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. Hey, welcome to another episode of Real Talks. Scott, welcome to 2026. I know it's crazy to think about it's a whole other year.
And we get to talk about this for a whole other year. I'm very encouraged by the fact that we kind of got renewed. And this is super exciting. I mean, there's so many cool things that I think are going to happen.
So many things that have happened. And maybe one of the biggest things that's happened in the last year. I think you're going to talk about news in the day. That's that's right.
I hope you had a really nice holiday season. It was good to take a couple weeks off. And I am excited to sort of get back into it. As we always say, it is just astounding.
As even as we both watch this, feel very closely astounded by just the pace of things. And just the other day, chat GPT has launched something called chat GPT Health. And you know, this didn't come out of nowhere. There are apparently 300 or so million conversations, conversations, a week that have to do with health care.
I mean, I know I do it. I know my parents do it. Relatives do it. My patients, of course, definitely do it where they put their symptoms, medications.
And what does this mean? What are my diagnosis with the uploader, their labs or whatever? People are already doing this. So chat GPT, create chat GPT Health.
So what does it actually do? So what it does is it actually lets users connect their medical data with like their labs and medications, visit some rays. And here's the king. Here's a key thing, Scott.
Like I was surprised by this. They are connected as far as I know, I mean, information is still sort of coming out and bits in the area. Yeah, a little bit. It was just I just found out about literally today.
I was just announced today, first part of January. And it connects with EHR ecosystems like epic and large lab providers. So it actually plugs in. It's integrated with EMR systems or EHR systems.
And so it will plug in. You could say something like, hey, Remind me some rays of my last three visits. And they have a force. They say they have guardrails.
There's no diagnostics that are in there. It doesn't make clinical decisions. It doesn't prescribe. It doesn't write back into EHR.
But this I think is a huge first step and also it connects with wearables. Think like for Apple Watch or your, your, your cometer that people, you know, see people wearing on their arms. This is, I think Scott, when we look back on it, you mentioned this could be the major news of 2026. This is, I think this is big because.
This is a recognition from AI companies. I think healthcare is going to be one of the most disrupted fields that are touched deeply within AI. This is just. The chat you beauty health is right is right.
In the next year, I think, we're next year, such a monumental shift. Yeah. And so. I'm really curious how this plays out.
I'm very curious to try it. Scott, what are your thoughts? You know, I, I, you know, I, when I read that come out this morning, I was on another conference call and one of the other guys in the conference call, one of our friends, Easy on Yamaha was on the call. And he's like, look, which just came out.
And I literally stopped the conference call. I'm like, oh my gosh, really, that's pretty cool. But then I read what you just said. And it made me go, hmm, which is it's going to integrate with all these EHRs.
Well, we know that it's almost impossible to integrate with I carry HRs across the board for anybody. It's a really long time. And I got to be epic and all that kind of stuff. And it's going, well, you're going to be able to plug in and get your data and blah, blah, blah.
I'm like, yeah, there's so many security, cybersecurity, hippoprivacy issues with that. We'll see, you know, I, what I don't want it to be what I hope it's not is Dr. Google with a different name. I hope it really does do what it promises to do.
Maybe call me a tiny bit of a skeptic on this one. Because I think I'm a tiny bit of a skeptic on this one. I hope it's the first step towards integrating health data across systems. You know, we talk about in many of our podcasts, we've talked about these APHIS systems, which are personal adaptive personal health information systems.
And you know, the reality is is maybe this is that first step towards that. I think that this is going to be one of those things over the next three to six to nine months ran. We'll probably report a lot in the news and go, is this really working the way we think it's going to work. I don't know.
We're going to see. I, I, I encouraged, but I don't know if you're saying he is. And you probably remember I, we talked to Dr. Perique about this last year, we talked about the difference between Dr.
Chattsy, PD and Dr. Google. And there's a difference. And you see it with your patients.
Dr. Chattsy, PD is much more confident and authoritative probably and not necessarily for, you know, for any good reason, but it really sounds that way when you, when you punch something into chattsy, PD or Gemini, it comes across as the truth, they grow on truth. And so that's, that's the difference. And so I wonder about this.
My thing is, I think people will use this, but then put it right back into it. If it doesn't give a diagnostic diagnosis, you could just then use regular chat to you to check, chat to you to get diagnosis. So I wonder if people are going to start doing. I'm not referencing there.
Yeah. Yeah. And I get your form of the HRs, but I think we're already starting to see EHR systems, even in the optical space, integrating and using some type of LN for, for doctors and I think that's definitely going to be coming. I mean, I think it still comes down to a data problem, right?
