The Radical Impact of AI on Health: Power, Emotion & Expectations (#53)

Nowhere is people taking control of their health more radical than with artificial intelligence.

Beyond the headlines of faster drug discovery and AI companions, Christophe and Mo Zouina explore how AI changes our expectations, emotions and the balance of power in healthcare.

In this episode:

• AI speeding up drug discovery and creating new medicines
• AI stepping in as coach, advisor and companion
• How AI reshapes patient expectations and emotions
• The messy in-between: when AI promises more than is possible, or names a disease with no cure
• What this new zone does to the balance of power in care

With Mo Zouina, human-experience specialist.

AI doesn’t just speed up medicine — it shifts power toward the person. Health, reshaped from the outside in.

Subscribe to Healthusiasm newsletter:
https://www.healthusiasm.com

-- -- --

Book Christophe Jauquet for keynotes:
www.christophejauquet.com

Transcript

Estefanía 'Nia' Escobar-Kölle: 0:00

I mean, how do you how do you define care exactly? Does it care in a way that it's gonna give you interventions and it's gonna try to support your mental health or your physical health or any aspect? Or does it care that it has to fulfill specific requirements that are have been designed as part of a care path and you have to stick to that? So it really defines how you define care. But I think because it's a system and it's completely unemotional, it doesn't care. You're completely right. But we can train it to care in specific ways.

Christophe Jauquet: 0:31

Welcome to the Healthusiasm Podcast, a panel discussion on innovations in health and self-care. My name is Christophe Jauquet. I'm a health business expert, author, and keynote speaker. And every two weeks, actually, I discuss with a panel of experts the positive changes that are shaping our health and happiness. And for this discussion, we have again a very full panel, all five of us. Calling in from Barcelona is our digital health connector, Aline Noizet.

Aline Noizet: 0:58

Hola.

Christophe Jauquet: 0:59

From Scotland, but living near Brighton, medical expert in digital innovation, Dr. Keith Grimes. From Colombia but living in Berlin, digital health expert for startups, Estefanía 'Nia' Escobar-Kölle. A little and last but not least, from Ghent, Belgium, human experience expert Mo Zouina.

Mo Zouina: 1:19

Hello, Kis.

Christophe Jauquet: 1:21

Hey, hey. so every two weeks we talk about the health innovations of the previous months, which we just finished for the month of May. And then we always look for a specific theme, something more like something which we can deep dive into. We had a big talk about whether we should do man's health, and I think in the later part of the year we will have a big focus on that. I think that's a very good thing. But for this session, we decided to talk about a subject nobody ever talked about, which is AI. So, more specifically, we want to talk about AI in healthcare and see how that could, should, would or should not maybe even change the way that we do healthcare. What can we expect? Where is this going? Why is this important or not important? And how should we go about this? Let's start with you, Mo. You always the more philosophical one. Maybe it's good to start with these type of touches before we start thinking about it.

Mo Zouina: 2:28

I feel a bit typecasted, but I'll gladly go into it.

Christophe Jauquet: 2:31

You did it to yourself. You did it to yourself.

Mo Zouina: 2:34

I'll gladly go into your invitation. AI, it's all over the place. It's about unlimited speed, scale, complexity, made simple. And then our friend Keith said it's about capability. I call it agency, the ability to act, right? And in healthcare, that kind of speed is kind of revolutionary, right? But here's the thing I want to say out loud the health system, the delivery of health is not keeping pace, right? So it's not fast, it's hard to scale, it's often complex, and resources are dwindling. Very important resources are dwindling. We have less people and less resources per person to deliver care. So on one hand, you have diagnosis sprinting ahead, and then treatment is jogging behind and might even be taking a break because sometimes there might be strikes where the system is kind of shutting down. And I think this is where it gets dangerous. Listen to that formula that unsettles me. A single click, fast answers, plus consumerism, plus slow care, plus limited resources, plus social media squared. Right? What does that add up to? I think it's a dangerous cocktail. When people experience technology that feels this fast and this intelligent, their attitudes and expectations might follow just as quickly. And yes, I'm hopeful. This is Halthusiasm. Yeah, the future feels hopeful of the decentralized drug production, 3D printed pills, pharmacy kind of producing their own medication, maybe even home bioprinters. And it doesn't stop there. We're also heading, we talked in the previous Halthusiasm, we talked about regeneral cell therapy, compound-free therapies, gene editing, living medicines, and even regenerative medicine. Bold and exciting. But where I think we will suffer is in the messy transition period between the fact that we can taste speed right now and we're not getting it from the healthcare system. So imagine you finally are able to diagnose something that wasn't diagnosed before, right? AI pinpoints the condition. You know, it's that exactly, precisely. And then you hear there's no treatment, or there's nothing we can do, or it's incurable, or worse, there is treatment, but you have to wait four months because the system is kind of slow. And I don't think that's hypothetical. I'm here in Holland, and as you probably know, the Dutch hospitals they have a quota for cancer treatment. They only can spend that much on cancer treatment, they can only treat that much, that many patients. I had a friend of mine which was working at JNJ. His therapy was way better than the existing leukemia therapies, and the hospital said, but who are we saying no to then if we have to pay more for your? We can treat less patients with your drug. So I think in the messy intermediate side, we'll might be exposed to a tsunami of frustration. And I think I predict the erosion of tolerance for the expressions not yet, not now, or even no. I think we'll be very we'll have a difficult time hearing those words because we get the impression of speed and that everything is possible. So, yes, innovation is coming fast, but expectation might be moving even faster, but delivery is not going that fast. And that sounds like TNT to me. That sounds very explosive. So I think you know, these soft skills for clinicians, for caregivers, and things like that to be able to develop the culture of maybe patience, of matching expectations. And I think if we if we don't address that, we might be moving into the age of unreasonable expectation and unseen impatience. That's I think a radical challenge we are facing with AI. Speed on one hand, inertia on the other hand, and then unreasonable expectations in between.

