How AI Is Reinventing Elder Care | Chia-Lin Simmons of LogicMark

0:00To figure out how to use an Apple Watch for an elderly person is a big ask. Everything is so tiny and the touch screen is overly sensitive. But I want a real person with, you know, intuition, mortality, fragility, experience of having been fragile and listen, right? Where do you see this going? where people can stay in their homes much longer than they can currently because there will be enough monitoring around them that people don't have to worry as much. For people over 50, more than 90 % of them want to age at home. One in four percent over 65 will experience a fall. So how do we help you not have that fall?

0:44How do we get predictive? And that's where AI plays a huge, huge role.

0:51Shailen, could you begin by introducing yourself and give a little bit of your background and how you got to LogicMark? My name is Shailen Simmons, CEO of LogicMark. I joined the company in June of 2021 as their Pivot CEO. The company has been around since 2005 and was actually created by an innovator in technology who had a firm belief that personal safety should and should always be available for all seniors. And so at the time, subscription related medical alert products were very expensive on a monthly basis. And so he created a product that if you push the button, it would go straight to 911, much like, you know, giving your cell phone, old cell phone away to emergency services, sort of connected directly to 911.

1:44and thus providing personal safety and medical safety for a lot of low-income and fixed-income seniors. And so when I joined the company, the company was in a bit of some major changes, certainly was in a bit of a pickle, hadn't created new products in, I would say, since 2015, and did not really have a connected product in a world where 80 % of Americans had a Netflix account. And so subscriptions were definitely going down in pricing, became more prevalent. But the company hadn't really been caught up in that. And so the board brought me on board as a Pivot CEO. And I have much more of a product background, having done work in connected IoT, connected vehicles, and in AI.

2:40And so I was very much interested in this category space because like a lot of people, the one in three millennials and more than half of Gen Xers out there, I'm a sandwich generation caretaker. So I actually have my own personal experience, much like everybody else, with a beloved family member having some not wonderful experiences with a medical or product. And so I remember filing at the time I was working at Google thinking, somebody needs to change this. And it turns out in June of 2021, I was given the opportunity to do that. And so I took it upon myself to sort of join the company. So and that's been where I've been ever since.

3:23Yeah, well, very, very timeless. I'm sorry, very timely, also timeless, because my generation, I'm not the sandwich generation, I'm at the tail end of the boomers. My parents have passed away, but I'm going to be needing something like this, and I have siblings that need something like this. as a matter of fact I've talked to my brother about this because he he's fallen a few times but let's talk about what Logic Mart does it's a it's a personal emergency response system and it's connected it has some AI capability maybe that's what I'm particularly interested in is the ai capability so if you could describe i think the product's name is freedom alert max uh describe what it does but particularly how ai is is used in the product absolutely so um you know i grew up like a lot of us with you know the iphone and kek it up understanding what medical alert is you know you kind of fall and you know you have to be conscious and moving in order to basically push that button to get the help that you need and it makes a lot of sense this industry came out of home securities and so it became sort of a thing where they're protecting your home with all these devices and sensors and so why not slap a device on your you know parents and be done with it and you're using the same call center, right?

5:10And so the reality is, you know, for a lot of people, especially, you know, late stage boomers like yourself, you're active, you're out there, this is not kind of what you want, you do not want to wear a garage opener hang around your neck. And the reality that it makes a lot of suppositions around this concept that you're going to be conscious, and that you're capable of moving. Half the time people fall in a shower. And, you know, If you look at common communities and senior independent living facilities, I think their solution is a product that mounts to the wall and has this really long string.

5:45And you have to actually crawl to it and be conscious and crawl to it and then pull that cord because it's also very difficult to access that fault detection. And so, you know, the feeling was coming out of IoT as a business and AI, you know, as a business, as a tech executive, my feeling was like, wow, like this whole entire business is totally stuck in the 1980s. And so everything else is evolving, even connected cars. I mean, those folks are working in cascading like, you know, waterfall models of like seven to eight years. And even they're moving faster with connected cars. Like what's happening here?

