The Crackin' Backs Podcast

AI in Medicine: Are Doctors Still Making the Decisions?

• Dr. Terry Weyman and Dr. Spencer Baron

Share episode

Are medical students and young doctors really becoming better clinicians — or just better prompt engineers with stethoscopes? In this episode of Crackin’ Backs, we go deep with Dr. Ti Pence to unpack AI in healthcare, how it’s reshaping medical training, clinical decision-making, ethical responsibility, bias, and what the next decade of medicine might look like.

From daily AI use in study and documentation to hospitals quietly embedding tools into clinics, we ask the hard questions:

  • What happens when medical training prioritizes AI fluency over clinical judgment?
  • How can future physicians use AI to augment — not override — human expertise?
  • Where does bias pose the greatest risk in real-world medicine?
  • Who should be held accountable if AI harms a patient?
  • And what does an ideal AI-powered healthcare system look like by 2030?

Whether you’re a clinician, medical student, health professional, patient advocate, or curious about the future of medicine and technology, this episode challenges assumptions and shows the human side of AI in health care.

About Our Guest — Dr. Ti Pence

Dr. Ti Pence, DC is a clinician, healthcare strategist, and thought leader at the intersection of artificial intelligence, clinical judgment, and medical education. With advanced training in healthcare administration and cutting-edge insight into how AI tools are shaping practice, Dr. Pence explores the balance between technology and human expertise — especially in how doctors are trained, evaluated, and held accountable today.

He’s been featured talking about AI’s impact on manual therapy, chiropractic education, and the future of clinical skills — challenging traditional norms and offering practical, big-picture insight into where medicine is headed next.

What You’ll Learn in This Episode

 Why AI is already embedded in how medical students study, write notes, and plan care
 One real example where AI impressed — and one where it missed the mark
How bias in healthcare data can harm real lives
What professionalism with AI should actually look like
The “rules of engagement” every future physician should follow
A hopeful — and a cautionary — vision of medicine in 2030

Whether you’re preparing for med school, navigating clinical practice, or just curious about AI + human judgment, this episode makes complex ideas feel understandable and urgent.

Listen, Subscribe & Share

Available on Apple Podcasts, Spotify, Amazon Music, YouTube, and wherever you get your podcasts.

If this episode made you think differently about medicine and AI, please subscribe, rate, and leave a review — it helps other curious minds discover the show.

We are two sports chiropractors, seeking knowledge from some of the best resources in the world of health. From our perspective, health is more than just “Crackin Backs” but a deep dive into physical, mental, and nutritional well-being philosophies.

Join us as we talk to some of the greatest minds and discover some of the most incredible gems you can use to maintain a higher level of health. Crackin Backs Podcast

Dr. Spencer Baron:

If you're a med student or even a young doctor, odds are you've already used some sort of AI to study, write a note or check, check a treatment plan, which is oftentimes becoming the most fashionable thing nowadays, but probably didn't tell you're attending that what's going on. So meanwhile, hospitals are quietly flooding clinics with AI tools all over the place, yet they still hand out CDs after an MRI when nobody has a CD ROM. So a lot of sense that does not make but so here's the real question, are we training better physicians or just better prompt engineers with a stethoscope tonight or today, we go deep with Ty pence. Dr Ty pence, fresh out of Harvard, AI in healthcare, from strategies to implementation to expose what's really happening behind the curtain. Welcome. Dr Ty pence to the show. Thanks, guys. So listen, after completing this, I'm very fascinated Harvard's AI healthcare program. What was the moment that made you realize AI isn't just another tool, but a true turning point for, you know, medicine?

Dr. Ti Pence:

So I work with engineers all day. So that's, that's a piece of this is every engineer I work with. They're all my every patient I have that's an engineer is being asked to integrate AI into their systems. So it, it started clicking there. And then I think just before the interview I had with you guys last time, I had taken that first course through Stanford that was kind of like a seminar on a virtual seminar on, like, on AI, and where medicine is going to be with AI. And that just made me really aware of how much I did not know and that this probably wasn't going away. And so that's why I was like, I need to find how do I at least understand the concepts that are going to be core skills in five years in medicine and chiropractic. And so that's what got me started on this.

Dr. Spencer Baron:

Well, I'm already seeing our our health records are, are using AI for, you know, some of the ways we can audio tape, you know, a conversation between a doctor and a patient. But let me ask you, you know what? You know, what, from your perspective, what excites you about this and what are you worried

Dr. Ti Pence:

about it? Oh, yeah, this is the wild west right now. In AI, we do not know where it's going. We know what we've achieved, what folks have achieved so far, we can guess what might be possible as the next steps, but we're just guessing and trying stuff out. And like some of the stumble onto ideas are amazing, like I would never have thought of that. Others are are just as humans. We're using this stuff and coming up with it. What excites me is, man, we could get back to just being in a room and talking to a patient and treating a patient and not having to worry about any of the documentation, very little of the documentation at all, possibly where, when you just walk in with your phone, you tell the patient that you have an AI scribe with you, and you just put the phone down, it Does all your notes. It then formulates them to your whatever insurance billing, there'll be another agent that then formulates and sends that out with the pre auth that's already done before the patient walked in. It took all of the patient's notes from all the patient records, and now when you walk into that room, you have a list of the patients like problems on your phone so you won't miss something like, Oh, you just showed up in the ER for dissecting aneurysm two days ago. You would have all of that on your phone right away. You just see that. You put that down. It records for you. It documents for you. It sends out the bill. You get a check. That's that.