Christophe Jauquet: 6:27

Floor is open. Does anybody want to piggyback on that?

Dr Keith Grimes: 6:32

I mean, we I've spoken about this, but you know, we're talking about something that I spend such a lot of time thinking and talking about, and but just try and boil it down. I mean, there's the there's a whole bunch of points there. I think I think Mo's right, but it's a kind of variation of a problem that we've always had, but we've got it's kind of happening faster, and people's expectations are high, you know, what we want versus what we need, and who's making that decision, and do they fully appreciate it? Well, when we talk about AI and in healthcare, I previously talked about healthcare, I sort of see it as the two components, there's health and care. And the health part is the in some ways the kind of more technical transactional bits. And that's where AI seems to offer the most promise. If you take an ethical approach, you know, beneficence doing good, non-maleficence not doing bad, autonomy, the ability to act with agency and justice, that it's available to everyone. You can see very quickly how AI, you know, actually ticks a lot of boxes in terms of positivity, you know, the beneficence. There's we've heard about tremendous positives that AI can do in terms of diagnosis and the like. But then we worry about non-maleficence not make, you know, if it makes a mistake, what harm comes. The bit that I get excited about is that democratization of access to knowledge and capability, which speaks to autonomy and justice, that people getting this. So, you know, there's a lot of ticks in those boxes, a lot to work on there. But then also as Mo was speaking, I was thinking about what we want as humans from health and care. And we can't we want our health fast, but we want our care slow. And I don't mean waiting for it. I mean we want, you know, care delivered with time and patience and attention is what we want. And in the rush to deal with the health side, that kind of pressure to get things done faster, better, cheaper, more accessibly. My concern about AI is we can go two ways. We can either use AI to sort of whip us even faster and drive us further into the health side of things, and we can succeed on the surface of it, but at the cost of time and care. And so my message is always about, you know, we use, and this is a message other people have said too, we use the technology, AI, to give us back time to care. That's where I think we need to go. And then building on what Mo said, I think a lot of doctors, I'll speak about doctors here, is we've defined ourselves very much about the technical side of things, the health side of things, the capability side of things, at the cost of the care. And doctors and healthcare professionals feel it, patients feel it too. So we can make choices. We can choose to go one way and maybe feel worse, or we can choose to rediscover something that we've lost. And so, yeah, I think AI is forcing us to face it more now than ever.

Christophe Jauquet: 9:22

Yeah. I think to the point of that Mo was making, the sentence I always use is that if the speed on the outside is way faster than speed on the inside of the system, we go straight into an infarctus. And I think in between, I love the way that you put it, unreal expectations are definitely the reality there. Now, let's go back on what it could really mean. So let's just having that fact, which is I think a reality or soon-to-be reality. Let's try to motivate our listeners and maybe the ones that are working in the healthcare industry by providing them health reasons and positive things that AI can do, let's try to convince them to speed things up. I mean, you were talking about democratized democratization key before. You're also talking about giving back time to care or recover something that we've lost. But I think Nia, you had a thought, I believe, if I'm not mistaken, that kind of fits in that giving back time to care or at least recovering something that we've that we've lost in the system, maybe.

