenAI just announced yesterday they're getting into hardware. We're not sure what. But where do you see the future of healthcare data specifically with technology oh, this is huge.

Look up digital twins. And basically Abash and I took you and I created a digital twin. And then I can help you. Let's say I had your genome. Obviously there's trust and all that, but say we get through all that, make sure it's safe. I have you, a digital you, I have your genomic you. And then I can start creating drugs that are specific for you. So imagine a world where maybe I'm having to increase your blood pressure meds. Well, now I know your genetics. I know you. And I can. I forgot to mention this. Now with AI, we can lower the price of drugs by 90% to create that drug. So if this thing cost a billion, now it's 100. 100 million was still a lot, but still I can do it. And so imagine I'm starting to create drugs that are special for you, maybe certain diets that are special for you. And so being able to use this large amount of data like a genome, now I can put it all together and create a digital twin. And then I can even start practicing saying, okay, this drug didn't work. This drug. And then when you come to me, I'm like, okay, hey, digitally, this one worked. Let's try this on you. And now, boom. I'm giving you maybe one tenth of the dose, your normal medicine because it's so specialized to you. Now, before I forget, I got to show you this real quick. This is my last prop. This is AI Coke generated. You need to check it out, taste it. My wife hates it, my son loves it, and then my other son hates it. So let's see if you can break the tie. So anyway, I had to show this because Coke made the formula using AI and used the label. I have no ownership in Coke stock or anything like that. I just had to share it. Just because this is an AI show this.

So many. I. I think you're also the only person to ever bring props, which should be a requirement from now on. This, this is so fun. It's just like, what is Dr. Castro going to pull from behind the desk that is AI related. So, so fun. All right, one. Another question here, a great one from Tom. Thank you for the question, Tom. So he's saying, I told. I'm told we will hit a legal barrier in our attempt to help users explore their health data with GPT. It's a huge missed opportunity since it's. Although imperfect, it is quite accurate. So do we need to wait for GPT 5 or 6 before we can offer it to our users?

No, I, I don't think so. I, I think like that example of the discharge instructions, we could do that. There's other stuff that I'm helping companies with that we should be able to do. You know, the answer is, and I know this is going to sound weird, but there are, I think the future will hold some people way doing waivers saying, you know what, pros versus con, I'm going to sign this off, I'm fine with the risk and I want this done. The last little weird ethical thing that I say is I predict that we're going to start seeing crazy cool health care applications in Africa before we see them in the US because of this politics and legal and lobbyist groups and all this saying no, no, no, we're not going to do this. And then other countries that are risk versus benefits that are a little different, they're going to say oh no, you know, we got parts of our country that have no health care. We're going to use Elon Musk Sad light, we're going to bring some computers, get some PAs, and we're going to give healthcare to the masses because that's the only way we can. And, and people are gonna be like no, that's not gonna happen. I, I predict that will happen.

Oh wow. It seems, it seems like an obvious fit, right? Like for, for other areas in the world that don't have access to medicine. And, and here we have access to medicine like we, we have access to technology. So that does seem like a, a pretty natural fit there. Another question here from Babar. Thank you for joining us. Saying Dr. Castro, how do you see AI playing a role in resource limiting limited settings? So kind of like what, what we were just talking about. But here's, here's a great one. Places where even the use of electronic record systems are not in place.

Yeah, that's a good question. I, I think I, I love old school. I think it's a balance of that human plus AI putting together, but I don't think we're far from that day you mentioned it. Apple already has a 34 million parameter large language model in their phone. I predict that the memory will go up and the size of these large models are going to start training them to be smaller. And so I really think it's just going to become an inefficiency thing. And here's the thing, for healthcare, you don't need ChatGPT, you know, whatever 2 trillion parameters you might be able to get away with. Llama, and Llama has three models. Their smallest model might Work for a lot of the things that you're already doing. Create like a mini electronic medical record that's already on your phone. So this thing is really going to change things.

Yeah, that's something that, unless you're following AI very closely, I don't know if people really understand the implications of having a large language model on your device. Right. So that means you don't need to upload any data to the Internet, you don't need to even send information to a, you know, an open AI, a Google, a Microsoft. Right. It's all happening, it's all processing on your phone. And you know, Microsoft, Microsoft just announced this, you know, with their new Surface hardware being able to run a large language model locally. So. Yeah, like, like how, how might that. And, and obviously we don't know. Well, maybe you do, but it's, it's brand new, you know, it's, you know, running a large language model locally on a device is, is very new. Is that going to be used in the healthcare setting? Is that kind of a, a temporary quote unquote solution for, you know, some of these issues with, regarding patient privacy and, and you know, giving data to companies? Is that going to be a thing?

Yeah, I want to explore that for a second. You know, I'm gonna give this idea away. This was something I was gonna do myself, but now I'm like, you know, what if someone's listening, they want to do it, go for it. I think the future is this, with this what you just said, a clear example. Imagine you were walking and you find someone down and, and they're passed out unconscious. Their heart just stopped right in front of you. They just went down. And you're not a doctor, what do you do? You just call 91 1. Well, guess what? It may take five to 15 minutes to get there. And if they don't get any oxygen those first five minutes, they're dead. What if your phone had already the large language model of CPR and everything on there? And now with augmented reality, you're showing the patient and it's telling you, hey, CPR is like this. And then someone can be beholden the phone or you can be doing it and it's kind of walking you through. And now you're giving cpr, now you're giving that first cpr. And if the latest technology changes now the phone already has it in, built in it. And then for that person that said, hey, what if you don't have Internet? That's the point. If it's on here now, it doesn't matter. You could be in the, in the middle of the Grand Canyon with no service and this thing will tell you what to do.