Dr. Spencer Baron:

Wow, that's big time. So, you know, here's, here's my thought. You know, I remember, what was it about 30 years ago, maybe 25 years ago. Do you remember Dragon? Dragon software was still around. Was the first voice to text? Well, one of the first that you could, you could dictate your notes, and now look, you know, way ahead at where we're at now, and you know, we're so advanced. But what are some of the challenges or difficulties that you're seeing that people are well, the doctors may be using the AI and it's not working. Mean, actually, because you're working with someone's health, where would that be a problem?

Dr. Ti Pence:

There's, there's a bunch of those. So first of all, it's a language model. It's, it's not a math model by definition. So so it really doesn't do well with numbers. It has to use other tools to generate numbers, or like timetables. It also is very as I think we discussed on the last one, it's very sycophantic. It will try to give you the answers that it thinks you want. So if you're not careful in how you're asking for something, it's going to give you the diagnosis it thinks you want, because it wants you to stay engaged. That's how it was trained, right? It was trained to be engaged. It wasn't trained to give you right answers, so you've got to be careful of that. Now saying that they've there's plenty of studies that say that these models are outperforming great doctors at most cases, some cases that the model completely fails at, but like that, they are outperforming a lot. So you can't just dismiss the whole thing and say, oh, there's no there's no value in it. There is, but you got to be mindful of like, is it trying to say what I want it to say, or am I getting the right information? Did I make sure that the prompt wasn't leading it to an answer that I wanted, and I'm trying to get the answer I need out of this? So scary parts, and most of it's from the patient side. Patients are putting all their information into these open models that like it's that that information is accessible, it's not protected, that's not that's not like you're putting it in into some secret nobody knows about box, you know, like that, that that's a huge computer system that is accessing that data. Who knows what, how that's going to be used in the future? Like, that's not a good idea, and it's not always the best bet, but often it does come up with diagnoses that you wouldn't have come up with otherwise. So we as doctors have to, like, entertain like, if a patient comes in with that information, we have to at least somehow entertain it and go, it might be, because these systems aren't always wrong, sometimes they're very right. And so we have to entertain that and go, Okay, well, if it is that, then we have to run it down.

Dr. Spencer Baron:

There was a patient that i Young. Well, not young. She was in her mid, mid 50s, and she was on multiple medications. You would never think she has these, the challenges that she does, physically, mentally. And I went ahead and took all her medications. Well, actually, there were three or four medications that I knew just felt that weren't good for her. And, man, I put them through something as simple as chat, GPT, and asked for, you know, what are the contraindications, plus this patient also drinks a glass of wine every once in a while. Ty it talked about that this woman could have, she could die, because she was, you know, at night, she would take something to sleep, and then that would be on the, on the, you know, piggybacked on a glass of wine. And so that was valuable. But would a would it a program that's used by the medical students be able to do that automatically, or is that something you have to ask for it? I mean, what? Because there I was telling my son that, and he told me that he spoke to a woman that has taken 26 medications. How is that? How is that okay? Is that something that AI could automatically generate a contraindication?

Dr. Ti Pence:

So right now, like, if you have a lot of us in chiropractic, don't have access to the kind of systems prior to AI like just electronic medical record systems. It's vastly different when you're working in a hospital setting where you have access to something like epic or Athena or Cerner, versus what we're usually using right we're all we have in our notes, is what we've generated from our office. But in these places, they already have all of the medications, as it all already gets pulled into this system, yeah. So they see all of the problem list on one side or somewhere on the on their program, they see the problem list. They see the medications. They can now ask the chat thing like, hey, pull up the medications and run contraindications like, you can ask it to do whatever you want that exists now. You can ask it to like, hey, summarize the past, whatever, however many notes, or give me a patient summary based from this date to this date. You can do all that right now, and it'll auto just go into your notes.

Unknown:

So like the TY, a question. You know, we used to be nervous about when patients were going Web. MD, and pick up all their symptoms, right? And then they come in freaked out. I'm gonna do this and this now they're going on chat. GBT, you know, what did the person listening, whether it's a healthcare provider or the layperson. What's the person listening need to worry about or be cautious about when they're going online?

Dr. Ti Pence:

Yeah, oh man. Okay. So this is one of this is one of those, like, long conversation kind of pieces. That's okay. I feel like I've just got started on this crazy, crazy rabbit hole, but it just, I looked around and there wasn't anybody else diving down this rabbit hole that I could see. So I'm like, someone needs to dive down this rabbit hole, right? So I'm trying, I'm trying to see how far I get down this thing. And it's funny, as a chiropractor, being in this space, like there are not chiropractors in this space, so I definitely it's wild, but everybody's super. They're like, Oh, cool. You're into this cool. We don't care that you're a chiropractor. Like, I get to talk to people that I would not get to otherwise. But back to the question, all right. First off, as a doctor, you got to know how AI fails. You have, you have, you have to have enough knowledge of AI to understand when it's useful and when it fails right now, which will not be the case two years from now. So we have some sort of AI literacy is very, very important to our profession, and unfortunately, right now, our profession is just trying to get students not to cheat with it, like that's really all. We've got no one. We're not teaching new students how to use it. There's no continuing ed courses that are like pseudo mandatory where you're actually learning. The best case is like people who have businesses in AI trying to kind of teach you about their business. That's not you need some literacy, because you have to educate people on where AI fails, and part of that is that it's going to try to give you the answer you want depending on what you ask. You can prompt different answers just depending on how you ask the same question. It's not like a robot like that. It's a it's a language predictor, like an auto correct in a lot of ways. So it's predicting, so it's trying to give you the answer it thinks you want. And so that can change each time you ask. So you got to be mindful of that. You got to also be mindful that it's not human. We tend to assign human qualities to this thing, like empathy, and it doesn't have any of those. It's not human. So like, the problem is we'll treat it like a human. It'll give us answers we want. We'll start depending on that thing for the validation we want. It will start giving more and more validated answers, meaning it's backing up our beliefs. And so if you're like, a worried well, or if it's a mental health issue, like we as doctors, should be asking, like, how much time you spending on this? Because just the time spent could produce worse outcomes. Just, just because we're we're, we're believing our own nonsense and talking to a thing that's reinforcing our nonsense without anybody checking us that says we're full of it, like we don't have, we don't have AI that just goes, you're an idiot. You're wrong. Go to your doctor. We don't we don't have that because nobody would use that, because nobody wants to be called an idiot. But in some cases, that's exactly what your doctor needs to tell you, like you're completely off your rocker. What you're worried about is not it like, let's be empathetic. I understand you're worried about this. Here's why you don't have to worry about it. Interesting.