Estefanía 'Nia' Escobar-Kölle: 10:25

Absolutely. And it's very interesting because a couple of weeks ago, my LinkedIn was recently flooded with this really interesting infographic from the hardware report on gen AI use cases in 2025. And the top three cases were therapy and companionship, organizing our lives, finding purpose, and then the next one was enhanced learning made as the top on the list, right? So if you look closely, all of those use cases of how we are using AI or Gen AI rely on one unexpected ingredient, and that's emotional intelligence. And while it might sound a bit paradoxical of machines and empathy and emotions, we're growing, we're seeing a growing evidence that AI can actually model emotionally intelligent behavior. And there was a new study from the University in Geneva and the University of Bern, where they actually tested Gen AI models like ChatGPT, Gemini, and Claude on standard emotional intelligence assessments. So the models actually scored an average of 82%, while human participants averaged just 56%, which is even more impressive that ChatGPT 4 later created emotional intelligence tests from scratch, which were then validated as equally reliable as expert-made versions, a process that usually takes years to generate. So, what does this mean exactly for healthcare? So, in a helpful one in the health enthusiastic side, is that with thoughtful design and proper oversight, Gen AI actually can support, for example, when it comes to healthcare providers, to have these emotionally charged conversations, help clinicians handle difficult topics, coaching caregivers under pressure, or even provide a new layer of support for patients who are waiting to be treated or unable to see, for example, a therapist, right? I mean, I think we all have seen use cases of Gen AI in healthcare, but the fact that there is evidence and that we're looking into this from a scientific perspective means a lot. And once that AI is entering the spaces, once thought to be exclusively human, as we're of course facing a massive car gate care gap in our friends, this means that emotional intelligence isn't just really human anymore. And in healthcare, it might be exactly what we need, considering that, for example, one in eight people globally live with a mental health disorder. And in the US, for example, over 160 million people live in designated mental health professional shortage areas. So they really have no access. And 60% of individuals with mental health conditions receive no care at all. So, yeah, some people might consider this a threat, but as we can see, there's actually hope in terms of receiving care and getting the tools that we might need to treat mental health conditions, not just for patients, but also for healthcare providers.

Dr Keith Grimes: 13:11

Yeah. Just on the back of what Nia was saying, actually this week, I was reading and also actually speaking to people working in the emergency care services, the ambulance service, and they've been quite excited about the use of AI for ambient scribes, you know, transcribing the conversation and summarizing it. But within the package that they're looking at, it also has a kind of sentiment and semantic analysis, a sentiment analysis function as well, which doesn't require generative AI. It's been around for a little while. But the reason that they're interested in that was more for the mental health and wellness and support of the call center staff, because if they're taking a lot of calls, they could be quite challenging, that would be able to flag to those people that are leading the teams and so on to be able to schedule time breaks, give them extra support, move into different type categories of calls as well. So even though that's kind of functional, as in we're trying to get the most out of our team and make sure that the service is delivered, it's underpinned by a need to care for the staff as well. So I think people are aware of that and it's coming in. But then I think about how people have been reporting on the ambient scribe research and particularly impact and return on investment. And the overwhelming conversation is like, well, it saves time, which is great. And then people say, well, it doesn't save that much time. It saves like 30 seconds a minute or something like that. And then it gets translated into, and that's not much of a return on investment because you can't see that many more patients. But what's always second or third is that the clinician feels better, the patient feels better, they feel less burnt out, they can sleep better at night. And it's it I find it sad that this enormously powerful signal about how this technology is helping people just feel better doesn't have enough value to justify the spend on that technology because it's not returning something else, which is the ability to just see more folk. And yeah, we really have to do better, I think, to appreciate that. I appreciate we live in a real world where things have to be done, but we have a solution to something that we've all been screaming about, but we then like, yeah, but that's not enough.

Estefanía 'Nia' Escobar-Kölle: 15:13

I think what makes me what I wonder is sometimes when we talk about AI in healthcare and the use cases, we focus so much on just operational efficiency or clinical efficiency that we actually forget the sociological and emotional component and psychological component of people, both providers and patients, right? And this is a very clear example of how we can take technology, embed it into processes, create operational efficiency, but also have wider impact that can actually go a very long way for all people involved.

Christophe Jauquet: 15:45

Yeah, I would say that I think it's already two or three years ago, is that the emotional AI will really make digital health fly. I think that is that is the one thing that was sometimes missing with digital health, is that you didn't have that emotional capability and within the tools. And I think that might be a big difference maker in the in the very near future. Anyone to piggyback on that any further or otherwise?

