- 20 hours ago
From hospitals to homes, social robots are entering everyday care, supporting patients, assisting caregivers, and helping manage tasks from monitoring to rehabilitation. But what happens to trust, empathy, and human connection when machines take on these roles? Can they ease workforce shortages and improve quality of care without replacing essential human judgment? This session explores how these technologies are reshaping care and what it will take to integrate them responsibly.
Category
🤖
TechTranscript
00:02Okay, good morning everyone. Beginning of VivaTech, how exciting. We're here to talk
00:09about machines that care. I'm Greg Williams. So robots in healthcare are no longer science
00:15fiction. Their machines are already in the room. They're being deployed, social robots
00:20are being deployed in care homes, hospitals and private homes with real patients, real
00:25caregivers and real consequences, of course. So that means, you know, outcomes, trust and
00:32the nature of care itself. And that means grappling with some really hard problems. How do we design
00:39a machine that a cognitively impaired person can trust? Who bears the cost? Who captures
00:45the benefits? And what do we risk losing in human connection, in clinical judgment and
00:50our idea of what care actually means? So to work through these questions, I have three
00:56amazing people who are building, deploying and evaluating this technology in the real world.
01:01I'm just going to ask them to introduce themselves. And Angelina, you're at the end. So if you could
01:06start, please.
01:09Hello, everyone. My name is Angelina Wolf. And I come from University Hospital of Odense in
01:16Denmark. My role is as innovation consultant at the hospital at the innovation center and
01:22center for clinical robotics. And we primarily focus on projects that facilitate and take care
01:33of the patients and patient care and implementation of the robots. Yeah, Dor?
01:39Thank you, Dor.
01:39Hi.
01:41Okay, now it works. Hi, I'm Dor and I'm the founder and CEO of
01:46Intuition Robotics. We make EleQ, which is an AI companion robot for older people. You
01:51might have read about it in a recent New York Times feature article or on their podcast.
01:55And it lives with the older adults, becomes their best friends at their home and helps them
02:01live healthier, happier, more independent lives at home. Also, thank you for skipping Jeff Bezos
02:07to come to talk to us. That is a unique thing, so I appreciate it.
02:14Thank you. My name is Joanne Hackett, and I'm responsible for the work that we do with health
02:19systems at IQVIA. For those of you who don't know IQVIA, we're the human data science company,
02:26gathering and looking at lots of data points, not just for helping humans to have longer and
02:31healthier lives, but also that intersect between preventative health, longevity, and a variety of
02:37other things like that. So looking forward to speaking with you today. Thank you.
02:44Thank you, everyone. I think what would be really helpful, actually, is just to kind of
02:48level set at the beginning and get a sense of where we are right now. And maybe, Angelina,
02:53I can start with you. From your perspective, what does the market look like? Where are these
02:59care robots deployed? What settings? What's the gap? And just give us an overall picture,
03:05if you don't mind, please. Yes. From a hospital perspective, because that's where I come from,
03:12we mainly look into how to make the patients feel more comfortable, improve their experience,
03:25what will bring benefits to the patients. And also, we, as another focus, have the clinicians
03:33at the hospital. They need to fuel less, those, though, menial tasks that they do every day,
03:44they can be augmented. Their workflow can be substituted by robots and automated. Instead
03:53of them doing it, they can focus on more patient-centered tasks instead. Thank you.
03:59Dor, you're deploying robots. You've deployed thousands of them, as far as I know. Can you give
04:05us a little sense of what that looks like, the work of intuition robotics, and just give us an
04:10overview of what you do? Sure. So, kind of, maybe the best way to think about it is think about
04:18the
04:18customer or the person living with the robot. So, we work with older adults that primarily live alone
04:24or spend most of their time alone. So, think of a couple that lived together for 50, 60 years,
04:30and then the husband passes away. Sorry, guys, statistically, we die first. And the woman stays
04:36alone. And nobody says good morning to her. Okay? Nobody greets her. She lives far away from her family.