Dr. Spencer Baron:

What is, what would you say AI is most useful for?

Dr. Ti Pence:

I mean, the medical war in the medical world, all the stuff that we don't want, that we didn't want to do, that we don't like to do, that is causing us all the burnout. I think that's where AI is going to be super useful. I think it's also going to be very useful in connecting us, if we want to be connected, if we want to connect with all the rest of the care team for a patient, probably pretty soon, we'll be able to do that in a very fluid way, where, like, Okay, after those notes get generated, they can get auto sent to their primary care providers, right? And the primary care provider, and you could have conversations now, and and there are companies where like, Oh, you want to, you're in the VA and you want to book an appointment with a chiropractor. Oh, well, the scheduler just pulls up your schedule and it auto populates into your schedule. You don't nobody on your side answered the phone. You just have a schedule full from people booking people are going to find you because they're going. Go, we're right. We're at agentic AI right now, which means there are you can ask to go do things for you. You wouldn't ask it to like, Oh, hey, when is this When? When does this doctor have openings? You could say, hey, schedule with this doctor and put it on my calendar. And so if you have back pain, you go, I have back pain, and it goes, All right, well, here's three chiropractors in the area. Which one would you like to book with? Okay, we'll book with that one on a Thursday. And here's this calendar and blah, blah, blah, done like you're just if you're in the loop. Man, it makes it a lot easier to do all of that.

Dr. Spencer Baron:

Everybody is transitioning from Google and using chat, let's say, but now Google's firing back with their own AI. You know, is there obviously a difference for people that when using purely an AI system versus what Google's offering,

Dr. Ti Pence:

I don't think any. I think it's very rare to see somebody actually using Google as any search engine, as it used to be, prior to the prior to like generative AI, because I don't before that people wouldn't read past the first page of a Google search. Now people won't read past the AI summary. It's the same thing you're just you're getting. I don't think people, most folks, understand just how much they're already using AI in their life. If they did, it might change their tune on things. But, I mean, there's, there's pros and cons of this. But which one of those chatbots do you use? I think it's, I think it's dealer's choice at this point. They say that Claude is a little less sycophantic, but really it's just which one do you? Which one gives you the information that you want. It'll kind of learn what you want. As far as data, you'll learn how to work with it. Doctors that work with these things are better at working with these things. And I mean, as you said, you entered in medication, like a list of medications. When they train doctors on this. They had a study where they put doctors like, you get AI or you get Google, and what they were doing with AI was just typing in the same thing they would type into Google, and the performance wasn't much different. But then when they taught the doctors, you don't need to just type it in. You could just take the whole patient file and dump it in and ask it the questions. Once that happened, then they're like, oh, that the performance went way up from that like, Oh, yeah. I didn't even think that I could just dump whole patient files in there and then extrapolate data off of there. So I don't think many people are even using whatever the rate right now of Google being used rather than AI, it's probably going to just dwindle, although frighteningly, it did seem like there was some studies that using Google before you used AI was better. And also, there's some there's a study that I just read yesterday or day before, where they where, if you used AI first, and then you try to come to your diagnosis conclusion, or you came up with a diagnosis conclusion and used AI as a second opinion. Using it as a second opinion was better so that there's that, like, Okay, think first, then, then use aI don't like Don't limit your thinking, and then try to agree with it,

Dr. Spencer Baron:

just to as a chiropractor, just to salvage the fact that profession and how important it still is, people still want to have their hands, have a doctor's hands on their area of complaint, versus just, you know, getting a prescription or being told what their problem is based on AI. Is it safe to say that chiropractors will always be have a secure have a security in their profession, because we do the one thing that AI can't do is touch.

Dr. Ti Pence:

I would so I'm gonna, I'm gonna answer this in a two part answer. One of those is, right now we have a skill that cannot be duplicated by a robot or AI that hands on. Skill is something that is decades away as far as a robot. But our schools aren't turning out awesome hands on doctors right now. They're not like, I have not talked to a doctor that hires associates. It's like, wow. They're really doing better than five years ago at turning out new doctors? Nope. Nobody's telling me that. Everybody's telling me the opposite, that it's harder to find recent graduates, is know what they're doing with their hands, if you've got that skill, which is very important for our profession. Going down very little, very few people are into AI who want to do rehab. If you want to do rehab, you you better learn how to augment it with AI. And if we're not doing that, then, then we're also behind the curve on the rehab piece. And if you're not careful, and you're not teaching people how to think with AI. These new students are actually using AI to make their job easier of getting through school, so they're cognitively slower than those that had to learn earlier. Eventually, you're going to have robots catch up like the curves are going the opposite way, right, like eventually the robots, you're going to have robots that don't even have to look like human, but can mimic a movement. And if people are willing to go to the joint versus go to someone who's been practicing or skilled for a long time, there's going to be a part of the population that's okay with having a robot do it, yeah.