We always, we always talk about date, your famous phrase is, data is the new gold. I think data is the biggest challenge we have is, you know, GPT has got to learn from data. And I really think this is a smart, strategic move by open AI to start gathering data to learn and teach from. Because we know it can't really get into HRs and really take the data yet.
But they're kind of around the table in a very openly sneaky way, saying consumers give us your data. Well, they very explicitly say they're not going to use health data for training sets. I see you smirking on video. And so we know how that sometimes, I mean, turns out this was after all, a company called open AI, which, you know, was supposed to be a nonprofit.
That's a whole separate conversation. But very interesting. I think we're starting off 2026 with the bang. And I, and like you said, I think we're going to be talking a lot more about this.
I agree. And that brings me the subject. No, so today for our knowledge bite. And you know, we've gotten a lot of feedback from our listeners that they really like this piece is a way to educate, get educated about the different terms and concepts.
And many times we try to provide something to think about. So. I think I don't think it was our last podcast. I think it was the one before that we got talking about synthetic data and I alluded to three people after that when live go.
Scott, you need to explain what synthetic data is we're not kind of getting it. And I was kind of like, okay, maybe that's a good talk. And so. You know, and kind of as we just talked about, you know, to build smart AI, we need data and we need a lot of data and we need standardized data and we need to have been cleaned and scrubbed.
There's just a lot of data. Now that's granted that saying that we stay with the large language model process, like what chat GPTL is doing. Saying let's lose a large model process and I think maybe in the future we're going to do the future of analytic AI, which may make LLM's. Absolutely already, but for future reference, you know, we need a lot of data and you can think of data as.
Clinical experience in our case, right, the more clinical experience we have the smarter we get, ideally, you know, and we learn from that experience and we try to use real or data. We look at pictures and we look at ultrasounds and we look at scans and we look at corny as we look at retinas and we look at other medical records and we go, okay, here's this isn't we learn from those things. But there's two really big problems when it comes to getting data for AI ones privacy kind of got to just mention and the other one is scarcity. You know, we can't just hand over millions of private records to a computer, private medical records to computer and say, oh don't don't learn from this don't don't find stuff.
And so I think we're at that really kind of paradox right on where I'm to the audience where we have to have this data to train the models. But the data is not that good to train the models and nobody wants to give it up so we have the scarcity privacy issue that's really a big hurdle for us to get through and this is where the terms synthetic data comes in well, synthetic data is for all intensive purposes in air quotes fake data that's mathematically real. I know that sounds a little weird. This is data that the math holds up, but it didn't really come from any one person.
It isn't collected from the world. It's actually manufactured by an algorithm to mimic the patterns of the real thing. And there was we're making and I think I think it was Rania that did this a digital twin of the data. Basically synthetic data saying we're going to make a digital twin of what we've learned out there.
And we're going to apply math to it and we're going to make a new model so like imagine this if you want to teach AI to. I don't bake a cake, but you don't want to show it photos of a real cake that people made. So what do you do you have to create this 3D digital simulation of what a cake looks like and you know you try to give it texture and height and chemistry and. But you build all this but there was never a baker there was never flower there was never sugar and there's no frosting what's the point eating the cake.
You know it's a digital twin that looks and acts like the truth without all the negatives the baggage of reality. So you know a real world example would be like I have one of my patients is he works on radar light our excuse me for self driving car and he was we were talking about the subjects of the day said you know self driving cars be safe car and he's to know what to do. And in a case that's odd like a toddler chasing a ball in the street I think we talked about a couple weeks ago during a thunderstorm at night well if we waited for the exact thing to happen in real life to record it it takes decades before there were enough of those the car could learn the system could learn from instead engineers make synthetic simulations of that data they build a virtual city they make it almost very hyper realistic they create a toddler and a ball and a story. All in a storm and they run the simulation through it now we think all that's cool that self driving cars bubble up but if you think about even a better real world experience example what do you think that pilots for the airlines train on they train on simulators that give them if we had to wait for every pilot to crash a couple planes to learn how to not to crash a plane aviation would exist.
Simulations that we train pilots on well all this is doing with synthetic data you make a synthetic simulation well is it that far of a stretch to start making synthetic simulations of a health care situation that we can train our AI on instead of waiting till we have a human biological being that if we make a mistake something bad happens. So at some of the thing about the synthetic data is just a simulation that we can learn from without putting anybody in danger and the reality is I believe by 2030 you know most of this AI did training data will likely be synthetic rayon because we can't get it from the EHRs it's not standardized as we said it's not clean you know it it's not regulated I think that most of our AI day is going to come from synthetic situations that we put our algorithms through and then human supervised learning goes yeah you didn't make the right answer there that wasn't the right idea you know I think that we're going to train models that are more private less biased. And maybe then we can start creating situations humans haven't even encountered yet or they haven't understood that that was what was out there right it's not just going to be fake data. Or fake information I think synthetic data might just be the fuel for the next generation of secure non scarce intelligence that algorithms can learn from.
