Robo Sitters

pretty cool, pu pu. He reported in the surgery went well and he is in recovery doing fine so well. So while he is convalescing, we thought we would do a listener request podcast. So today's topic request comes from our listener Brian in Chattanooga, Tennessee, and Brian wanted to hear about what's the deal with robot childcare? Yeah, I mean, because we've got robots doing all kinds of things. I mean we kind of technically have They can weld car doors, yeah, yeah,

which relates to childcare in many ways. They can they can wage war on humans. I guess it's usually have human controllers involved there. I mean, we already have baby monitors. We technically have limited robots watching your children. Maybe kind of you might call that a robot. Okay, that might sort of fit into a framework that we can put

up and discuss in a bit here. Uh but okay, let's give a little disclaimer before we get into this discussion, which is that from my perspective and I'm someone, I'm not a parent myself, and neither of us our parents, So we should get that out on the table. We don't necessarily speak from experience. Um, but I at least observe that discussions about parenting are fraught with subtle judgments and sometimes not so subtle, so times really really out

there in your face judgments. Yeah. So we just want to stay here at the beginning that we intend none of that whatsoever. We just want to have a discussion about technology, the possibilities, and not make any judgments about the way people parent or the choices they make, uh, raising their kids right, because everyone does it differently. It's certainly not an easy task for anybody. I think that everyone is doing the best that they can and we

would not know. Yeah, and as a robot would not judge, we will not judge. We want to emulate that sort of like cold steely in difference and neutrality that are that our robot overlords have modeled for us. So well, Okay, I think that our first point actually goes with that pretty well. Yeah, I think it's a little bit cold, is it. Yeah. So, as much as we love our children, I do think it's worth saying that for lots of people, many aspects of childcare are not necessarily the most highly

prized activities. And that's not necessarily something I think people should be ashamed to admit, Like, you can totally love your kid and still not be thrilled about changing his or her diapers, right, or having to make them eat food when they don't want to eat food, or take them screaming through various public spaces, right. And in a lot of cases, I'd say that the idea that I really wish I could have some machine to help me do this, it's not even necessarily a question of preference.

It might be something about necessity, like you have limited time resources, right. Sure. A thing that's pretty frequently talked about is the expense of raising kids um in terms of actual monetary dollar values, and as of twelve in the US, from birth to the age seventeen, it's an average of two forty one thousand and eighty dollars um to raise a child nothing and that's that's a kid born in But the time expense involved raising a child

is a lot more difficult to calculate. Um And that's putting it a little bit coldly, but it's definitely a thing to think about. I mean, it's it's certainly a factor in whether or not people decide to have children at all. Right, Well, there are a whole lot of really important decisions that come into play once you've had a child. Um, Like the question about whether one parent will stay home with the child to help raise the child, or whether both parents are going to work. That's a

thing that's a lot of people worry about. Now, that's one of those areas where you see people making judgments at each other and oh, certainly, um, I find that kind of strange, but I do certainly recognize that it's a serious question. Oh yeah, you know, and it's not always feasible for to have any parents stay home in

these are modern economic times. Um As in the US, of the families that included a married a married couple and children under the age of eighteen had both parents working, and of all mothers UM seventy point five percent we're looking for work or we're we're working or looking for work, And in some other countries it's it's different, right well, you know, here here in the US UM, I think it's partially due to the lack of of national paid

family leave UM. American women are guaranteed twelve weeks of unpaid leave, but the US is one of the very few nations in the world that offers no paid maternal leave guaranteed to its citizens. Canada, for example, offers fifty weeks, Pakistan offers twelve weeks, Papua New Guinea and Swaziland are right there with US at zero UM and and some countries offer paternal leave to Iceland and Norway lead that

worldwide with about ten to twelve weeks. Yeah, So if you don't have parental leave and you're in a situation where you need to return to work in order to have enough money to support your family, UM, you're probably going to have to turn to something like daycare. I mean, if you don't have to have a close family member who can provide that kind of support, then absolutely, and it's not always easy to procure that kind of daycare in Japan, for example, some parents have to wait like

two years on a list before getting a spot. It's such a problem that as of fall, the Prime Minister Shinzo Abbe promised to create four hundred thousand more daycare spots. By seen it's one of those politician promises that and it's such a huge number. I'm like, oh, no, this is a this is a problem, y'all. I don't know why I laughed. That's that's not that funny, but it just seems so. It seems literally right right the fact that I mean, you know, it's I think it's that