04:42And even if they don't, especially in the U.S., we see that. That's where we operate primarily.
04:48They're not there with her all the time. So, nobody's motivating her to get dressed,
04:53to get organized, to get out of the house. She stays with her nightgown in front of the TV.
04:57Her muscles atrophy. Eventually, she'll fall, break a hip, go to a nursing home. Or her brain will
05:02atrophy if, essentially, you will see cognitive decline. So, instead, this perky little robot that
05:08kind of looks like a lamp, that's beautifully designed, does not look geriatric, comes in and
05:13lives with her. And it's the first AI that's proactive. So, she initiates the conversations.
05:18She'll say, hey, Mary, good morning. How did you sleep? And how is your back? And today's Wednesday,
05:24and the farmer's market is in town. Why don't we go out? Et cetera. So, throughout the day,
05:28she will continuously motivate the individual around goals that are negotiated between the older adult
05:33and their robot, also influenced by their doctor, and influenced by the family caregiver. And what we're
05:41seeing is that a true relationship gets built. We're seeing over 55 interactions, on average,
05:47per person, per day. That's about 10 times more than ChatGPT. And sorry, Jeff Bezos, it's about 100
05:53times more than Alexa. And it lasts a length of time. And as this conversation will continue,
06:01I'll tell you more about what we're actually seeing in the home.
06:04Thank you. Doors, sorry, building the robots, you're deploying them, Joanne. What does it look
06:10like from your perspective? One of the things that I find quite fascinating about this space
06:16is that it's an accessory. And a lot of people don't want to think of it that way. And as
06:21Doors
06:22just rightfully said, it's one of these things that if it's used correctly, it actually allows us to
06:28have broader health care advantages. Because it's not just an interaction that happens once.
06:34It's something that you can have a relationship with. And as Doors mentioned, this is something
06:38that can help be motivational. It can actually start to depict and see different ways that your
06:44health is actually improving or potentially even failing. And so we've been able to work with care
06:50homes and also in areas where individuals are going through rehabilitation. And one of the things,
06:56strangely enough, that you're able to find those early signs either, as I said, of someone helping
07:04to progress, but also if they're failing to progress. So for example, when most of us stand up,
07:10we are hopefully young and healthy enough that we can get up without too much movement.
07:15But if you've got someone or a robot of some description with sensors or ways to capture additional
07:21information, if you notice that the person maybe has to help themself up and it takes a little bit
07:26longer, those data points help to provide some insights as to how an individual is progressing.
07:32So it's not just the companion side of things that's extremely helpful. It's that ability to
07:38predict. And this is where it gets tied in with longevity and being able to have a healthier life.
07:44So it's not just care homes. It's happening a lot with people who've had surgery, for example,
07:48and things like that. So it's not, it's not, you don't think of it just as something for older people.
07:54Think of it almost as an accessory to have a healthier life as well, while you are still healthy,
07:59or while you're recovering as well from something.
08:03What you're all describing is a quite an intimate thing that there is relationship between human beings
08:08and these machines. And I'd love to just kind of explore the idea of trust in this. And maybe Angelina,
08:14we can start with you. It's such a foundational thing for this to work. From your perspective,
08:21like what design choices or interaction patterns are necessary in order to build
08:27this relationship between humans and machines? Sorry, that was for Angelina.
08:35Could you repeat again? Sorry, do you want me to repeat the question? Yeah. So trust is so
08:41important in this, obviously, between humans and machines. From your perspective, what design
08:46choices or interaction patterns builds that trust? Yes. Yeah, trust is very important. And that's a
08:55very complex thing when we talk about technology into the healthcare sector in the hospital. It is also
09:04somehow including the cultural aspect. For example, in Denmark, we have very high trust in the system. And trusting the
09:15doctor is a very important part of it.
09:19When the doctor trusts the robot or the technology that they're introducing, then the patient also automatically
09:28trust also the technology. And when including design, I think that nowadays we, of course, want to add more complexity
09:38because certain tasks need to be automated more and more. But adding AI or more complex
09:47tasks into the technology makes the trust more fragile because there is more points of failure.