Dr. Spencer Baron:

Well actually, my rebuttal to that, or just even not a rebuttal, but in concert with what you're saying, is that those who would allow a robot to do any kind of manipulation are usually the ones that are on a McDonald's a McDonald's diet, so to speak. You know, they're at that level. They just want to bump up their they'll have health care. Whereas, you know, take a take an athlete, you know that really is finds their health and wellness important, they're going to go to somebody who's immersed in that skill of hands on and treatment. And that's one, one part that's important, the other, the other is that you know, you're absolutely right about some of the students that are graduating right now. A lot of them lived through the covid era, where they were learning how to put hands on through a zoom call on a stuffed animal. So that is fascinating. And I, you know, I've been seeing that, although you know these, these, these students need to be rigorously trained and put their hands on somebody.

Unknown:

Well, don't you remember when, even maybe 510, years ago, right around this, you know, the holiday season, you see personal massagers, and they were like, You laid back, and it was just two rollers, and it would go on your neck, and you just lay down, they roll it. Now you're seeing this thing that looks like a hand that's going like this, that is working their neck. I don't think we're too far away from the people that will choose a robot working on their hands. I think it's the same people that buy a massage chair instead of going to massage therapist.

Dr. Ti Pence:

Well, let me, let me, let me break the glass even further. There's not a line in between that there are. They have something exists right now that is another thumb on the end of your finger that's all robotic that you that will attach to your brain, that you can teach how to move. Right now that exists human. Well, you don't, you don't sew it on. It's like a it's like a glove. You like, put the glove on, and it has a thumb that then connects to your brain, and it's like, you're thinking with the the thumb on the other side, and it moves, and you can get it to pick up stuff, or, like, do soft tissue work. I mean, wow, it's gonna get gray, right? Like, so where, how much augmentation Are you cool with?

Unknown:

Yeah, wow, I haven't heard that one. What did I do with an extra thumb?

Dr. Ti Pence:

I mean, I mean, anybody that's doing a R T type stuff that just gets tired of their thumb all day, and you can, just like, use your other thumb to do that. It wouldn't take that long. I mean, if you've used any type of Graston tool, I'm sure it's probably not that hard or factor tool. If you're using that, and you feel like you can get a feel through that metal tool, I feel like it's probably about the same amount of time to get used to another thumb and how to think through how to move that thumb.

Unknown:

Wow, wow. I even think about stuff like that.

Dr. Ti Pence:

So, so if you can do that like, it's not a line, right? It's super Gray, and then can that thumb learn from what we're doing?

Unknown:

Probably, I mean, don't you think that people are getting, you know, with AI access? I mean, it's the same thing when, I remember when they brought the Mobile. Phone out, and they're like, oh, it's gonna simplify our lives. No, it's just been more complicated. Had more access information before. Yeah, they look at the pretend Cyclopedia, the Britannica to look up something now it's on your phone, right? And so is it gonna get to the point where doctors, or lay people are actually smarter than the doctors now, but they lack that one thing. They lack that intuition, that touch, that that experience, where do you see the future, even in the next couple of years, where you see that? I had never heard about the thumb, but I mean, wait, where you see some of some of this dot, where you see the grain, and where you see it, where you think it, it, it needs to go. Okay?

Dr. Ti Pence:

So the easiest way to start this talk, because this is, this is where everybody has these talks like, Oh, my God, are we going to become irrelevant, which is what everybody's worried about with every new, crazy, modern invention for like decades, if not 100 years, right, right? And it's never turned into less of a thing. There's those that learn to use the thing, and then they just augment what they're doing, and they just keep doing what they're doing. And then there's those that don't, and those that don't, it tends to be harder for them. So I don't think that there's any worry that we as a professional, go somehow, go somewhere. There's already too high a demand for us, and we can't provide enough care to the people that are out there right now. We're not turning out enough students to care for those. The demand is just going to get bigger people. People are not becoming more healthy. People are becoming less healthy. This is, this isn't helping them. This is going to be if we we have no advocates within chiropractic related to AI for us or for our patients, and we really should be doing both, because if, if part of what we've watched when exercise became optional. Now nobody exercises, and it's so hard to get people to do the basic amount of exercise. If thinking becomes optional, very few people will think that that combination is going to be a real problem going forward. And how, how do we advocate for helping our students and our patients navigate that without becoming just wrecked in the process.

Dr. Spencer Baron:

Thinking is optional, interesting.

Dr. Ti Pence:

You'd never have to if think of think of your day, and then think of how many times you could just ask chat GPT to answer your thought. And college students are doing that right now.

Unknown:

Yes, it's funny. You mentioned it because I heard Elon Musk one time when he was talking about the neural link, and I think it was Joe Rogan. And Joe was like, well, that's scary, that we have this device that can talk to somebody else without opening your mouth and going on. And Elon just looked at me, goes, well, it's already there. He goes. He goes, you see that thing called a phone when you text somebody, aren't you talking to them without opening your mouth? And Joe just went, I never thought of it like that. You're carrying that phone just because it's not under your skin or implanted. Nobody goes everywhere without their phone. He goes, it's already connected, and you're already talking to people without opening your mouth. Yeah. Opening your mouth, and you're already searching for answers without ever thinking or open your mouth by Google searches while you're holding your phone. You know the only difference is now you don't have to carry it's just implanted.

Dr. Ti Pence:

Yeah, you want to get you want to get worried. Ask, ask your chat, GBT, or whatever one you're using regularly, what it knows about you so far?

Dr. Spencer Baron:

Oh, wow, I haven't done that yet.