Dr Keith Grimes: 16:13

Well, maybe it's important just to flag something. This is really important. And I've seen many papers talking about you know the empathy of AI models being greater than that of humans performing the same task, and people, and as you said, yeah, about the Harvard Business Review reflection of that, that people are turning to chatbots for reflection, self-understanding, therapy counseling, and so on. But we do have to remember one thing is that remember how these models are built. They are simply predicting the next token, the next token, the next token until they predict they should stop. Now they do that. That's a very reductive way of doing it. But AI does not feel anything. You know, it is a prediction, it is a simulation of this. Now that's a mimicking. Yeah, yeah. But it's tremendously useful. I'm not saying that's not useful, but we do have to remember this because we as humans cannot help ourselves but attribute human characteristics to things that look as though they're behaving like us as well. And sometimes that's very convenient. And actually, if I go into an encounter knowing that's the case, I think that's fine. My worry is that when we forget that and we start to relate in a way as if this is a thinking, feeling entity and barring some discoveries suggesting otherwise, it isn't. And does that matter? Well, who knows? It's clearly doesn't matter to many people, and it's got really positive outcomes. But we do have entities that think and feel, the humus in here as well. And I feel and I suspect that remains valuable now because we all feel that it's gone, and will probably become increasingly valuable. That's my optimistic take if we can lean back into it, but always bear this in mind when we see it.

Christophe Jauquet: 18:00

Yeah, I fully agree, I think, for now. I mean, obviously, I wouldn't say I don't really believe it will be a sentient being in a way, but it might be that good or even better in replicating emotions and humans. Okay so the question then becomes what is real and what is not real? And is sentient really necessary to provoke the real emotions with humans? Obviously, there's a lot of you know, whatever, what if it triggers a wrong emotion and what if there's a risk to it and so forth.

Dr Keith Grimes: 18:34

But I but yeah, I think it's just coming in on that. Sorry to come in as you're thinking through, but it then comes down to accountability. Like, you know, my an expressed emotion from me and the effect it causes in someone else. I am accountable for that. I am the person, you know, if something goes wrong, they come to me with a machine. Is it the manufacturer? Is it the machine? Who is it? And of course, right now it isn't unless the manufacturer's, you know, has been negligent in some way. It's not the machine's fault, it's not the manufacturer's fault, it's the it's the person who's experienced the harm's fault in some regards because they've engaged with it in that way, you see the terms and conditions. And they are that's a really fundamental thing that people also forget. And maybe that's the that's the that's the difficult one to cross, because I think people will potentially relate to this mimicking machine, which is doing a fantastic job in that way. But if something bad happens off the back of this, who you know, who do they go to?

Christophe Jauquet: 19:38

Yeah, exactly. And then I mean it's always that way with anything in healthcare. but here it will be even more important. I'd like to piggyback on something you said earlier, Keith, and that was the democrat democratization of you know information and so forth. I think we share exactly the same, or I suppose, similar opinion on that. I have this one keynote I recently gave for Metscape in Basel to a whole bunch of medical affairs and pharma. And I think the one message there, because I always want to have like a simple message, which is very often a little bit exaggerated, but the idea behind it is to really make sure that people understand how big a shift some things are. And I think the message that I wanna that I wanted to convey was that before the 80s, there were two forces that were driving healthcare. And that was, on the one hand side, the unequal distribution of information, and the second hand side, the unequal distribution of solutions. Now, as from the 90s, 2000s onward, for sure, we had Dr. Google, we had social media, we had technology like the wearables, who are very often medically great right now, and we are getting into the AI era. And so, information, even knowledge, it's not just about information. I think knowledge will become more evenly distributed. I think in the past, it was very hard for a patient to understand anything. Everything the physician was the one with all the information and knowledge, and that will radically change. It's already happening right now. I mean, we saw the Dr. Google Mugs saying, like, don't confuse your Google search with my medical degree. I don't think you can use that for you know with AI now. Don't misinterpret your AI search with my medical degree because the AI might be as smarter, even smarter in many ways than the medical degree that a physician might have obtained over in the past, even. But the same is true about with solutions. And I think in one of the previous episodes, we talked about patient built solutions, right? Where AI is actually is going to drive a lot of power towards patients, allowing them to do a lot of things themselves, which otherwise they would have searched for within the healthcare system. And so to me, the biggest Shift that we're seeing with AI, it's a radical acceleration of letting go of the two forces from the past, the unequal distribution of information, knowledge, and solutions, towards a situation where basically the information and the solutions will be almost a hundred percent equally distributed if you exaggerate it maybe a little bit. But that changes like literally everything in healthcare, if you ask me. And my talk when I when I when I try to convince the healthcare industry about yeah, but what will really change is actually that. I mean, it's like everything will change, and you need to you need to be ready for that. And I think there's a couple of good things, definitely for patients, because it gives them a lot more power and insights and so forth. And obviously there's a lot more challenges as well. But happy to hear your thoughts on that.