Exactly right I think and I think synthetic data is great especially for surgical education this is something that's been used in ophthalmology for for some time the first time you do a complex case should be on a on a simulated diet like a flight simulator like you very regularly mentioned. And so I think that's very helpful for our audience I definitely learned a ton there thank you for that alright so this is our first podcast of 2026 and so we're going to end this with our fireside chat about prediction times Scott predictions for 2026 the crystal ball we always give crystal ball which you know look if. If I really had one that work that I probably wouldn't be doing this podcast with you let's just be honest i'm going to be somewhere but we probably would be still practicing medicine man yeah that being said my prediction for 2026 is that. Akelamics is going to go mainstream when I'm on my LinkedIn profile i'm just shocked by how.
This term which appeared like you know maybe five or six years ago and you know we've known about it for a long time as I care providers of course the eyes window. I really started talking rayon until about the last year that's when we really started like it became almost mainstream conversation it and so my my so what i'm actually predicting because it's sort of right there. Is that my primary care physician colleagues are going to be coming to me and saying hey do you know what I could do with a retina photo now I can. Did you know and so I think it's going to go starting finally the people who are actually going to be using this because it's not truly for i care I mean I think that's one way but really this is for primary care and really it's for retail care and screening environments and if that that's.
Information the people who really need to know about ocular mix are finally getting the message sort of loud and clear and that's what I mean about this going mainstream I think the claims are going to be. Rightly so that maybe narrower it's not going to be about just general age or anything like that it's going to be specifically on cardiovascular risk kidney progress you neurologic maybe some neuro degenerative stuff oh yeah and I think that's going to be huge. I think reimbursement will be a major bug a boo I think value based care is going to be a big big player in this and maybe even just. Patient direct patient payments so that's my prediction for 2026 I think we're going to see see it become more mainstream and the biggest one is that I think the FDA is going to give a clearance slash approval for one of these.
Types of devices that are currently being studied I totally agree and I kind of want to jump on top I want to kind of add to that I agree 100% but you know ran I also look as I think and as maybe a term on kind of making up because I don't know if there's a really good term. Tele diagnostics and maybe we can wrap remote diagnostics as that wrapper but I really think we're going to see a major move kind of like you said is that cameras might start showing up retinal cameras won't just be in I care offices I think we're going to see tele diagnostics not necessarily tele medicine like hey we're going to treat patients completely over the internet. But I do think tele medicine remote diagnostics you know like you talked about earlier in the news of the day chat GPT is going to link with your Apple watch and your or a ring and you know all your biometrics. You start looking at biometrics and start doing you know remote visual fields with heads up displays and you know I think that's going to be the biggest single shift we see in the next year is wait the data that we're getting won't always be coming from.
I care providers office. I think that's going to be a big one. I think there's some other technologies like I'm super interested in seeing where and we're starting I'm starting to hear the bubbles coming up again on this one about the smart contact lenses. I happen to know three different companies that are actually working on these that are these integrated smart contact lenses and even a company is working on a smart.
Intraocular IOP measuring device that you implant and leave in the eye. Now you feed that with Bluetooth. I think the definition of glaucoma and you know ocular pole sample to and what's happening is a completely different beast. You know we I'm super interested to see where that technology is going to go.
Contacts aren't going to be just about vision anymore. I think they're going to be about diagnostics and you know very similar I think that you know the heads up display that that revolution is in full full March right now right I mean all the big fish the apples the apples these apples they're going to get all colored and dark circles the you know the mesos and 25 other ones are all working his heads up displays. I think that we're going to see a lot more and how we integrate with the optical and you know marketing and all that I think the whole diagnostic tech I predict that's going to be a really big thing in the next 12 months. Yeah.
Yeah, both of our predictions have to do with AI diagnostics that are looking at the eye, but maybe giving diagnoses well beyond the eye. Well, and I have a dream to have a hope. Well, you know what my biggest hope for 2026 is? What's that?
The hope that the industry starts to realize that AI is here to make our life better. And you can't ignore it. It's coming. It's here and it's going to change the way we work.
That's my that's my that's my goal. That's a perfect way to end out the first podcast of 2026. Got appreciate appreciate everyone's time. You could find this and all our content at AI in I care dot com.
And thank you again for listening looking forward to chatting with you soon. You've been listening to real talk an AI and I care your weekly podcast to keep you informed about AI technologies revolutionizing I care.