Unknown:

Yes, that's crazy that, yeah. And so, all right, so you're immersed in a what things to this day, that when you look at AI in any especially in our in the medical field, that just blows you away, that just shocks you, and goes, Wow, that is so cool. And on the flip side of that question, what does AI miss the mark that that people don't realize? Oh, man,

Dr. Ti Pence:

you're talking about a realm full of extremely smart individuals that I'm just trying to keep up with right now. So missing the mark is, like, it's just exploding out as far as what it can do, and it's learning very fast. And what, what happened was, what's happening right now is, every time it expands out with a new version of something, there's a little bit of a lull where then we get to look back and see, oh, well, we could also use this for that. You get to play with the same tool for a little while, rather than getting the next tool super helpful. What I see is the most exciting, cool parts, man, we could where we could be as a profession. Could be amazing if we decided. To embrace this, you could and think about the learning curve on this. This is expensive now, but won't be expensive soon. You could take video of somebody and have aI analyze it, just like it's analyzing any of your other imaging techniques give you angle measurements, force generated by the person do all that stuff that's not too far out. And then when you do that, you go, Oh, you want to work on your hip. Click. Then it just takes that and generates a rehab report on, like, here's my rehab report for that. Oh, I don't like this one. Let's up that, because I know you're stronger in this motion, and let's down that because your balance, your balance is kind of wonky, we need to like up the balance for that hip, and then just send that that goes to the patient, uploads into whatever the patient wants to do for their exercise files or their at home items. It's crazy. What we could be doing. We could be turning that on ourselves and learning how to become way better at hands on adjusting just by videoing US versus videoing other people. And if we want to look like somebody, we can just have that other person and work until we look just like that other person.

Unknown:

I mean, you're so you're saying that I could, I could video Spencer walking and running, and then download it into an app, and he comes in says, I have left hip pain. And then AI will analysis everything on motion, 3d everything that we're usually doing right now with our eyes. And they can go, Well, this is why he has hip pain. He's got this and this and this, and then turn around design a rehab program for him,

Dr. Ti Pence:

not yet, but soon. Wow, because to back up a little bit generative AI, to simplify it for here, it knows the whole internet. Agentic AI can take action on the environment. So it can, if you give it access to your calendar, it can, like book appointments on your calendar, for example, if you give it access to something, it can, it can, it can act in that environment. So it can take action. So you can give it a goal, and it'll take action to achieve it. A simple one would like, hey, I want a pre auth approved. It could just annoy an insurance company until you get the pre auth approved, which again, turned against you, if we don't embody this stuff, insurance companies are using that, and they're just telling you no until you get through the filter. How are you as a human going to do that without having some sort of AI helping you do that? Wow. So each of those things that you described that I just that, I told you and you asked about, they exist now, if you know how to prompt and start looking for things, but you got to put it all together, it will start getting put together. They have they have ones where they can watch you move and turn you into a stick figure and start figuring out math based upon that, the speed at which that will become way more problem. The only reason that everybody isn't doing this already is because everybody's focused on more high value dollar things. It'll come here soon.

Unknown:

What are some of the dangers that people have? People have to pay attention to because, I mean, if somebody's going on AI a lay person and emptying their symptoms, and it tells them what will it tell them what to do? And could it lead to death? Could it make some liability? Can increase liabilities. Can Can chiropractors, doctors enter wrong thing and be and and can it lead to death? What are some of the downsides that people need to be really aware of,

Dr. Ti Pence:

okay, so, I mean, what? We have some of this now where people will be wearing wearables and and get a number, an HRV or a or a pulse, or something that's or that says that they're not sleep. Who knows? They get some wearable data, and they diagnose them, then they input that into chat. GPT freak out come into you, they actually have no symptoms. It's just a synthetic error because the wearable is not that reliable. But that's not what they're telling you, because they've just read chat GPT. So they're giving you what chat GPT just told them. So it sounds like a problem that you've got to, like, track down to, oh, you just actually looked at your watch and it told you something crazy. So there's all of that. Then there's the liability part. First off, there's no liability right now in AI, like, like, you'll have to worry soon, but it is a race. To who gets to the top of AI first between Well, the world. So we're not slowing that down at all. The federal government has made it clear you're not, you're not going to get punished for this. But things are changing, and who knows where that line is going to be. That line might be that if you don't use AI for certain things, like we're using AI from mammography right now, if you don't use it as part of your practice, you could become liable because you you had access to knowledge that you didn't use. So it's where that sits. I don't know it's not, it's not now, but it will be at some point, could be. And there's, there's to say, AI sounds like there's a singular tool, or like a terminator type consciousness. That's not really what there's. There's millions of tools that are being used based upon different platforms, and they're doing different things in hospitals. So there's just, like, lots of things going on with different risks in that that's going to have to be filtered and figured out how it's being used, but big organizations are trying to figure that out now. Yeah, interesting. Yeah, you guys look like like I looked the first day when I started digging into this stuff was just like, Oh, my head hurts. It really starts to make you think differently when you start digging into this.

Dr. Spencer Baron:

I wonder if a doctor, if, if the day, well, first of all, I wonder what this conversation will look like a year from now, like you were telling me about. I can imagine how advanced it will become in a year from now that we'll look back on this. And then this conversation was like age old compared to what it would be like 365 days from now. So that's how fast

Dr. Ti Pence:

much of what I'm saying Could be, could be wrong or radically different a week from now, two years is considered a long time in the extent of this, and the people that are pushing the edge of this right now, all of these models Train off of language. They're large language models. So they're, they learning based upon words. The next step is to have these systems learn off the world. Oh, fine, yeah, what that looks like and how that's gonna what, what that's gonna do for us. Man, that that's just guesswork right now. But, but, but to think of the the infinite of that, and if people have that in their pocket, just whenever they want it, it's crazy to think of like what could be possible.