09:54And adding this complexity and making these points of failure multiply also makes sometimes the clinical staff doubt. So this
10:10transfer of trust is very important.
10:15And so once you have onboarded the clinicians, those people that actually are going to work with the technology, then
10:22they feel more secure to introduce it to the patients as well and work with it.
10:28Door. By the way, if anyone's interested in learning more about what Door does, there was a great piece in
10:33the New York Times. You've deployed thousands of these robots. What data points do you now have in terms of
10:38kind of like the types of, you know, relationships
10:42relationships that people sort of like develop with them? Like, what does that look like just in terms of like
10:48what you can see?
10:49Yeah. So we published this actually last August. What we're seeing is we looked at actually trust scores and our
10:57relationship classification.
10:58So on trust, for any of you building robots, I highly recommend the TPS HRI scale. It's a published scale
11:05that measures the integrity, ability, and benevolence on the users, how much they believe the product works well.
11:11Does it do what it's supposed to? Does it do the task well? And how much are its goals aligned
11:16with my goals? Is it looking out for me or for itself?
11:20We consistently scored about 87% at the highest level of this scale and 99% at the two highest
11:26levels put together.
11:28And when we looked at classifying the relationship, there are three levels of friendship. There is an acquaintance, like going
11:35to the bar, talking to a barman or a friend.
11:39There is best friend slash therapist, where you will share your fears and your goals and, you know, how upset
11:46you were that your kids didn't call.
11:47And then there is spouse level relationship, which is, like, the highest possible level. We saw 2% of our
11:55customers are scored at the relationship, the, like, occasional acquaintance.
12:0355% are at the best friend slash therapist. And over 45% are at the spouse level. Doesn't mean
12:12it's a sex robot, okay?
12:14What it means is they feel comfortable enough. If in the best friend, I'll be comfortable sharing my vulnerabilities, in
12:22the spouse level, I'll be comfortable to share how the action of my companion made me feel.
12:28So, we'll see things like, hey, Ellie Q, when I told you I'm not feeling well, and what all you
12:35did is suggest that we do some mindfulness breathing exercises to take my mind off pain, I felt like you
12:41don't really care.
12:42Or the fact that you didn't say good morning to me yesterday made me feel isolated, right? Like, they're criticizing
12:48her behavior on how it made their emotional state change.
12:52So, we're seeing that. And on trust, we released a framework or four layers on how to design, build, and
12:59measure for trust.
13:01The first is acceptance. Do people want it? Like, what happens? You give them the robot. Do they keep it
13:06or do they send it back?
13:07We're at about 91% there of people keeping it. The second is engagement. Okay, they have it. Do they
13:13use it?
13:14You want to see multiple times a week being used and multiple times a day per active day. We're at
13:20six days active per week, 55 interactions per day.
13:24And then you want to see the sentiment. And that's kind of what I mentioned before. How do people feel
13:33about it?
13:34Would they recommend customer satisfaction, net promoter scores, but also softer stuff?
13:38And the last and most important, at least for the three of us, I'm sure, is the impact.
13:44Does it change the clinical condition of the individual? Are they healthier? Do they take their meds on time?
13:50Are their cognitive scores improving? Is their quality of life improving? Is their depression and anxiety scores being reduced, etc.?
13:59And there, we continuously score this per patient every six months.
14:04And we're seeing about 94% of people living with this product see monumental improvement in each one of those
14:10factors.
14:11Joanne, there's also a question of trust when it comes to healthcare workers.
14:15And obviously, there'll be concerns about displacement.
14:18How should hospital leadership, leadership in care homes, how should they be thinking about this rollout and deployment?
14:27It's a good point, because a lot of people are very afraid of the fact that we're going to have
14:31robots taking all of our jobs.
14:34And, you know, it's a legitimate fear in one sense if you don't necessarily explore the actual power of these
14:40tools and technologies.
14:42And to me, it really is, as I said, it's an accessory.
