Welcome to Real Talk on AI and I care with your host, Dr. Scott Morris. And I'm co host, Dr. Rehan Ahmed.
In each episode, we discuss current news, adding a little AI education, and debate innovative AI tools and topics that will change the industry. Hello everybody, this is Scott Rehan today in our news of the day. I'm going to cover something kind of interesting and I don't know if any of you guys have played with this. They just had the Google IO event, 2025 event, so that kind of dates us what day we're at.
They released Google, meaning they release what's called VO3 VEO3. And it is their newest AI video generation model that makes these little eight second videos. And we've been playing with them in our office and through my artificial intelligence company. We've been playing with these for about a week now.
And I'll tell you what it is shockingly realistic videos that they create. You can prompt just about anything. We use a whole field of actors or actresses, different languages that you can put different backgrounds. You've actually I've started seeing some of them and now I kind of know what to look for on some of the social media feeds.
I'm like, oh my gosh, that's not really a human doing that. That is an AI avatar. And now they're starting to make movies out of this. I mean, some of the filmmakers are taking and building a whole series of eight second clips with the same model, same actor, same whatever, same background.
And they're building entire movies in days with AI with a zero basically a zero film budget. And it makes me big like that kind of technology is pretty spectacular. And like I said, I mean, right now it's getting to the point where I'm starting to get to the start and have a hard time telling is we've been playing with these in our own office and building some of these in my staffs getting in and out now. They're like prompted different stuff about how we use these for own advertisers, for advertisements.
It used to be me getting up and filming something and then putting in our social media. The reality is now we can have some other AI generated one. Do all the stuff we want and people going to watch it in the language. I think as we were doing our pre-meeting, you know, you brought up a great point, or I honest that, you know, this is.
This is almost getting the point where where it's coming a year, we kind of wonder who's going to survive. Like who's going to be the AI generation company, AI video generation company? So maybe there'll be a few, but I think the tools keep getting better and better and better and better. Right now, I think V03 is definitely the most advanced tool that we can access publicly.
But I'm sure that in a month now, we're going to probably go, oh, there's a new one that's even better, but that's kind of the news of the day. I think it's going to change and it kind of prompted us to think a little bit of what we're going to cover tonight and we'll kind of lead into that. So next I always we have our knowledge by it and round I'm going to turn that over to you and kind of cover what you want to cover for our knowledge plate. Yeah, I know excited talk with the not just to sort of a double down and a lot of fast things change.
I remember doing deep dives on on these types of video generated avatars. And a month ago, I was about page and other from the agent, another couple of tabs, these are companies, wow, and then. At four weeks later, it's something else and it's just amazing how fast these things are getting better. I made one couple, you know, called months ago and I'm excited about and it was good, not great, but I'm excited to see and I know the one I tied, that's the worst it's ever been.
I'm only going to get better. And so yeah, it's going to be really interesting and think of all the I am I care applications, it's just models of my thinking about how. You know, you sometimes want to felt there are only two of me, how much more work I get done or how much more I could, well now there could be two, three, four, ten in different languages, doing doing all sorts of the behavior, we shall see. But yeah, I wanted to go into the knowledge right and this actually is related to the knowledge right today is about conversational AI and specifically something called natural language processing and the best we had talked about is probably to do an example like an eye care example, something I think about a lot, you know, you know, I'm madly and slow play that night and you know, I'm not the model so you get a call from a patient after hours and you know, it's something like the bad questions about their eye drops after cancer, too.
But it gets rather to you because this is a H5 sensor, would it be nice? It's possible if that patient to be triaged through a conversational AI agent and this is where NLP and AI agents it's good to know something about so that these topics come up, you have some context start thinking about it. So NLP, this is a subtlety of AI that remains computers to understand interpret and respond to human language in a sort of a meaningful way. So we probably start you thought you remember clicky back from the Microsoft.
Oh yeah, first that was probably one first that was not really a conversation, but they were trying, my stuff was ahead of like, I had it occur back then. But they were that was sort of like the first iteration, of course, now we're you know we see really and of course, chat GPT now all in the region. So it sort of seemed like second nature now and away, but man, this has been a long history. So NLP is kind of like the cocktail some brain to analyze light structure and provide relevant responses.
So the way it doesn't super interesting it transforms words and sentences basically into numbers into numerical form and these are called vectors and these vectors are a math, because it's computer writes all 0s and 1s, they represent what I'm called represent E.D.s and our relationships between words. So allowing the AI to sort of recognize similarity and sort of come so for example, the vectors like I vision, the biology of country are all closely related and they're different vectorization algorithms, certain that there's several different ones and they sort of chunk them in different ways. That's sort of the way to think about rather how it does it. But but conversational AI is way more than just NLP.