Dr Keith Grimes: 22:53

Just as you were talking about the mug, you know, don't confuse your Google search with my degree. I've drawn a little scribble here of a mug, and the line on it now would see AI might know, but it doesn't care. Yeah, that's a good one. That's actually a good one. And I think I think that symbolizes it. Yeah, I've got it, I've written it here. Copyright me right now. But yeah, no, it's there's something just actually as you were talking, something simple about that. Is it kind of does know all the information and knowledge and even the capabilities in there? But it doesn't care. So the question is do you care that it doesn't care? Sometimes you don't. And that's well, it could care.

Christophe Jauquet: 23:32

I think it's yeah, that's a philosophical point. Yeah, it could care in many ways. I mean, this is what Nia was saying previously, right?

Estefanía 'Nia' Escobar-Kölle: 23:39

Yeah, well, I mean, how do you how do you define care exactly? does it care in a way that it's gonna give you interventions and it's gonna try to support your mental health or your physical health or any aspect, or does it care that it has to fulfill specific requirements that are have been designed as part of a care path and you have to stick to that? So it really defines how you define care. But I think because it's a system and it's completely unemotional, it doesn't care. You're completely right. But we can train it to care in specific ways.

Dr Keith Grimes: 24:10

Yeah. Do you care that it doesn't care? there's a difference about care as in a feeling that I have or a feeling that you have when I'm telling my story or expressing my pain, and I have a sense that the person or the thing at the other side of the room is sharing in that pain and experiencing or having an emotional reaction. I always sort of think back to the servers at a restaurant. You know, if you get great service at a restaurant and they're very smiley and happy and they look like they're having a great time, and you remind yourself it's like, well, maybe they are, that'd be lovely, maybe they're not, but I still feel good at the end of it all. And you know, there's an element of that, I think. It's just intruding into a very sensitive area, maybe. Intruding, entering to that area.

Christophe Jauquet: 24:51

Yeah. And I mean, all of that may sound like healthcare providers could become 100% redundant, which is not necessarily the care. I think the essence would be that in as you said, Keith, to try to find ways to care in a different way that an AI could do. Because I think the human value, maybe, maybe we don't have the monopoly on empathy, maybe we don't have the monopoly on sharing our emotions or dealing with emotions. But having this human interface definitely will make a difference in many ways. Now, piggybacking on what Mo said earlier, the way that I referred to it was the speed on the outside needs will be faster than the speed on the inside. So it's very important that physicians need to have some speed and adapting or integrating AI in their way of working. I think one of the things that I've been mentioning a couple of times as well is that, and I haven't read a lot about it actually, but it's something that I see happening in some way. Listen, listen to it and see if you if you agree with me or not. But we see more and more, if definitely if you look at pharmaceutical use of AI, we see more and more AI use in the creation of compounds, finding the right molecules, even doing the early tests. We saw the in silicomedcine, I think they're even you were using alpha fault, and they gained about five years to going into the I think phase two already, right now, if I'm not mistaken. The reuse of existing medication for other diagnosis, other treatments, other diseases, let's put it that way. What I expect AI to be doing in the next five to ten years, maybe not two to two to five, but I think five to ten for sure, is that we risk seeing a tsunami of new solutions coming into the system. New compounds, reuse of reuse of existing compounds, new studies, because with synthetic data, we will be able to do a lot more studies more quicker as well. You know, we'd be able to work through a lot more data faster. So I really expect within five to ten years a tsunami of solutions. And if you if you worked anything with physicians, and Keith, you are a physician, it's not always that easy to be, you know, on top of all the new things that are happening in healthcare and are coming out. So what if suddenly there's a hundred times more solutions that come to the market? How will physicians deal with that? It'll it'll definitely be a good thing, but if you don't, if you're not using AI yourself as a physician, then what?

Dr Keith Grimes: 27:27

Yeah. I mean, Aline, you spend so much time looking at new solutions, you know, you're across all of them all the time and you'll bump into the manufacturers, but you'll also bump into the physicians. What are you seeing in terms of their responses?

Estefanía 'Nia' Escobar-Kölle: 27:40

With AI, you mean?

Dr Keith Grimes: 27:41

Yeah, with this sort of what Christophe was talking about, you know, this profusion of new technologies with these highly hyped capabilities that may or may not be delivered. You know, what's your sense of how they receive this?