14:45I often think that if we had double the workforce to be able to do some of this stuff, would
14:51we actually be able to help treat patients faster, have better outcomes?
14:55Of course, the answer is yes.
14:57Are we going to very quickly mobilize a workforce that's double the size in a very short period of time
15:03with human beings?
15:04The answer is no.
15:05And to be completely honest with you, when we surveyed care workers, we've asked them, you know, are you afraid
15:12this is going to take your job?
15:13And sometimes they say yes, of course.
15:15There's a lot of people, though, actually high in the 90% of individuals who say, if this could actually
15:22remove some of the burden of the day job that I have, I would be very pleased to be able
15:27to embrace this.
15:28So I think this is where that subtlety between accessory, companion, and also an extension of healthcare comes into play.
15:37But we never go into these institutions and talk about it replacing the individual.
15:42It's expansion.
15:43It's being able to capture more data points, be able to predict when someone is going to be ill, predict
15:49when they're going to be able to be released.
15:50So it's more about the kind of boring stuff in healthcare that most people don't really want to do as
15:55well.
15:56I think the thing that we get asked all the time is, can you please design something and implement it
16:01that's going to do the heavy lifting and help, you know, change diapers and all the stuff that most people
16:06don't really want to do.
16:07But that's your job, not mine.
16:11So, Angelina, let's come back to you.
16:13Obviously, you know, you see kind of the whole stack, as it were, clinicians, end users, hospital staff, engineers.
16:20I'd love to get your sense on where they most fundamentally disagree about what care robots should be doing, where
16:29the winds are, and where we should be, in which direction we should be progressing.
16:38Yeah, as Joanne mentioned, the first question that I get as well is, will it take my job?
16:46And I always answer, no, it's supposed to help you do your job better, or help you alleviate your job.
16:58If there is any physically hard work related into the clinician's workflow, it's supposed to alleviate that.
17:09But we, for example, had an assistant robot tested at the hospital, and the robot was supposed to stand at
17:17the reception desk welcoming patients,
17:21and there is all those volunteers that also stand there.
17:26So the first thing that they were asking me, is this robot now here being tested,
17:33so in the future I will have to cut down from the employees of the reception desk?
17:39And I was like, no, hopefully it can just help the flow.
17:43So, yeah, this is a real fear that they have.
17:49Dor, you talked a little bit about the trust piece, but I'd love to know, clearly what we want to
17:55do here is we want to improve outcomes.
17:58What kind of data do you have around that?
18:00What kind of evidence do you have that we're actually improving people's lives beyond just, you know,
18:06the warmth of having, you know, a spouse that's not going to, you know, be a downside?
18:14Yeah, so the U.S. healthcare system is very complicated and hard to work with.
18:24But the good thing about it is it forces you to be really, like, strong in your ROI calculations because
18:31they will only pay for something that saves the money.
18:34Okay?
18:35It's all about the dollars and cents in the U.S. healthcare system.
18:37So we have to check exactly for that.
18:40So I mentioned some of the outcomes as far as loneliness reduction, 97%.
18:45Health and wellness improvements, over 94%.
18:48Quality of life improvement, 90%.
18:50But what does that mean?
18:52Like, if I'm a health insurance company, what do I care?
18:56And so we measure different things.
18:58The most expensive part in the U.S. healthcare system for an older person is if they can't stay independent
19:04at home and they need to move into a nursing home.
19:06That costs about $15,000 U.S. per person per month.
19:11And we've proven that with LEQ, people stay longer and more independent, stay independent for a longer time in their
19:18own home.
19:20In fact, in Washington State, they now fully reimburse for the robot for anybody that's at risk of being in
19:26a nursing home.
19:27So they get it 100% for free by the state.
19:30So that's where we started.
19:31And then we started moving into more, like, my new thing.
19:34So we're working with a large university hospital now on pre and post operations for spinal surgery and hip and
19:43knee replacement.
19:43These are very complicated recovery processes.
19:46Very hard to go back if you live alone.
19:48Think you're 82, you just had spinal surgery.