Dr. Spencer Baron:

Well, you know, it's interesting. You say that because you think about other countries that are way ahead of us in their health, their health and wellness, or their longevity and how America spends a ridiculous amount of money on health care, but yet, we're the unhealthiest, one of the unhealthiest. If we could tap into those other systems in other countries, I wonder what, how that alone would change health care if we can. I mean, politically, we're already starting to model what other countries are are doing as far as medication and surgeries and things like that. I think it's fascinating, especially with this Advent, the advent of the functional medicine doctor, who's now, you know, I don't know about I'm sure Terry heard the same thing when we graduated from chiropractic college. How? You know, Doc, medical doctors were saying, hey, just you don't need vitamins. Just have a good diet and you'll be fine. And they didn't believe in vitamins, period. Now, it's such a cash model for them, because insurance sucks there. There's medical doctors becoming functional medicine practitioners, and that's in our country alone. So when I, when I use chat and I upload a lab, you know, whatever it might be, could be a stool sample or saliva test for hormone, and I'll say, as a functional medicine doctor, what would you suggest, as far as nutritionally, you know, maybe even with peptides and so on that to take care of this. You know, what I'm finding on this lab, and it's giving me, you know, cool stuff, especially if you say that the patient is interested in getting off medication, because that model needs to hear that and give you recommendations in getting off a medication, which could be bring a lot. Ability to that doctor as well, if they're garnering their information to give to that patient you know, and a man you know in his lawsuit, happy world, you know, the medical doctor might go, oh, I asked chat GPT, and it told me to do that. You hear, yeah. You imagine, really tough. Yeah, a liability,

Dr. Ti Pence:

of course, of course, yes, the liability is going to fall on the person that has the license, because there's it's a long way off before we could claim that code was written wrong, or could have been written some other way, or that a generated answer was a lie and should have been something better, or like, there's so many of this, but all of those models, all of these models, could they technically help our healthcare system? Yes, right now we're being we're using them to fix the problems with the healthcare system as it exists right now. Whatever we train the models on is, I mean, it's just like teaching a kid. You know, if you teach a kid a certain way of doing something, they're going to reference that as the way they do it as an adult. The same thing with these models. If you, if you train it on, on some of these healthcare systems that are doing better, then it might have different responses. As far as, then what it recommends on the other end of things, as far as prompts go, yeah, there's, there's a lot like, research on prompts is crazy, and I am definitely not a master of this. There are some good webs I could, I can send you some YouTubes as far as, like, good ones that will kind of teach you how to write prompts better, which is just whatever you type into the box. But clarity is the key piece of that. Like you want to be clear that the whole, like, talk to me as a functional medicine doctor, not really as much anymore. You would want to know like, Hey, what are treatment solutions for this patient with these given they don't like meds? Wow, I want it in this format. Here's the other details. You want something like that, like love that.

Unknown:

So, hey, hey, Ty. So listening to this what stops? What stops a patient listening to this show from when they get their their blood work, their stool samples, and they get their report in the mail, what stops them from scanning in and asking the exact same question that you just did and bypassing the doctor?

Dr. Ti Pence:

Oh, nothing that's happening right now. Yeah, and that's not only that's only hap if you have cancer. I just watched a demo on a on a startup where you don't even write your you don't have to upload your files. You just tell it where you're going, and it'll go there and pull the files in. It'll tell in patient language, what the treatment options are. It'll it'll give it everything the person needs, as far as, like, what are my next steps with cancer, right there.

Unknown:

So All right, so piggyback on that question, if, if they're bypassing a doctor and students now are using more AI you said the word earlier, the cheat and do all that. What? What do healthcare providers need to really focus on to still be valuable?

Dr. Ti Pence:

So the easiest way I can describe this is it frames back to one of the schools that someone's got going, where they're using AI and to teach young kids. And it's, I think it's called the Alpha School, where the teacher isn't in front of the in front of the students, teaching the information. The student is the the teacher is the guide for the students. So there's one teacher in a room that's helping the students guide through all of this data to to grow as they want to grow, and what that does is allows them to excel rapidly in their learning. Because you're you're not the you're not the bottleneck of knowledge, right? The person can input it as much as they want, but you still need to be like, okay, basic human ideas. Are you going to get up because a robot tell, like, something on your phone tells you to get up and go to the gym? No, you're not. Like, that's not how people go. That's going to be the same thing as like, oh, it auto enrolled you in a gym membership and told you to go to the gym membership. You just turn off the auto alert three or four weeks in, and that's that like you feel no responsibility to that. But if you have a trainer at the gym, you're going to show up. If you have someone in health that's that's your advocate and helping you understand your health and where you want to go, that's super valuable. Still, you just have to, you have to understand that. Your role, and you're helping this person in that way. Wow, good stuff. You're not the keeper of knowledge anymore. You're not like, I am the keeper of knowledge. I am the only one that can speak in Latin. I am the only one who knows medical talk. Like, you're not that guy anymore, like any, anybody can come in with that information. And if you don't, if you don't know more than what just a regular search on chat GPT gives you, or where the fallacies are in that, then the value in you goes down. Yeah, interesting.