Aline Noizet: 27:54

What I say is that AI won't replace physicians, but AI, the physicians using AI will replace those who are not using it. And I think it's clear for them that it's coming. And maybe my role, what I do and what I love doing is actually sharing with them the benefits of AI and why they should use AI, how we can help them in the day to day-to-day improve their efficiency and helping their patients as well. But maybe Christopher, what you're saying, like I think it's a good news, but we're going more towards personalized medicine. So, yes, there will be more solutions, but personalized solutions for me, like for my disease or for my treatment. And when you were saying so many solutions, I was like, Yeah, AI will help the physician choose the right the right solution. That will be all integrated in the in the system. Hospitals are starting already.

Dr Keith Grimes: 28:50

Yeah, I think I hear it so often as well. I mean, I came to you because you see the different areas. But when I try and get doctors that I speak to or nurses or pharmacists or any and the clinicians that I speak to about this, they have that they have a sense that they're already running flat out, they're already exhausted. They you know they're struggling to keep up as it is, and so and everything they've been given before now is to try and make it easier for them to run faster, you know, like nothing's come along that has that has alleviated that, despite some of the promises. But what I try and explain to them, and this is the sort of happy future that I see, is that in any doctor, I'll say, there'll have been a time when they really enjoyed what they did, or there'll be moments when they really felt that they were being a doctor. And when they talk about that and you go back to it, it's generally conversation time, you know, they always they always had more time to do something. You know, they didn't say I managed to do occasionally they say I managed to sort of save a life in a half a second, but most of the time it's like I remember that breakthrough conversation with a patient, or I remember that time, or they go back and it always comes back to time. And so I try and talk about well, what can be done to bring that back. So we talked about nostalgia actually in the last thing. You know, there's this kind of nostalgia for a time when care felt better and or simpler. And so maybe the solution for AI is to allow us to be very complex, but for that complexity to be kept away to allow the human part to feel simple again. And I think if we if and that's why scribes people like them, because it feels simpler. Everything else feels complicated. So let's push to simplicity. And AI absolutely can do that, but it takes it takes care about the solutions we're building.

Aline Noizet: 30:43

But I also agree with you, Keith, because last week I went to the doctor and she put the scribe between us, and she was so happy about it. Like it makes me it made me happy to see her so happy. She's like, Yeah, I can really focus on you, we can talk, like I don't have to worry about that. And we had a great conversation. I'm like, well, that's fantastic.

Dr Keith Grimes: 31:04

It is amazing, and this is it, it all the studies are saying this. People feel better. There are some people that find it harder, of course, and some people that like to type, but the overwhelming signal is people feel better. What is wrong that feeling better is not enough for this?

Christophe Jauquet: 31:21

Yeah. And I think I mean that is one of the AIs that we're seeing right now. In my keynote on how it will change healthcare, I ext I try to explain that patients will use a lot of types of AI. They will have, you know, an AI that can run through the data that comes from their devices, they have AI that might suggest them N as one type of solutions, maybe even nudge them in certain ways or guide them through the healthcare system. But the same will be true with physicians. They also will have a couple of AIs that will help them in different fields. And I once did an exercise, and I think the transcribing one is an easy one. But you also have, and I gave them names, they're like there's Dr. Scout as an AI agent that they can use, which can help the diagnosis without even being very doing a lot themselves, but it can help that doctor specifically. Dr. Scout can inform the physician. There's Dr. Architects, which can help the treatment planning. There's a nurse purse that checks the execution of everything. There's a watchtower, which has a very specific way of understanding when things change, even before they change. obviously, there's a translator, which is the transcriber, but also maybe to other types of patients, education levels, digital health, digital literature literacy levels to explain everything. There's a clerk bot that do all the compliance stuff. There's the quality maven that checks all the quality regulatory stuff. And there's even the mentor at the end of the day that can tell you, you know, this is what we've learned today. Maybe next time, if you care for your patient, you should be doing it in this way.

Dr Keith Grimes: 33:00

There is all those are fantastic examples. There is a sort of future that I can see, or at least I think I can see. And maybe it comes from the experience I've had of what when I've been happy being a doctor. And I've explained some of that about that connection. But there's another time that I've been happy when I've really felt that I've co-created the solution with the patient. Now that can just be coming up with a common approach to what's going on. You know, we agree that we're going to do this and so on. But in my case, it has quite literally been co-creating, like building something with a patient. So at hackathons, for example, I really enjoy those too. AI gives us an opportunity to have that. So when I'm with a person, between there's a sort of team, there's the kind of the doctor, there's the patient, and then there's the AI, and there's sort of shared capability, which means that there is a world in which AI or healthcare can move towards a much more co-created solution world. And that hasn't been really been possible before, at least not in this way. I think this is something unique. And that really fascinates me because I mean it speaks to the kind of person that I am, and maybe it's not for everyone, but I get tremendously excited about the idea of having the time to sit with someone and truly kind of like work with them and their information and their wants and needs and build that tool with them for them. I think we're seeing the very beginnings of that just now. And that could be a very interesting future. I you know, I'm a digital health doctor, I practice medicine through the products that I build, but wouldn't it be great if the products I built, I built with the patient? I was co-creating. That's the role of that person going forward. Kind of makes sense, right? Yeah. I mean, I think it will happen regardless, but it'd be nice if we engaged and thought about it from the kind of medical way as well, because it changes how you're practicing the medicine.