It speaks recognition, it's natural language understanding where you're sort of term, you can text meeting, it's dialogue men, it's natural language generation, for example, I mean, because you got respond. So those all work together to create what we call conversational AI. But when you know, in truth, Scott, I think you agree, like when we're talking about it and in short practice NLP at sort of our level, NLP conversation AI, certain moralists anonymous, what would you agree with that? Yeah, I mean, I think yeah, I mean, I think NLP is a part of conversation AI comes as I as the big umbrella of it and NLP is an important piece of it.
I think you know, so through translational services into that group, you know, I mean, something I think you're going to have empathetic AI that falls into that conversational piece. I mean, I think it's a big subject, but it's and you know, it's I want to tell you like this, it's text to talk and talks to text. Yeah, but it's more than just talk it soon, it's going to be as we just mentioned in the news thing, it's going to be video, it's going to be interactive or in your relationship kind of stuff. I mean, I you know, NLP is language that's at the end of the day, it's language that we understand whether we're listening to it or we're speaking it or we're interpreting it, it is, you know, I always think of all of AI rayon is kind of like the human body, right?
We have machine vision, which is visually eye, we have auditory and speech, which is NLP conversational AI. We have biosensory AI, we can smell that's a, you know, chemical testers blood analysis, that kind of stuff, kinesthetic testing. I think that all of that is, you know, AI is taking on all of the senses of what we do and this is an important part. And I think that communication is what brought humans to to be the Elf is on the planet conversational AI is what's going to help AI.
Yeah, be successful as well. Exactly right. And so thinking about an example that we could use in clinic, right? So imagine you had a chatbot that was a conversational AI agent and it was it was customer facing to respond to basic inquiry.
So Scott, you may have a you may have a conversational AI agent at patient calls and ask, hey, you know, are my son's glasses right? And so the way this will work work through the conversational AI, well, first the NLP would analyze to determine the intent of the message. Okay, well, the intent is checking in order steps. Then it extracts critical details.
The critical details son's glasses. And it could be other things right so we're in talking about vectorizations. The person said, are my son's glasses ready? But they could have said something kind of awkward, kind of Yoda S.
say, are the glasses of my son. Right. And it would it would turn over to the tosses of vectorization. If I just describe it would sort of chunk them as the same or very very somewhat.
And then of course your conversation like a AI agent would respond to generations. And yes, my post losses are not words pick up, etc. Now for instance, just the basic example. We all know of other examples of me.
I feel like every time I call my bank now, they're getting better with conversational AI agent. And the possibilities are endless both, you know, of course, in health, Jared outside of health fear to handle routine questions. What I am most excited about this end of the same. This is from the example early on is the patient triage idea is that family, they are questions that people will have that can be answered by an AI chat.
Or there are questions that should should be escalated to have a human and right. And then having the AI sort of filter out what really needs to be escalated to what doesn't. That's where there's so much value. Right.
And one of the bonds to consider right is kind of a lack of a lot of nuance, a lack of empathy in sky you mentioned that that you know AI as a yet cannot fully replicate the nuance understanding or the demands of the patient right. The patient makes it okay. I know there's an AI, but I really want to serve it to be answering this question. It's that important.
And of course that parlay is the second thing. There are a lot of medical devotees, especially if the patient thinks they're talking to you. Right. But in fact, it's just the agent.
And so there are a number of issues, HEPA issues compliance issues, you know medical board really should that need to be addressed that I think we're still really getting some but it's a conclusion. This is a huge part to have some understanding of what conversational AI and LPEE and all fit in. And so that was the point of this knowledge rights to give some little bit of background on this that were we have a little context when this topic comes up super excited about this is a core part of AI and future and happy to share this knowledge right. Awesome stuff.
It kind of brings us into our fireside chat that we do and you know I want to take this conversation. I once that further go I've been thinking about that a lot, you know in terms of how well marketing or awareness campaigns work in any field but definitely in medicine. But right now it's always been a traditionally we're all sick of watching those damn 15 second commercials that are 14 seconds all the things that can go wrong and one second what the drug is with some, you know, pretty actor or actress. And at the end of the day, it's like I didn't really want to watch that anyway.