part of the equation is a lot easier. Here in the United States, we have pretty plentiful child care services, but but they're expensive. I think that I think that I saw it listed. I don't have the fact in front of me, but as as one of the top household expenditures outside of the physical house that you are

living in, yeah, I can believe it. So with all these logistical and economic pressures on parents today, I can really see why somebody would start to think, Man, if only I had a machine that could provide really good care for my child when I really need to be doing something else, you know, that becomes sort of understandable why somebody might earn for that. Um So, okay, So I want to kind of put together a framework for thinking about the different kinds of things that a childcare

robot might do. Um. So, the first one, and what seemed like the easiest thing to me, would be basically entertainment. This is the thing I think a lot of parents used to sort of buy themselves sometime, Like, you know, I really need to wash the dishes, so I can put the child in front of the computer or in front of the TV, or give the child some kind of activity to do, right, so that, for the love of everything Holy, I can go take a shower, right.

Um Yeah, So that one doesn't seem like that much of a stretch, actually, And you can imagine some very rudimentary robots that maybe wouldn't even necessarily meet the definition of a robot. But there's some kind of machine that's at least engaging enough for the child that it really holds the child's attention, and there's um not a lot of risk of the child constantly becoming distracted and wandering

over towards the knife block. Now you might ask in what sense, if you're just talking about entertainment, would a robot be better than just like say, a computer game or a TV show. Yeah, And I guess one advantage is you could say that an interactive robot playmate, if it was sufficiently realistic in mimicking human behavior, that might be and I'm not necessarily sure it would be, but it might be more pro social and more engaging to

the child, more kind of modeling behavior. Sure, And because you know what, one of the one of the rejections of television and computer games and stuff like that, for for especially very young children, is that they're not learned in any kind of socialization skills and that they're kind of you know, rotting their brain I think is the highly scientific term that gets tossed around a lot. Yeah, well, it's passive, and and that they don't learn any skills

on how to interact. Um. Okay, So the first one would be that sort of entertainment thing, But the second, I guess would be the sort of child maintenance tasks like the you are in some sense when you're a parent, you're sort of a child custodian, So you have to change diapers, feed the child, wash the child. I'm sure

there's lots of other things I don't even know about. Um. It's sort of like the physical mechanical activities of taking care of the child to help keep him or her healthy, clean, comfortable, happy, um. And this seems like more of a stretch than the entertainment, but at the same time not necessarily outside the realm of possibility, right, because we do have robotics that are capable of are acting very delicately with things, although we

are just starting to develop those. Really, I think that safety concerns would be at the top of the maybe not so much list for this category exactly. So you can put a lot of trust in a robotic arm that is, you know, welding a car door thousands of times a day. I mean, obviously, car companies that use these invest in them and trust them enough to risk messing up all this equipment if they don't work right. But still you feel like more is at risk if

your your baby is there. And if the diaper changing robot, you know, accidentally, even just nudges your baby too hard, that's something you really really don't want, of course. And also babies, I imagine squiggle around a whole lot more than that car door that's that's being welded. There's probably a lot more variables to worry about in terms of

children than there are in terms of cars. Yeah, and I actually imagine that it's a harder robot programming problem to do that look at touch that is just precisely well something. Yeah. Well, well there's there's all that conversation about how difficult it is to get sensory feedback in a robot. Correct, so that so that you can pick up,

for example, a styrofoam cup without crushing it. It's something that we as humans take for granted because we've got a really complex feedback system in our brains and sensory apparati haptic feedback. Yeah. Um, okay, so those are sort of items one to entertainment, and then child maintenance three is where we start getting into the Wow, this would

be amazing, but I don't know. And this is keeping children safe, right, So that's the that's the keeping them like physically preventing them from running over to that knife block, Yeah, running into the street, crawling onto the stove. I mean, you never realize how many deadly things are in your house until there's a baby there. Uh. Actually, you know, my wife and I had a friend who brought his child over to our house, and suddenly our house, which

normally seems so comfortable, seems so dangerous to me. I was like everything that has a corner on it, and you're just like, oh crap. Yeah. Um. And so taking care of a child and keeping that child safe obviously is something that's a very sensitive task and requires attention.