Christophe Jauquet: 34:45

Yeah. I also think that it's not just about the creation of the solution, but it's also about create finding the solution in the moment. And I what I really think is that while you were saying that care will become more important, so a lot of the communication that's that today is happening is about you know finding doing the administrative stuff, asking the basic questions, and the care gets left out. If this could be automated, then you have a lot more time to spend some caring time with your patient as a physician. But what I do believe is that there will be a lot more communications between patients and physicians, but not the actual physician and the patient, but their AI agents. So I really believe that there will be this back and forth between the different doctor AIs that I've just mentioned and the different patient AIs that I briefly touched upon as well, where they will share information before they involve the real human beings. And when it when time is due and when it is necessary, then the real human beings will talk more focused, more maybe even humanly, because the administrative stuff and all the information sharing, which very often does not happen enough. Let's be honest today, will happen a lot more via those agents. That's how I look at things.

Dr Keith Grimes: 36:08

For Aline and Nia's sort of take on that, there's one other thing I wanted to add just to sort of throw in the mix, is we're seeing a rise of AI assistants and agents that have memory. That was the, you know, the fun fundamentally large language models don't have memory, conversation to conversation, nothing is passed through. But through advancing the technology, there are ways in which the memory of what went before informs what happens now. And you see that if you use Chat GPT just now, you'll have that already. You can opt out of it. But there's more and more building up deep contextual knowledge about who you are and your previous conversations, which informs the current conversation. That's a very human thing. And it's done to you know tailor how it operates. But that's a key component of what you've just discussed there as well, Christophe, isn't it? That I have, you know, my doctor's agent speaks to my agent, and then they work the stuff out, and then when they need to bring the patient and the doctor in, in you come.

Mo Zouina: 37:01

Yeah.

Christophe Jauquet: 37:04

If you don't have anything else to add here, I have a couple of things that I've that I'd like to share with you and have a quick thought on it. It's like the unusual ones, but not the examples in this case. Is that okay? Anyone has some other thoughts you want to share?

Aline Noizet: 37:20

Maybe I'm gonna share something that so yesterday Google announced that they're introducing the video calls where you can speak in different languages. So for now it's English and Spanish. And when you were talking now, you know, about like AI agent talking to AI agents. For me, it's also AI will improve. That's like further away, but on the human side, AI will help in that case. And I was thinking of that now there will be a more equitable access to care. Like anybody can talk to healthcare professionals. They can, you know, a patient that speaks like a dialect in an isolated region, they would be able to be understood, to communicate properly, because you know it's an issue, and they can access healthcare professionals from anywhere. And the same for second opinion. So maybe that's more basic AI, but I think it's very important advancement. Back to the humans, so humans talking to humans, not age on AI to age on AI, but being able to understand themselves without a translation in the middle.

Dr Keith Grimes: 38:25

There's also it's not just the human language, but there's also the professional language as well. There's an awful lot of talk about, you know, doctor's notes being translated to patient readable languages in different reading ages. But equally, the work of a doctor is constantly trying to translate what the patient is saying into medical language so they can operate with it as well. So you can imagine a situation where a patient at one end is describing their symptoms, but what the doctor is hearing at the other end is as if a doctor is relaying those symptoms in medical terminology to them to close that gap between understanding as well. So there's a lot of interesting things that can happen here.

Estefanía 'Nia' Escobar-Kölle: 39:01

I think one interesting component from the discussion that we're having is the fact that, and you started with this, Aline, when you when you said that you've talked to healthcare providers and doctors, is the fact that we cannot fully replace this interaction. So I think the fact that having this human in the loop will always be essential, at least until the technology reaches a pivoting point. But we will need the human in the loop because we want to talk to humans. And it's exactly what you were saying, Keith. Sometimes it's not just about the conversation, but it's about the body language. It's about the double negation that sometimes we know that LLMs actually have issues with the way that physicians take notes or the way that physicians can infer things that are just subtext or just through observation. So for me, when I see doctors using, sometimes I saw a meme the other day of someone who went to a consultation and said that the doctor was using Chat GPT, but they were using the same chat for all patients. Yeah, it was terrible. And they noticed and they asked the physician, and the physician was like, Well, I mean, I still get the answers, but then we're missing this contextual information. So I think we need a education on the physician side, because I think all patients, we all want to have better care delivered, but we also need to educate physicians to trust AI and to have a good use of the tools that are available to them. And how can we complement that with body language, with double negation, with casual conversations, with languages, with cultural differences and nuances? And I think this will be one of the keys to really unlock the value of AI in healthcare delivery.