6:22And so when you really look at it, it's because, you know, there is a real sense that this is a category area that's not too sexy. And so a lot of development against it. But in actually, you know, in a real sense, like this, you know, this concept of personalized, hyper personalized, you know, health monitoring and care has been really prevalent already. You know, anybody who has an aura ring, an Apple watch, you know that it's tracking remote sort of data on you to help you get better health. And so this is really not that far of a step from it. And so what we've really done is to look at what does it mean to bring that sort of personal safety and security and health monitoring to a category area that people, I think, with the aging boomer population have an expectation for that.

7:10Right. You guys are living through an experience where you have your eyewatches and your auras and your connected cars. Right. And so why should you be, you know, taking care of something for yourself that's so personal and so important, you know, your personal safety and suddenly go back to the 80s. And so when we talk about the education of AI, really what we're talking about is a shift from this industry of being a reactive technology. And let me just state that we have to be really great at it. So applying our AI to fall detection and really understanding, like, is this a fall or are you just sitting down too quickly?

7:45Right. And not triggering the false positive is really important. And why that's crucial is everybody will tell you, hey, so like, why would you bother with one of these things? Right. You have an iWatch. And so my feedback is like, well, there is that New York Times article about how, you know, an entire sort of towns 911 got shut down because everybody was skiing in that ski town in Colorado. And so when you're going at a high velocity, you know, not to geek out here using an accelerometer that's on your watch and you stop. Right. That's crash detection. That's what iWatches are great at. Right.

8:19And that's because you're trying to serve an eight year old all the way to like a 70 year old with an iWatch. That's a huge audience. And like no first 80 year old person I know was moving at such a fast velocity and stopping so suddenly that they could potentially trigger that type of fall. Right. The reality is, is that when we speak to the Veterans Administration, it's the largest hospital network in the United States. Right. The clinicians tell us stuff like, hey, like, can it predict a slump? because typically when people fall, it's kind of slower. And so having great AI apply to understanding what's a fall and what's not a fall and then actually tailoring, using the AI to tailor it to your personal experiences and understanding, for example, from a reactive technology that, you know, on Tuesdays and Thursdays, what looks like a fall and we're contacting you like, hey, Greg, like, did you have a fall?

9:13Are you okay? And you're like, no, no, I'm just in a yoga class. And we look at your GPS location. We look at this and we're like, oh, you're okay. And then Thursday happens again, we're like, hey, are you okay? So we see this happening. And then now your digital twin, we created something via AI to say, Tuesdays and Thursdays at 10, Greg's doing yoga. Looks like a fall, but not a fall. So now we're not going to bother you and we don't have these false positives, right? Making you not want to wear the device. So second joke, so that's us being really good at fall detection and reactive technology.

9:42But what really I think AI is so exciting in terms of application in this industry is predictive analytics. And that's where I get super, if I don't seem excited enough already, I get super excited about this because, you know, when you think about, you know, falls, one in four person over 65 will experience a fall. and, you know, it actually has catastrophic sort of like, you know, responses after because typically people will just say, oh my gosh, I had a fall. So I don't want to do too much walking. And so they'll be less active and there's much muscle atrophy and a lot of things that can happen, you know, from a cardiovascular perspective as well.

10:22And so what you really want to do is encourage people to walk and be active, but that initial fear then triggers. And so how do we help you not have that fall how do we get predictive and that's where ai plays a huge huge role so when you look at fa max like our freedom alert max we put things in such as you know medicine reminder and we do that because you know poor medicarens you know it alone contributes about you know 500 billion i think in avoidable health care costs and affects about 125 000 like sort of thousand like predictable deaths right so meta-adherence is important but meta-adherence also gives us a sense of like whether or not you're ill you're much more prone to like being dizzy and falling it gives us one sort of anchor of a sensor data um on understanding whether or not you're more prone to that we put in you know what is very typical of this idea of like activity tracker so you usually wake up at 10 and 8 8 o 'clock in the morning and you know you take your device and you sort of carry it with you and then you go to sleep at 10 and you typically take around like 5 ,000 steps a day, then suddenly we're slow attrition.

11:37You know, you're getting up at 10, going to sleep at eight, you know, slowly but surely. And then your steps went from 5 ,000 to four to three. Sometimes it's very hard for caretakers for seeing every day to see that sort of gradual change. You're not there all the time. And so being able to see a longitudinal sort of like analysis that's personalized to you to say like we're seeing this pattern and then taking that and building it into a digital twin and saying hey so when i've seen this on a 65 year old woman who's on coumadin which is a blood thinner and saw this pattern of recognition what happens with a similar sort of age and sort of demographic of person what happened to them three months after this pattern, six months after that pattern.