Like a lot of what this parenting investment probably is a lot of times you're not even necessarily doing anything directly, you're just keeping a really close eye on things to make sure nothing dangerous happens, all right, allowing the child to to explore and play and interacting with it, but but also yeah, making sure it doesn't kill itself. Yeah. Now,

this is where the robotics starts getting really difficult for me. Um, because unless you're in a very controlled environment, say like a safe room of some kind, you would need incredibly advanced AI on a robot that was designed to keep children safe. It would have to have something way beyond any kind of AI that that robots are capable of today, and very advanced motorization as well. It's really difficult at this juncture and robot express to create something that moves

around with that kind of ease. Yeah, so just imagine the scenario is the child is running out into the street in a car is coming. So now you have to combine not only a robot with enough good sensory input to see that the danger is arriving, but good

enough AI to realize that this is a danger. Um the delicate touch we were talking about in the you know, the diaper changing robot, because it would have to be able to scoop the child out of harm's way in a way that wouldn't hurt the child, right, because if you just hit the child with a speeding robot instead of a speeding car, you're not helping the situation. Right.

Once you get to this stage, you're really starting to get into the i think, the kind of sci fi area of robotics where this is way way out there if it's ever possible. So, so, what's the most what's the most ridiculous thing that you've thought of? Okay, yeah, those are levels one through three. I would say that probably the hardest one is the idea of nurture, and that's sort of an abstract concept, but basically I'd say

it's a synthesis of education, encouragement, and emotional bonding. It's that other thing parents do for their kids that's not sort of a mechanical task or involved in physical safety, but it's what parents do when they're having a relationship with their child. And and that's also maybe one of the things that that people pass the most judgment on of of whether or not a human being is ever doing a good enough job teaching their child and loving

their child and paying attention to their child. Oh yeah, I mean, it's this incredibly deep, human nuanced kind of task that I don't know if we even really understand it well enough to codify what humans should be doing. Oh, we don't much less to have a robot do it for us. There's a lot of you know, psycho behavioral kind of studies out there about this sort of thing, and we're we're learning more about how brains develop, but

it's it's all it's all very research based right now. Yeah, and then there's the other question of once you get to that level, you're probably talking about tasks where I'd imagine most parents wouldn't want a robot to replace them in this sort of nurturing category, even if it would free them up some time, Like that's the kind of sort of like the core of being a parent. That's

why they became a parent. Yeah. Um, okay, so we've talked about this framework now, but what's actually out there, Like, are there any robots that are anywhere close to being a childcare robot? Well, there's already some robots for older adult companionship. Um and I and I say companionship because it's it's not caregiving kind of thing really. Um. You know you've got that that peroh the oh yeah, the seal Pepper robot. It is so cute, umunster fuzzy little

electronic monster. Absolutely. Um. And there's others in testing that are that are more humanoid shaped. There's one called the Moby serve um that's being tested in Europe right now. But um, but you know, the tasks involved in childcare and and companionship for the elderly are so incredibly different, you know, but both involve monitoring and sympathetic interactions. But but companion robots are more like safety and remind your

kind of based things. But what they're talking about, there's something that's going to go like oh, Hey, you left the stove on, Please turn that off? Or oh, hey did you take your medications today? Or hey, you seem like you're making that face that I have learned as sad. Would you like to call somebody and talk to them about how sad you are? You know that that kind of stuff, and that's not the kind of an interaction

that you can have with a baby. Some experts think that it might be twenty years before we see reliable senior care robots on the market and the kind of care that that you were talking about in in your list of possible robot tasks there, what's required for for all of those is going to vary widely based on any given child's age, moods, and in particular needs. Yeah, okay,

what about the soft touch. That's something I'm pretty curious about. Uh, the kind of very delicate handling that any robot dealing with a child whose bones are still forming and all that, or even if the fully formed bones, you still probably need to be delicate. I guess right, Um, material science is going to start helping out with this, and we're we are just brushing the surface of that kind of thing.