Christophe Jauquet: 40:37

Yeah. Very powerful. Now, going back to unusual ones, what we just did was we're in this situation, we see what's happening and we're thinking forwards. I did an exercise a little while ago and I wrote an article about it. It's called the what was it again? The quiet AI revolution. And what I what I did is that very often if things change, it's only afterwards that you really realize how it changed our lives. So, what I tried to do is together with AI, by the way, I looked at what within 10 years, how we will, how will we look at things that maybe we didn't realize today. And I came up, I tried to translate it in a couple of quotes just to make you understand how it feels. So imagine that it is you saying this within 10 years. And I think the first one is pretty easy. I think it's about we prepared for AI to change how we healed, but we didn't expect it to redefine who delivers the healing. I think a lot of the things that we talked about already goes in into that direction. I think a more and also perhaps a little bit obvious one, but if you think about it, it hits me differently, is the following. Our belief in human wisdom about health was rooted more in ego than in evidence. That's a good one. Because we really believed physicians based on their human cognition, but there's a very there's a big limit to that, right? What they recall from what they've studied and so forth, intuition maybe is a was a big part of a lot of decisions. But AI will probably outperform that within 10 years, let's put it that way.

Dr Keith Grimes: 42:15

We're all having to think very deeply about what it means to be a patient, a person, a doctor, insert whatever you hear, which is why it's such a tremendously exciting and slightly scary time. You know, I don't think many of us have lived through this kind of thing before. Even I've gone through all these things. The internet, you think, is a big thing, but that happened quite slowly. You know, I went through a lot of early stuff. It was very slow compared to this. And even then, it didn't really redefine what it was to be a person.

Christophe Jauquet: 42:44

Yeah. Another one which is a bit more deeper, I think, is we thought that AI would make health and wellness easier. Instead, it changed what learning to be well and healthy or healthy and well even meant. So, what I try to say here is that we eliminate the struggle. And so the self-discovery that very often comes along with healing and personal growth and wellness will be automated. So, will we still have that kind of struggle or not?

Dr Keith Grimes: 43:16

Well, maybe the way of putting it is like, you know, 10 years from now or in the future, I didn't, you know, I now understand that I am responsible for my own health and wellness. It's not someone else. It's another way of putting it, you know, that enablement at the point at which you can personally access it, you can't help but face the fact that you have the responsibility ultimately lies with you.

Christophe Jauquet: 43:45

But here it goes automatically. And so is it still then still with you? Are we still struggling and doing the effort, or is everything just automagical?

Estefanía 'Nia' Escobar-Kölle: 43:56

I think if we talk about full integration in the future, we there will be prevention programs that where you just follow the care path. And you don't even have to wonder what do I have to eat, how do I have to exercise, what type of exercise is that good for me, what kind of routines do I need? But the information is already there. So this is basically done for you, and all you have to do is follow the steps, right? And I think this is something that we're lacking right now. Of we talked in the previous episode about Alzheimer's and people discovering the risk of having Alzheimer's or developing breast cancer or ovarian cancer through genetic testing. But a lot of the people were not acting upon discovering that there is a risk of having the disease, right? I think in the future this will be just automated, hopefully. So there is no question of what to do next and personalized.

Aline Noizet: 44:48

Absolutely.

Christophe Jauquet: 44:49

And the very last one goes back to the emotional part and the caring part, Kate. did AI support our emotional well-being? Or did it quietly redefine what it meant to feel understood and cared?

Dr Keith Grimes: 45:06

I think yeah, that's a really interesting one. And I definitely think we're seeing that happening. And you see the Harvard Business Review showing exactly that.

Christophe Jauquet: 45:14

Exactly that. Companionship number one use case for all AI. Already today, and we're nowhere near what the capabilities could be. Anyways, time to wrap up this podcast. A very nice discussion, a bit philosophical, a bit challenging, a bit positive, negative. Everything that we needed to have in this discussion. Thank you for listening, you listeners. If you like the show, don't forget to hit the subscribe button or go to Healthusiasm.com. Thank you to our panelists for their Healthusiasm. Thank you, Aline Noiset, Dr. Keith Grimes, Mozowina, and Estefanía 'Nia' Escobar-Köller. My name is Christophe Jauquet. We are the Health Enthusiasm panel, and we'd love to see you again next week for some more Healthusiasm. Ciao