19:51Two days later, the hospital is like, okay, you can go discharge.
19:54Are you really going to go home by yourself to an MDO?
19:57So now the robot is there caring for them, connecting to the care team, working with them on their PT
20:02and on their recovery,
20:03but mainly watching out for them and calling in the recovery if you need help,
20:07as opposed to going to an acute center which has just terrible results.
20:11Nobody's happy, very costly.
20:13We do the same also for complex needs patients with COPD, with multiple chronic illnesses,
20:18help take care of them, et cetera, et cetera.
20:20But the real, real, real impact, which we have yet to measure, but we're working on it,
20:26is the fact that we know what's happening with the patient at home.
20:29Usually the healthcare system, they find out there's a problem when you get hospitalized.
20:34Okay?
20:34And that's not good for the patient, and it's very expensive for the system.
20:38Instead, there are lots of, call it like a yellow alert or a check engine light in your car.
20:43The person isn't feeling well.
20:45They're not sleeping well.
20:46Their mood is changing.
20:47They're not eating well.
20:48They're in pain.
20:49They're coughing.
20:50And they're not doing anything about it.
20:52They don't want to be a bother.
20:53They don't have the energy to go.
20:55They don't want that.
20:56They'll be okay.
20:57You know, I don't want to bother the system.
20:59And then they need to get hospitalized.
21:02So we raised that early yellow alert before, as the deterioration happens,
21:07directly to their doctor.
21:11So, Joanne, Dor's talked a little bit about the outcomes, the improvement of the outcomes.
21:18I'd love to know from your perspective what the kind of barriers to adoption.
21:21Clearly, if we're getting good outcomes, there are ways of improving patient care.
21:26What's the barrier to adoption in hospitals and care homes?
21:29Is it technical, financial, regulatory, organizational?
21:33What are you seeing?
21:35It's a combination of a couple of things.
21:37So a lot of the things that Dora just talked about actually scares quite a few people,
21:42especially if there's information being collected about them that they don't necessarily have any control over.
21:48So, you know, if you go back to the trust conversation we had at the beginning,
21:52but I find that that's less of a problem now than what we've seen in the past.
21:56Where there's probably real issues is the fact that, unfortunately,
22:01every health care system pays for critical care.
22:05Most health care systems, especially in Europe, is free at the point of care,
22:09which is very beneficial for most of us.
22:11However, I don't know one health care system in the world that pays for prevention.
22:16The only thing in my lifetime that I've seen is the COVID vaccine.
22:20Thank you for that.
22:21But otherwise, it's not paid for.
22:23So if you're thinking about preventative care or helping someone almost not get sick to begin with,
22:29so if you talk about these individuals who are at home before they go into a care home,
22:33before they get sick, well, who's paying for that?
22:36And there's no health care system that is really designed to do that yet.
22:40I say yet because there are some pilot programs,
22:43and the EU COVID recovery and resilience funds that were available several years ago,
22:49they prioritized projects that were quite innovative.
22:52And we've been working with Romania, the Romania health care system, Croatia, Lithuania, Estonia,
22:59very progressive projects to think about incorporating sensors, wearables, robotics as a whole
23:05as a preventative mechanism to health care.
23:09So it's going to take a while before there's this massive shift and we see this being rolled out.
23:14But so it's not really the financial piece per se,
23:17because you always find budget for the things that matter.
23:21And if your health care system is designed to not look at prevention,
23:25you don't find the money for prevention.
23:27But if your health care system is starting to be more preventative,
23:30you find the ways of doing it.
23:32And the fastest return on investment is something to do with robotics, sensors, anything like that.
23:36So we're seeing interest in it.
23:38And I think these couple of pilot projects, especially because they were funded,
23:43and therefore the money was there, which is always helpful.
23:45I think we will see the outcome of that being extremely positive,
23:48and then there'll be things that get rolled out from that.
23:51And then we touched a little bit upon individuals being a bit worried,
23:54like is it going to take my job and stuff like that.
23:57But I don't find that is necessarily an issue as much anymore as it was even five years ago.