12:25And so perhaps we see six months after that, that they had a fall. And so that gives us to your caretakers, both professional and personal to say, Hey, so this is what we've seen in aggregate. So is it time to put in a walker? Is it time to get in a cane? Is it time to do more physical therapy to strengthen the time to take a look at, you know, the mediterrancy issue? Because mom seems to have missed like, you know, three of her blood pressure meds, you know, every week. And so what are the other ways to do it? We've reminded her and she doesn't take it. And she doesn't tell us like, we're gonna.

13:00And so all of these things in combination gives us a sense of potential predictability and looking at it longitudinally and hyper-personalized to you, as well as using a hyper-personalized data to then compare it to a larger anonymous aggregate to say, hey, are we seeing a pattern of recognition here so yeah yeah and it's interesting because there has been a lot of research about these longitudinal studies that that can show patterns that lead to declines. I spoke a couple of years ago with Fei-Fei Li, who was working on ambient intelligence, and the idea there was that rooms would be fitted with sensors and cameras and learn the patterns of people in their homes.

14:03And she was thinking about Elder Li. And over time, would be able to detect changes in pattern that are indicative of cognitive decline or physical decline. What models are you using to do this? And is it a single model? Do you have multiple models? and I presume the models are in the cloud or how much of this compute is happening on device and why could this not be an app as opposed to being a separate device? Sure, I think that's some very good questions, quite a few. So I'll sort of address it. So first related to Fei-Fei, so we actually are in current beta on a non-wearable sensor, as we've mentioned, and I think you mentioned camera, the privacy is a number one concern for us.

15:06We actually have a camera for the Freedom Alert Max, but in order to protect people's privacy, we do a couple of different things. So one is that we tend to process on the edge, right? So we're looking for pattern differential. And most importantly, because like, you know, for us, safety is one of the major drivers of people wanting to put anything into their home or wear anything, right? So it's important for us to feel that, you know, our customers and our users, that's the center of everything we do. Human-centric, you know, product development is very key to what we believe in as this company.

15:44And so, you know, in order to trigger some of that data, sort of, you know, pattern recognition, we actually do the edge processing. So you'll see us do things like processing in, you know, whether or not sort of things are changing sort of in the device. But then the data is that we're transferring is actually pattern and it's tokenized. So we have patterns around tokenized data traveling because we also have to get it to a call center and get that. The call center needs to be able to see it and then get it to the EMT who's advised that knows that you are on Coumadin and here's some of the issues.

16:19And then it has to disappear from the service. Right. So all of those things are from an edge computing perspective is really important, especially as, you know, for folks like Fei-Fei and myself, like we're looking at trying to put more devices into people's homes. Privacy is on our mind. And so Freedom Alert Max has a camera. You'll see us basically we're actually doing, can't really talk too much about it, but we're in the middle of doing beta for a product that is AI driven and it's multimodal sensing. So we're utilizing, you know, our own sound based AI algorithm to help us detect a large, very heavy or if it's a slump.

17:02And it's you'll see us actually start implementing some of these things that's more sensor oriented. That could be including like vibrations. It could include potentially in the future cameras. Cameras are very, very controversial thing to put into a home. So, you know, our initial sort of foray into a beta tested sort of experience is to try to make it more privacy oriented. But getting a hub in the home is something that we're rapidly working on and is currently in beta with a couple of different retirement communities and senior living facilities as well as they're evaluating and helping us sort of evaluate that data.

17:42And it would work with a wearable device because you're not going to be in your home all the time. So all of these things should be connected, the data process on the edge using multimodal sensor, sound, accelerometers, altometers, and gyroscopes, and all of these sort of vibrations, all of these things in combination to help us actually get a better understanding of what's happening from a health perspective. And so you'll see us moving towards things like, you know, we'll be launching a watch product and that watch product will be much like, you know, an iWatch, but we're really, really focused on trying to create a product that's really for the right audience.