Dr. Spencer Baron:

Okay, let's, let's look into the future. I Yeah, because, you know, I'm thinking to myself that the future could be advanced 100 fold by by next year, by March. Yeah, it's crazy. What What are worst case scenarios? And what are, you know, best case scenarios for, maybe what patients could be looking forward to or being afraid of with AI. Or maybe you can make some suggestions of what they should be, you know, the the average, I think, I mean, we do have doctors listening to our program, as well as a lot of lay patients or lay people. You know, what would be suggestions for you to that they should be careful of versus what they could look forward to,

Dr. Ti Pence:

first off, doctor or patient, be careful of just putting your all of your information into something like chat, GPT, be careful. That's interesting because you are especially there are folks right now that are having full on therapy discussions on their like they're there, that's their therapist. This is not this not uncommon. You got to be careful of that, because all of that information is going somewhere that's not private. It is not protected by any type of HIPAA or patient safety anything. This is going to a private company that owns this information. So you have to be careful what it can learn off you and what they where it decides to go from here we don't know. So having your information is could be problematic. If you're a doctor, do not enter patient health information into chat. GPT. Don't. You can unidentify it and enter, yeah, yeah, that's what I was gonna Don't, don't, don't add it with like, don't see an uploaded PDF that has the patient's name at the top right. Don't do that. Good, good, good. Be careful. Also, of biases within a lot of these models. The if you're a doctor, if they know your age and your ethnicity, it might make a different recommendation based upon another like it. It has the biases that we put into it, and so we be careful, be careful of depending exactly on what it suggests as a diagnosis, follow up, make sure that makes sense, often it won't. So when you're using this, use it as a way to have a conversation with your doctor, if you're if you're a patient, totally if your doctor isn't giving you the kind of information that you need, and you're not getting to a place where you want in your health, and you think that there's more information that needs to be discussed in a patient visit. Do it? Walk in and have that conversation? Don't compare the answers from the doctor to AI. That's not a good idea. That's going to get nowhere, and that's probably just going to upset the doctor, and you don't even know if what you're comparing is some sort of gold standard as far as the diagnosis that the AI came up with. So don't do that. But then on the other side, doctors, when the patients come in, validate that, like, Okay, this patient is trying to get to something they need. They've used this for a reason. Figure out the reason, then figure out, like, what part of this plays into their health goals.

Dr. Spencer Baron:

Let me ask you just a quick question regarding that just dawned on me about I've heard that patients can upload their they get their they can get their X rays or MRIs read through AI now is, are they uploading the radiologists impression already, or they're actually uploading? I can't. I mean, I don't know if they can upload films, but because MRI is digital, and now X rays can be digital, is there? Are they? Is that something that's being done?

Dr. Ti Pence:

I don't know if it is being done. Is it theoretically possible? Especially with like X rays. Sure, what kind of findings are you getting from that? I don't know. I don't know that I would trust that, but your the findings often, what they're doing is they'll upload the findings and ask it to speak in, give it a summary in like so that the patient can understand what the what the report means. Yeah, and that again, could be fraud. Could be problematic. It could be so often, often these models, the part that they miss in diagnosis is omission of something like they don't tell the person they need to go to the hospital for this, or they don't need to follow up for this when they actually do. So it's not that they got the treatment wrong. They missed part of it, and that piece is important. And so you got to be mindful of that kind of thing of like, okay, what got missed is it just trying to tell me what I want to hear.

Dr. Spencer Baron:

Well, interesting, because even now you know, a radiologist will read detail, and the patient sees the radiologist impression, and they think that they're, you know, they're stricken for life with this herniated disc, but they didn't realize that, because it wasn't run with contrast, that that's an old injury. It's not what's causing you the pain now, but they, you know. So you're getting radiologists that read into things which I'm blaming the radiologist, I'm blaming the patient for seeing a disc bulge and that, you know, oh my gosh, you know, I can't, you know, I can't squat anymore. I can't pick up my kid anymore, or something like that, because they're making their own interpretations based on something that we see all the time well.

Dr. Ti Pence:

And before you would do that, and you would just Google until you found something that validated your concerns, right? And now, instead of doing that, you put it in a chat, which something that is not human, but you think is human is agreeing with you that you're right on this, and it could be something bad. And then you start feeding more into it, and they start and you can get spiral. This is what I mean, like the worry, if you're concerned about your health and you're spending time on this man, that good old 10 Minute Rule is probably still a good 10 Minute Rule, don't, don't, if it's just agreeing with everything you're saying and you find yourself spinning out on this thing. Stop. Put it down and stop messing with it like it could be that it's just got you spun out. Go talk to your doctor.

Dr. Spencer Baron:

Yeah. Oh yeah.

Dr. Ti Pence:

Now, now they have, they have where there's not just one machine, you know, it's not just one AI. They can have teams of AI working to an answer. So, like, one of them is trying to come up with a diagnosis. One of them is red teaming that diagnosis. One of them is trying to make sure that it makes sense to the doctor. Like, and they all kind of argue until they come up with a right and come up with an answer, and then they so it's you're not the judge is checking the answer, yeah. And those things can exist all over the place, you know, like you could. Could it be that it listens into conversations with patients and says, Ah, well, you did this, right? But, like, you could do this better in your motivational interviewing or something like that. Like, you know, yeah. Well, that

Unknown:

is interesting. Don't you remember the times when, even back in Nixon, when everybody would be whispering because, you know, oh, we're being eavesdropped. There's got to be a bug somewhere. Now people actually go up and Alexa tell me how to do this, you know, it's like, totally switched to, we were so cautious. Now we just say everything. And then we go, you know, I was asking. I was just thinking talking to my wife about a, we need a new vacuum. And then I go on my line, and first thing, it pops up, Dyson's on sale, you know. So, you know, yeah,

Dr. Ti Pence:

wait until it's in your in wait until it's already in your purchase box, right? Or it's bought for you.

Unknown:

We need a new Dyson also shows up at the front door and your credit card is already charged.