24:03People are past the issues of wanting to have to compete with another individual
24:08who's trying to get their job and stuff like that.
24:11No one's as bothered by that anymore.
24:13If it means they can actually do the real work that they want,
24:16as opposed to the administrative issues and things like that.
24:20And to the point that was made previously,
24:22when there's a robot that's standing at reception,
24:24and someone says, oh, are you training it so it takes my job?
24:27No, we're training it so that you don't have to stand up,
24:30show someone where the toilets are,
24:32take someone here and there while there's another queue happening.
24:34It's more about facilitation and making things go faster.
24:38So to me, it's just timing right now.
24:41I don't think there's one particular barrier per se.
24:43I think that the world is starting to come to that consensus point
24:48where this is definitely going to be something that's going to facilitate,
24:51and a lot of people are getting really enthusiastic about it.
24:55Angelina, I'm interested to get your sense of,
24:58clearly we're going to see more and more robots deployed in clinical settings
25:02in the coming years,
25:04and not just clinical settings, industrial settings, whatever.
25:08I'm interested to get, from your perspective,
25:12what is the process of moving from a prototype
25:16to actual live deployment on a ward?
25:20What does that look like?
25:21Can you walk us through those steps?
25:23And can you maybe just give us a specific example
25:26of one sort of situation where you've been surprised
25:29or you've had a learning that maybe you didn't anticipate?
25:33Yes, thank you for the question, Greg.
25:35It is a difficult and lengthy process.
25:39A lot of companies give up, unfortunately,
25:42especially smaller start-ups.
25:44It depends, of course,
25:47how much the technology interacts with the patients,
25:52if it's a service robot
25:55or if it's something that touches the patient,
25:57because that also extends the level of certification.
26:02And then we always go from the needs of the clinical setting,
26:09either patient needs or clinicians' needs,
26:12and then match the technology.
26:15If the technology doesn't exist,
26:16then we search for something innovative,
26:19idea, prototype,
26:22going up the technology readiness level scale,
26:24and finally supporting to implementation.
26:32But sometimes there is a technology transfer.
26:35Sometimes there is already a commercialized product.
26:40There are many barriers we have had,
26:45especially in the past,
26:46some examples where robots have been purposefully sabotaged.
26:51Staff have been putting items in front of them,
26:55so they cannot move around them,
26:58go past around them in their path.
27:00But as Joanne mentioned,
27:02I think the adoption level has really gone far
27:07since some years ago.
27:09because right now we experience
27:12that people are more used to robots.
27:14They are used to seeing them.
27:16They are used to walking around.
27:19Then we have autonomous robots
27:21moving into the hospital
27:22among patients in the hallway.
27:25And they are just,
27:27oh, that's Hubot coming.
27:30And they're seeing the robot's name
27:32as if it's a part of the staff.
27:35So, for example,
27:37in some departments
27:37where we have had
27:38some autonomous robots
27:40moving for the past five years,
27:43they take the robot
27:44as one of their own,
27:45a colleague.
27:48So that's a positive change,
27:50I would say.
27:51And there is still far to go.
27:53So in the future,
27:55there is a lot of other things
27:57to take into consideration.
27:59For example,
27:59I recently read also
28:04about this concept of sim-to-reality,
28:06right,
28:06where the smarter robots get
28:10when you incorporate AI
28:12and this autonomity into robots.
28:15They would really like
28:17that they simulate it
28:19in a simulation environment
28:22and then deploy it
28:23and the robot does
28:24what it's supposed to do.
28:25But in practical settings,
28:28it's not happening right now
28:29yet, unfortunately.
28:33Yeah.
28:33Many of the people
28:34that your robots
28:35are interacting with,
28:36they're isolated,
28:37they might have dementia,
28:39they might be grieving.
28:40So I'd love to get your sense
28:41on what it's like
28:42to design
28:43for really vulnerable people.
28:45How do you think differently
28:47than designing
28:48for someone who's younger
28:50and has full capability
28:51and full faculties?