18:26I spent quite a lot of time looking like I'm some kind of loiter in an Apple store but trying to listen to what people are saying and I often hear people wanting to buy a you know Apple watch for their aging parent deciding not to do it because they ask the sales people like can you remove Apple pay and remove this this this and this and they're like no like this is part of the Apple watch experience and Apple experience and you know my feedback is I 100 % stand behind Apple for that because you know they're a watch for everybody we're a watch for um somebody who's more senior and our face is meant to do that and our services are tied to personal safety right and medical and so and that's you know you'll see us look at even like blood pressure blood ox potentially in the future and like bringing you know partnerships into a cloud um to help us um connect more sensors into an environment.

19:25I will say that I am a bit of a, you know, a Google girl, for lack of better word, a zoogler. I'm a believer in having an ecosystem that is very much encompassing partners who are very good at what they do. I will never, ever want to create a blood pressure monitor. I think it's very difficult. Blood glucose, very hard, not my expertise, but believe that those things are quite crucial for the success of monitoring people's health, especially in aging. And so you'll see us look at, you know, what it means to, you know, bring a hub into the home, do what we're good at, which is fall detection and, you know, personal, like sort of, you know, predictability of health and all of those things.

20:11But everybody has a piece of the pie. I mean, today, what's calling is that down to like McDonald's, everybody has their app and that data sits in like what the geek part of me says like isolated deposit data's like suck right like i hate that like because you're just kind of ruminating on your little piece of the puzzle and so what doctors and healthcare providers are doing is trying to bring together these little disparate pieces and so it's very difficult to do that and so i'm a believer in that like you know you put out the apis in a safe sort of a hipaa compliant environment you use this ai's to basically gather this data to actually help us do what they do best, which is AI is looking at a recognition, is looking at longitudinal sort of like anomalies, all of those things versus what I think we're seeing in terms of the application of AI is today.

21:03And when people talk about aging, which is like, let me get an AI, you know, caretaker in front of you. So like meet Ava, who's going to like answer your fault, you know, now. And like, did you have a fall? I'm like, you know, honestly, I've been in a customer support loop with AI. You know, when I'm in a time of crisis, I might've broken my hip. The last thing I want to do is talk to, you know, the Ava AI, right? I want a real person. I want to be able to detect a sound and say like, that sounds like it's probably a fall, but I want a real person with, you know, intuition, mortality, fragility, experience of having been fragile and listen, right?

21:51Because your AI will tell you, you know, I don't hear a sound. I don't hear utterance, you know, when I'm running my LLM, which we do as well. And I don't hear the word help in any like language format or what sounds like help. So there is nothing here, right? That's what the AI would do. That's what we teach it to do, it's recognizing, right? But a person probably say, you know, get on, like, once we sort of identify an issue, that person will sit there and will listen. And they will say, I don't hear any words. And there is breathing. But it sounds like it's heavy breathing. It sounds labored.

22:29That nuance is very difficult. And we will still be working on that from an AI perspective, you know, obviously, but I think, you know, that initial application of AI as an agent, I think is much more difficult in the initial. Yeah. So, so how, as it stands today, is describe the device. Yeah, I'm happy to. So, and I will do a little bit of show and tell. So it's small. It looks like a cell phone because it is literally a cell phone plus a medical alert product. And it's small because we want it to be portable today. When you ask a lot of older people to hold this phone, it's really hard to grip the older you get.

23:18And it's got a lot of functions. It's got a lot of things that happen. And it's a lot of stuff that you don't use. I mean, even as a former Google person, we knew that only like people only use like 35 % of their apps on their phones. Right. So we make it easy. You know, you could program your there's a caretaker app that happens on an Android and an Apple phone and you could program in all the phone numbers. I could say like, you know, Chris and like son and my grandson. And it's all sort of like easy and you don't have to sort of like fumble and like do all of that. So all easy to access and big screen.

23:55It has a emergency, you know, mental health crisis hotline because actually, ironically, seniors are actually been shown to be a little bit more prone to suicide because it's loneliness, it's isolation. So we have that sort of pre-programmed in. We actually have fault detection already built into this particular product and it has a camera front and back. So if actually fault detection is triggered, that's the only time that you as a caretaker can actually then see what's going on. And if you can see that it's down, you can sort of change camera perspectives and you can see that from your parents down.