Dr. Ti Pence:

Yeah? Now, and the flip side of this something that I think as doctors, we do need to be mindful of. We're worried about the bad things. If something happens, some negative thing that happens to our patient, so we're cautious. But if you're if you have criminal intent, you don't care. You just needed to be right, sometimes interesting. So so being smart about your personal security, and the security and all of your systems around your your your practice probably way more important now, because phi is still phi, and if and if some someone get breaks into your system, because now they've got an AI agent working for them. That's just trying to find where you you use the same login code for too many things, and it found it enabled the log into too many things. Like, now you're in trouble, right? Because now all your patient data now just got it exposed out that that's on you didn't even think of that. You there are consequences we have to be mindful of. That's why, I mean, like the EHR systems that we use, we have to be mindful that they're being good about all that and that they're actually up to date on real security systems are, are the rest of our systems in our office also like that? Do we? Do we personally make sure that we do not dumb things, right? As our staff not like opening links on on our email and things, because all of those simple things are more of a problem now, yeah? Well, yeah, I'll tell you. We know that, yeah, when you said future, I was like, but this could be

Dr. Spencer Baron:

just distant, not such a distant future thing. We are two years. Yeah, yeah, Ty, we're gonna enter our rapid fire questions this. You know how that goes. You've been through it before. Don't sweat this out, but answer the best way you can. And if your question for number ready for question one? Good, ready? Here we go. If AI could instantly replace one thing, that's one thing you do as a doctor tomorrow, let's say, what would what would you never let it take no matter how advanced it gets.

Dr. Ti Pence:

Oh, there's a there's a cheesy answer that, and that's just my ability to interact with a patient in person. But I feel like that's just an easy answer. The other part that I wouldn't take my ability to my ability to agree or disagree on the diagnosis and the goal for the patient. I think that would be where I would draw that line. I would let it come up with summaries, but, but I would, I would still need to make sure that I am the human in the loop there that's going okay, that makes sense, or that doesn't make sense. Good.

Dr. Spencer Baron:

Well, that's you, and that's how I believe. But I'm not sure there are doctors out there that would do that, but that's a good, good, good thought. If you had a theme song, this is question number two, if you had a theme song that the one that plays every time you walk into a clinic, into the clinic, what would it be? Gosh.

Dr. Ti Pence:

Okay, so my first one, last time was, was ramble on by Led Zeppelin. That's my number like, that's always the next one is the next one right under that is midnight rider by all my brothers. Those are the two that will roll in my head most of the time.

Dr. Spencer Baron:

Listen to that when I'm done here today. I love that one question number three, if you could, if you could invent a totally useless superpower just for fun. What would it be, and how would you use it in a clinical setting?

Dr. Ti Pence:

Useless superpower that could be used in a clinical setting that actually is a good one, because I feel like between Marvel and everybody else, they've invented every useless superpower I'm trying to think of one that isn't one of those.

Dr. Spencer Baron:

You could use one of those,

Dr. Ti Pence:

if you want. Honestly, the thing that I mean the most useful to me would be some level of being able to have that communication between the phone and myself, like, just in my head. Like, could, could it just be, and it's almost there now, like, the superpower could be augmented just by putting the right headphones in right now. But could it be where it just I get told, like, as I'm walking into this page, this page, like, I just get it in my head, like, and then I can just walk and be like, All right, here we are. Or I could try and travel, like, from right, from last visit to this visit. So the last time I saw him, like, I could just, like, immediately be back to so that. So it's just fluid. So it's not like I got to recall it. It's like I was just in there.

Dr. Spencer Baron:

That is a cool that would be a cool superpower the days of having travel cards and writing down,

Unknown:

oh, God, door God,

Dr. Spencer Baron:

all right. Question number four, what's one piece of advice you'd give your younger self with the knowledge you have now? I

Dr. Ti Pence:

what I did well early on in practice was I put a lot of time into making sure that I was good with my hands. I should have listened to that piece of me that also said I got I got this technology, seems like. Taking off. I should also be a part of that. In my head, I was like, it's just that they're two different things. I didn't understand that they could be the same thing. So, like, learn as much as you can about technology. It is anyone that tells you that they're an evidence based chiropractor that doesn't understand AI is not an evidence based chiropractor. It's just that the reality of the thing you are not, you won't be two years from now, or five years from now, like you are not keeping up with the evidence. Yeah, so true, because the evidence suggests that you should be using this stuff, and if you're not using it, then you're not following the evidence, so therefore you are not an evidence based doctor.

Dr. Spencer Baron:

That is great, because I hear that that's become a buzzword among all professions evidence based, and yet they're still doing surgeries from back in the 60s, but that's a whole nother interview. Last question. Dr, Ty pence, when your career is over and that white coat just gets finally hung up, what do you most hope your patients and students say you stood for?

Dr. Ti Pence:

Man, I'm just I've always just been trying to push this profession forward in the little bits that I could that that's really all it is. Did my patients feel like I served them in a way that pushed on past this, this generation like somehow it helped future generations, from patients to other providers like that? That's all I'm hoping, is that somehow me being a part of this helps future generations have a better profession and a better life than what we had, and we have it because previous generations did the same thing. That's all I want.

Dr. Spencer Baron:

Beautiful. Beautiful. I appreciate. It's very admirable that answer, and I feel the same way and strongly about this profession. I think it's really got a great future ahead, and I thank you for bringing us current on how we can augment what we already do with technology and AI and how we could really advance what we what we do with patient management. Thanks. Thank you very much,

Dr. Ti Pence:

of course, and that's that's exactly that last piece is the point that every chiropractor should be learning. We need to be augmenting with this. It's not going to replace us. It will replace those that don't keep up with it. Beautiful.

Unknown:

Thank you so much for your time. Doc, we so appreciate you. Thanks guys, it's fun to be back.

Dr. Spencer Baron:

Thank you for listening to today's episode of The Kraken backs podcast. We hope you enjoyed it. Make sure you follow us on Instagram at Kraken backs podcast. Catch new episodes every Monday, see you next time you.