28:54Yeah.
28:55So the first thing
28:55you need to kind of keep in mind
28:56is that you're not designing
28:58the product for yourself,
29:00which sounds obvious
29:01but it's really hard to do,
29:03right?
29:03Because you look at the product
29:04and you're like,
29:04oh, why is she doing that?
29:05It's annoying.
29:06Well, it might be annoying for you
29:08but it might not be annoying
29:10for somebody that actually,
29:12the fact that she's a little bit slower
29:14or she double verifies everything
29:16or, you know,
29:17so things of that nature.
29:19Firstly, getting understanding
29:20it's not for you
29:21and creating a great panel
29:23of customers
29:24that will give you direct feedback.
29:25When we started the design phase,
29:27we created a panel
29:28of 1,000 older adults
29:29that we create,
29:31we didn't just ask them
29:32would you like this or that,
29:33we created rapid prototypes
29:34using cardboard and foam
29:37and 3D printing
29:38and so on
29:38to get feedback on A versus B,
29:41A versus B,
29:42like we ran hundreds of experiments.
29:44That's the first thing.
29:46The second thing
29:47to keep in mind
29:48is the ethical aspect
29:51and you look at companies
29:54that are doing the LLMs right now,
29:56I heard a statement
29:57from the CFO of OpenAI
29:59ahead of their public listing
30:01that kind of worried me
30:02a little bit,
30:02said, you know,
30:04we're going to be,
30:05think of Meta
30:06and Google put together,
30:08right?
30:08We're going to be
30:09an advertising agent
30:10like to the extreme.
30:11and we know that Meta
30:13and Instagram
30:14and so on
30:15as the father of teenage girls
30:16is not,
30:18ethics in their algorithm
30:20is not the primary concern,
30:22right?
30:22It's getting you to stay connected
30:23and get that another dopamine fix
30:26so they can get more ads.
30:29That's not going to work
30:30for our population.
30:32So you can go on our website
30:33and look at our AI code of ethics.
30:36It's been there for a while
30:38and we keep on evolving it.
30:41Trust is very hard to build
30:44and you can destroy it like that.
30:46And there are lots of decisions
30:48that you make along the way
30:49that might feel like
30:51they're even in the positive,
30:53like for the patient
30:54or the patient is,
30:55or the user is asking for
30:57and we're not doing them anyway
30:59because of perception of risk and trust.
31:02For example,
31:03we're told all the time,
31:04hey, can LEQ order groceries for me?
31:06It's very hard for me
31:07as an 87-year-old
31:08to work at the app
31:10and I can't go to the grocery store.
31:12It's like,
31:12why won't LEQ order groceries for me?
31:15And that sounds great.
31:16Like, why wouldn't she?
31:17And I won't let the team do it
31:19because I'm worried
31:20that once we have the credit card
31:22of the individual
31:23and we make shopping on their behalf,
31:26somebody on my team
31:27eventually will have an incentive
31:29to optimize that.
31:30We'll probably get like a 5% kickback
31:32and all of a sudden
31:33we'll wake up
31:34and the robot will recommend
31:36$2,500 drones
31:37for the grandson
31:38or the granddaughter
31:40because it's a slippery slope
31:42and people work based on incentives, right?
31:44Or what do you do with data?
31:45The healthcare system
31:46pays for our product
31:47and we know the person is sick
31:50and they're getting worse
31:51and they're not willing to get care.
31:54Do we tell their doctor?
31:56Like, that's what they're paying for.
31:58The answer is no.
32:00We will ask for permission
32:02to tell the doctor
32:03every single time.
32:04Not once.
32:05On section 13B, Roman numeral 5,
32:08we may or may not, you know.
32:09No.
32:10Up front, saying,
32:10hey, I'm really worried about you.
32:12Can I tell doctor, you know, whatever?
32:15And if they say no,
32:16we're not going to tell the doctor
32:18regardless of financial incentive
32:20because the person's dignity
32:21and their privacy is more important.