Wow. Yeah, that's a deep one too, right? It changes where you can even practice medicine. You know, something, something, you know, I'm curious about because maybe, maybe I'm in the minority here but, but I don't care what my health care provider does with my data, if I'm being honest. And, and yeah, maybe I'm in the minority, but I would in theory waive my right and say yes, upload, upload my, my medical charts, upload this, upload that. I would do that for better, more accessible, more affordable care. Because at least here in the US sometimes to see a specialist list, you're talking two, three, four, five months. Right. Do you ever see that being an option or might there be a large company that, that, you know, shakes up the healthcare system and they're like, hey, come in here, you opt in. Yeah, we give all of your data to, you know, these, these large language models or whatever. But on the flip side, you're going to get way faster, way more accessible care. Is that, is that an option or is that that way too far fetched?

You know, I, I, I think the, the sad thing is risk versus benefits, we're getting there where, you know, it costs so much money. I pay so much money a month to just have healthcare and I'm sure I'm not the only one. And so again, the risk versus benefits, I, I, I would actually pay less and be able to get something like that. I'd be happy to sign my part away. And to give you a real example, people are traveling to other parts of the world to get healthcare. You know, it's not uncommon to someone to say, oh, they went to X country to get this procedure or that country to get that dental procedure because it's just cheaper. And so we need to do something because we shouldn't have our citizens have to leave our country to go get health care.

Yeah, absolutely. And you know, again, I know we have people, you know, tuning in from, from all over the world and thank you for sticking with us. I know this is a much longer conversation, so this is going to be my last question, I promise and then I'll give Dr. Casper the, the option here. But will that in the future happen? So, you know, and, and especially when we look at accessibility and in different countries and you know, yeah, people from the US specifically have to travel to other countries sometimes just to be able to, seen to, to be seen sooner and, and healthcare here is extremely expensive. So. So is that, do you think, is that a viable option where, whether it happens somewhere here in the US or else or elsewhere, because at least here in the US it seems like there's an opportunity for something like that because we have very expensive, very slow moving healthcare system, yet we are the only country that essentially has no guardrails when it comes to AI. All other countries, especially in Europe and in other places basis, they have very tight guardrails on AI. So it seems like that should happen at some point.

I think so. I think so. And hopefully someone out there is going to be calling us and say, hey, I got an idea, let's do X. Because I'm going to help out whatever it is. At the end of the day, my goal is the patient. We can help the patient. And all of us, including myself, we're all patients. And so let's do this as a community. One thing that everybody could do is honestly share this show because the more people hear it, see it, they're like, oh, that's a good idea. We should do X, Y. And it's not about us, it's about others. So let's help others out there.

Love that. All right, I took up way too much of your time, Dr. Castro. Thank you everyone for, for tuning in, for listening. But I want to give you the stage. What is your one piece of advice, your biggest takeaway? Whether it's for someone in the medical community or someone who is maybe just a patient and who is eager to learn about the future of AI and healthcare, what is your one piece of advice that you can give to both or either group?

Yeah. At the end of the day, I feel like everything's about education. Become a life learner. Take this and say, you know what, I'm just going to learn about AI. It's going to help me or help my family and just make it a hobby or make it a focus or make it an interest. And if it's AI and healthcare, great. But if it's AI and stuff every other day, AI, whatever it is, just do it. The main thing is I'm, I'm a doctor, obviously, so my main thing is let's help others. So whatever knowledge you get, just make sure you give it back to others.

The knowledge piece is important, right? Because there's, there's bad information, there's a lot of incomplete information. So yeah, it's important to, to study and to learn and to keep up. And I thank you for that, Dr. Caster, because you helped us. I think you just gave everyone listening, whether they are working in healthcare or they're just interested as a patient. I think you just give everyone a large, a large chunk of knowledge that they can take forward and improve their daily lives. So, Again, thank you, Dr. Harvey Castro, for joining the show. I very much appreciate it. All right, cool. And as a reminder, that was a lot. There's so many resources that I talked about. We're gonna plug them in the newsletter. So make sure to go to your everydayai.com sign up for that newsletter. We're gonna have it out in a couple hours. Also, if you're on the podcast, make sure to check the show notes. We're going to leave a link. You can come join the conversation on LinkedIn. Leave a question for Dr. Castro. He's always answering questions. He's always networking. He's always bringing people in and sharing knowledge. So make sure you do that. So thank you again for tuning in and hope to see you back for another edition of Everyday AI. Thanks, y'.

All. And that's a wrap for today's edition of Everyday AI. Thanks for joining us. If you enjoyed this episode, please subscribe and leave us a rating. It helps keep us going for a little more AI magic. Visit your everydayai.com and sign up to our daily newsletter so you don't get left behind. Go break some barriers and we'll see you next time.