So I wondered like how will conversation AI and whether it be through the use of our own personal health care agents I call them PHAs are personal health agents. So it'll be more of a customized approach where you're not necessarily using an agent but companies take a more personalized and customized approach with avatars whether they be humanoid like or anthropomorphic more cartoon like type of avatars. And I've been really thinking about that a lot lately and I think you know I almost envision a spot where somebody at some marketing company pick company ABC pharmaceutical device product optical you know doesn't matter I mean I was like they're sitting there going what if we instead of did it some storyboard where we're going to create a 15 second video that airs on NBC at the Super Bowl you know or whatever and that no one really wants to watch because if it's not about something they have or something they know about then no one's going to pay attention to it right and that's just wasted millions and millions of dollars not just in development. But also just in promotion like sending it out there so how could we more specifically target and engage the people and make people aware we're in need of a specific product service drug whatever at the time and I think about that is like you know I was going to do this other day and I thought like a headache okay so I had a day where I had like four or five people and wrote a headache and I thought you know what if what if instead of us telling patients about this medication of this process.
We instead have a agent or an avatar whatever that case may be that actually listens to them and says you know not really a chatbot not something that goes oh I just have a pre you know a preloaded message that I'm giving but something that more that's a good way to say this more than like it says I want to give you a symptom and if you have that symptom would you like to find out more things about this well first tell me more about your symptom in the agent now becomes their interface between a practice or a company or whatever and ask them all the right questions like you had mentioned this in the knowledge by ran about you know maybe they have some algorithm that they're going through that they're asking certain questions and getting feedback from potential patient because I think the days of us as portions of the day and say us but of companies going promoting this is our message I think with conversational AI that everything is going to become customized and it's going to start with the agent listening to the patient developing a storyline behind that pulling out the key pieces like you just said and then learning to ask certain questions to get more information to help customize that towards that particular individual. It's going to be way less than millions of dollars and the only people that watch it are the people that are actually care to watch it I think about this conversation and I think the entire structure of marketing awareness of what it looked like for the last 20 years I think from five years and five years from now where you look back and go can you imagine that we did it that way I just think that our avatars are the tone their empathy they're all that it's going to change. It's going to change how we interact with the world instead of it being about what we tell somebody it's about what they tell us and how our agents or avatars or however interact I think it's people want connection they want people to listen to their story and I really think the future of conversation live is having AI listen to somebody's story and distill out the important points and then help answer questions they ask. So we have about what are the next steps where do I go search for solution what are the solutions is this not my need to worry about is this something to go see somebody for today tomorrow two months or now I you know I think it's all different it's not it's no longer going to be a per sure video or something like that it's going to be a conversation.
I mean I'm open to suggestions open to open a rebuttal and now but I've been thinking a lot about this the last couple weeks and know you're exactly right and I think that first place where this has had a big role is obviously in a ensure mental health there are several properties that's been out with chatbot and really just get one. It's where because it was just textual and it was truly was conversational AI and that's what the knowledge but if we take it a step further and I love your analogy the human is that well it's just one type of that's just text and speech that's just mouth and ears but when you put vision computer vision AI then you're talking about man this is a conversation I had plus an eye doctor because you know you could say well you need to has an eye doctor you need to look at images and you need to be able to answer. Analyze doesn't analyze data. I saw just pure conversation all it's like well now you know we can't have it yeah I machine vision getting good enough to do that's right that's right and all gets packaged in with these series of agents that so all sort of work together with the sort of a super agent at the top that's sort of collecting all the information so yeah I I just like to see this you chair certainly in the same in the same way.
Well, but then you wonder you know will that you brought up this during your come during your knowledge by two is that you know I think about that is now you have an agent who's listening having a conversation and by no means do I am the first person to say I do not believe in any way say perform that any I is going to replace us as providers. But now you have an AI that can go hey this is more nuanced as you said then what my algorithms teach me to do it is time to upgrade you or to whatever you want to call that. And now you have the avenue of tell a hell yeah and taking it once that further what we talked about a week ago maybe you're upgrading it to tell a health to a human but that human doesn't have to be in your geographic area that human can be anyone. Who's there's your 24 seven practice right maybe i'm hey i'm a morning person right i'm up and moving at four o'clock in the morning i'd love to see patients from four to six and be done right but four to six for me is a European patient a patient from East Coast you know and that's what i'm doing my telehealth conference with them because it's convenient for me.
I don't know now you've changed not just how care takes place. And who is doing care and how that cares being provided but where that cares taking place act I mean I think that these conversation lay eye in these agents it's the next step into the evolution of telehealth telemedicine and taking care of people 24 7365 no matter where in the world they are no matter what language they speak. Yeah yeah that's a I love the concept of adjacent technologies of telehealth so you have a mode devices smart devices big data and the next row AI on top of that that really is i think those are all the ingredients for a major transformation that's a paradigm shift in the way care is delivered and maybe more importantly the way is received. Thank you everyone for listening to another episode of re-applied this is grayhanging Scott please visit us at our website AI and iCare dot on where you can read all our articles and listen to past episodes of AI and I care you real thought thank you.
You've been listening to real talk an AI and I care you're weekly podcast to keep you informed about AI technologies revolutionizing I care.