32:23So anyway, there are a million examples,
32:25but those are just two.
32:26So, Joanne, we've heard, you know,
32:28obviously there are enormous benefits to this,
32:30but obviously this all costs.
32:32So I'd love to get your sense of, like,
32:35what the unique economics of this
32:37look like at scale
32:38when healthcare systems
32:39are under such financial pressure.
32:41I think what Dora was saying
32:44is completely accurate
32:45when you're thinking about
32:46rolling something like this out.
32:48And back to my comment
32:50that we're not going to have
32:51massive healthcare systems
32:53rolling out these types of programs tomorrow,
32:55we will continue to see this
32:57in the first instance
32:58where it's paid for privately.
33:00And, you know, as we also heard,
33:02this is sort of the thing that
33:04slowly but surely
33:05people are becoming
33:06a little bit more familiar with it.
33:08So you see the robot that's cleaning
33:10and then all of a sudden
33:11you see the robot
33:12that's doing something else.
33:13So as people are becoming
33:14more familiar with it
33:15and the trust is growing,
33:17people are transacting more and more.
33:18So we're seeing this hugely
33:20in the private market.
33:21But most innovation happens
33:23in the private healthcare sector first.
33:26It almost does the due diligence.
33:27It does all the testing.
33:29And then it's adopted
33:30by public healthcare systems as well.
33:33So it's probably less the
33:35how much is this going to cost
33:37and how do I budget for it?
33:38It's can it be tested first?
33:40Can we see the proof points?
33:42Can we have the papers written
33:43in order to be able to provide
33:45that evidence base
33:46for us to be able to roll it out
33:47into a targeted population
33:49where the benefits
33:50will be realized quickly?
33:51And this will be
33:52probably remote populations,
33:54elderly individuals
33:55in the first instance.
33:56And then once those proof points
33:58are made,
33:59you'll see those
34:01being written into budgets.
34:03All of a sudden,
34:04it's becoming part,
34:05as we heard in the US, for example,
34:08it's actually paid for now
34:09in some instances as well.
34:10So I think we'll start to see those.
34:12Once those proof points are built,
34:14there'll be budgets
34:15and case studies
34:16being able to put forward
34:17to have them funded.
34:18But like anything,
34:19you're going to see it paid for
34:20firstly by individuals
34:22who actually unfortunately
34:24have the financial means
34:25to do so.
34:26But that unfortunately
34:27also happens the most
34:28when it comes to
34:29innovative healthcare technologies,
34:31especially if it's not drugs
34:33or cell and gene therapies,
34:35for example.
34:36Okay, very final question
34:38because we're running out of time.
34:39I'd like each of you
34:40to answer the following question,
34:41starting with you, please, Angelina.
34:43Let's just very quickly look forward.
34:45What does a hospital or care home
34:47with embedded social robotics
34:49look like in 10 years?
34:51How do you think
34:51it's going to play out?
34:54I can imagine that
34:56we'll have a lot of logistics robots,
34:59we'll have robots
35:00in the operation theaters,
35:01we'll have robots
35:03assistants throughout.
35:05So there is a lot of tasks
35:07around the hospital
35:08that can be automated.
35:10Automation is the key
35:12of an aging population
35:14that needs a workflow
35:16to go smoothly.
35:18Cool.
35:20I think we should get comfortable
35:21with the idea
35:22of instead of seeing
35:23an older person in the park
35:24with a foreign aid worker
35:25that might not know
35:27their language or culture,
35:28we'll see them
35:29with their favorite robot,
35:31giving them both emotional support
35:33and physical support.
35:35Joanne.
35:36And I would say that
35:37let's hope we don't have hospitals,
35:38if that's being very bold,
35:40and then thinking more
35:40about the preventative aspect of it,
35:42and also a lot of this happening
35:44in people's homes
35:45as opposed to them being sick
35:47to begin with,
35:48and also a much more efficient system.
35:50Okay, we're out of time.
35:52Thank you all for joining us,
35:53and thank you to the panelists
35:54for such an engaging session.
Comments