24:38The apps themselves for caretakers allow you to build your village, as I call it, because reality. A lot of us live far away from our parents, perhaps in another state. And so your village, people who love you and care for you can be the next door neighbor across the street from mom, my sister, who's, you know, 40 minutes away by car and I'm in another state. And there's nothing worse than when there is an accident and people don't know what's going on and somebody's not on call. So what we actually provide is what we call triple protection, which is, you know, family members can schedule and say like, you know what, I'm going to be traveling to Paris, you know, my family and the next door neighbor is like, you know, not going to be around.

25:23And so like my sister in another state is on call. And so they're the first person that, you know, such something happens, they get triggered and everybody gets notification, but they're the first person called. And then also our 24-7 monitor service, which is U.S.-based, is the backup as well. And then 911. So you have three layers of protection. And that's what we hope for, right? Yeah. Particularly on the Apple Watch and, you know, trying to build this into a phone. You know, I went through this with my aging mother. I mean, you know, to figure out how to use an Apple Watch for an elderly person is a big ask.

26:15Particularly, everything is so tiny and the touchscreen is overly sensitive. And so, yeah, I can see that, yeah, it makes sense for this not to be an app on one of those other devices. I mean, that's not to say that we won't put out a risk-based product. It just will be, the user interface just has to be tailored towards, I think, an audience that's different. My mother-in-law had, you know, Parkinson's, like early Parkinson's. And so she started out using an iPhone, but her hands were shaking too much. And so we really needed a screen-based, touch-based product that is like you're doing this versus trying to do this sort of thing.

27:00So we learn from, you know, really focusing on what we're seeing our, you know, customers go through. And we are a believer that your, you know, your experiences is just, I mean, there is a great product for, there is somebody who's developing a product that's meant to be for everyone. And so there's always a bit of a compromise when you're using an iWatch, right? It's meant for just as many people as possible. And I think the reality is this product, we will probably never, well, we'll roll out services that are really intensively focused. So geofencing is one of those, right? Six in 10 people with early memory care issues and Alzheimer's will wander.

27:41And so you can use our app to sort of create a geofencing experience so that when your family member wanders out, then, you know, you get alerted and you can actually find them. So I just literally read this morning an article at the San Jose Mercury about how silver alerts are on uptake. And it's because, you know, there's opportunities here to sort of try to find people, you have to find them within the first 24 hours. There's too many like scary stories about not knowing that your parent had wandered and that not being triggered. And so then they have been gone for two or three days and they pass, right?

28:21And so we think about that and we're like, look, you know, you should set up the geofencing and having the camera front and back, then, you know, assuming they're wearing the device, then actually, then you could see a little bit that they're definitely moving somewhere. And so now we can, you know, give you the GPS locationing of your parent, right? And they're definitely not in a home. And, you know, there was no indication of why they would be leaving a home. And so you could contact them, you could do a number of things, because this is also a phone. So then you can actually, you know, make sure that everything's happening and God forbid that you're in a meeting and you can't, then our 24 seven monitor service will go in and say like, hey, Mr.

28:57Smith, how are you doing? Like, you know, is there anything I could help you with? And so I'm lost or like, who are you? Like, and sort of, you know, we can start making some of those opportunities to ensure that you're not going to be, you know, a tragedy waiting to happen. Like we're going to find you within not 24 hours like silver alerts but literally within hopefully minutes if we can yeah and so that's yeah and where are you in the development because uh so you you spend a lot of time on on fall detection uh on refining that um and and you're saying that detection happens on device you're sending data, as you said, pattern data to a mother model.

29:50Is that right? So we have what we call the caring platform as a service or a CPaaS, like SaaS, like CPaaS system. And so that does pattern recognition, like analysis and all of those things. But the initial sort of analysis for like FALS, those algorithms, they reside within the device themselves. And so freedom alert max is already in market um it already has geo fencing it has all of those services that we talked about the you know um suicide prevention again loneliness you know if there's three pillars for success for aging it's like physical health mental health and financial health right so you know we're really barely touching the mental health portion but then you know but that's their um the capability to sort of you know jump in during an emergency when is triggered to make sure that your parents are safe.