Researchers are working on soft robots and flexible electronics, you know, things that move more like muscle and skin and don't have hard edges for baby to fall on. You know, it's thinking about how terrified I would be if something like R two D two showed up for my child to play with. That thing is just I mean, it's mostly smooth, but how many surfaces could it smack its head into? That's awful. Um, what I'm saying is that

R two D two would be a terrible babysitter. Um, I'm sorry, or too, I love you, Um, but yeah, it's it's robotic engineering has so traditionally focused on rigid materials that there aren't even really soft material simulation tools available to robotic developers right now. And um, and and even power sources would have to be totally rethought for that kind of thing, because a traditional rotary motor would be really restrictive to something that's trying to be kind

of amorphous and and movable and soft. Oh yeah, well, power is obviously one of the big issues, especially if you want a robot that's gonna follow your child around in the park. And again, like we said, make sure it doesn't go out in the street all right, All of that physical capacity stuff like like we were saying earlier, sensing and moving around sound like they're easy to do, but for robots they're very, very complicated, as we have

talked about before on the show. Yeah. Uh, then kind of kind of following down your list, you've got emotion recognition and simulation of emotion to worry about, um, because you know, in order that would play in in sort of one and four right, like in the like social interaction and in the nurturing. Sure, and all of that is being studied and developed, but it's it's so complex.

You know, input from humans is really nuanced and varied. Um, all of that stuff with with semantics and pattern recognition is so basic right now, and we have so much of it to learn about ourselves before we can teach a robot how to do it. Yeah. Well, I read a couple of studies actually saying that um, a sufficiently you know, convincing humanoid robot can get a child to want to interact with it and sort of ascribe some

human emotions to it and at least be curious about it. Right, it might it might be harder to work back the other way to get the robot to recognize like the emotions of the child and and sort of understand its needs to be totally honest. Whenever, whenever I read stuff about this, I'm like, there are many days when I'm not sure what's going on with other people based on their facial expressions. So therefore, I'm not sure how I

could expect a robot to do the same thing. Yeah, so how far away is just like Rosie the Jetsons, you know, you know, I don't think that we're ever going to have a Rosie. Honestly, I think that that's one of those things like like it would be nice, like the Singularity. That sounds terrific, But you know, I think part of what we need in order to have robot child care is is a basic change in the

way that we think about human robot interaction. You know, if if we have this concept that we're going to have this this robot, this kind of better than human, perfect thing that is more capable than we are, doing tasks for us that we just don't want to do. That's that science fiction, folks, At least it certainly is for the foreseeable future. And I think that a lot of our concepts about all that is really rooted in science fiction, which is a terrific model for for thinking

about the future. Obviously, I mean, I mean we talk about it so much on the show. Oh yeah, I mean when you whenever you talk about interacting with robots, you have to think about Asimov. Oh yeah, and and his and his rules. Those those three rules of robotics. Right, robots can't harm humans, right, number one. Number two is they have to obey humans, right. And then three is

they can't self destruct, right, but they're prioritized like that exactly. Yeah, and then two and three are each based on right the previous rules. Um. But but the thing is is that these these rules were written to be interestingly flawed. I mean that the story that he wrote happens because of the flaws, and so you can't really base reality

on inherently flawed rules. A couple of engineers, David Woods and Robin Murphy, have done some interesting writing about that, and and they say that better rules would focus on a realistic view of human responsibility for for safety and

ethics in robot deployment. It's true, once you get into the idea of a more complex robot interacting with the child, say anything beyond maybe one or two of the list, Like once you're getting into full on interaction, robust interaction that you know is uh, sort of free ranging and

that the child's safety is dependent upon. Right. Once you get into that realm, you're really talking about machine ethics, which is a whole other topic, um, and and one that we could do tons of shows about and probably will in the future. I'm sure that this will come up again, yes, but I want to bring it back to examples because I'm sure you know, uh, like Brian was asking about this, I mean everybody's wondering, like, is anybody trying to sell a childcare robot today? I mean,

is there anything like that? The answer is there some things are sort of like that that are prototypes. Yeah, they're there are related products, But I mean basically the answer is not at all. Um, let's talk about a few of the things that are sort of kind of yeah. Sure.

Back in two thou eight, there was buzz about this Japanese company, Japanese robotics company Team Masks Shopping Helper bought and this this is like a like a shopping daycare kind of thing, and this was a child size like like one point four meter or four ft seven robot that would help mind children in a couple of Japanese department stores. Like it had a little projector in one eye and a camera and the other and it could identify children based on little badges that you could put

on them. But that was for extremely short periods of time, hypothetically under human supervision. It was more of that category one entertainment sort of robot than an actual child care helper. Yeah. Along that line is something that's been experimented with by a Japanese company, an e C. Yeah, and so they've

made a sort of a communication companion robot called PAPERO. Um. Now I think it's not exactly paper oh we're talking about, but uh, I found one thing that was any C had a US patent filed in two thousand five published for quote Childcare Robot and method of Controlling the robot Um. And so ANYC has been experimenting with some prototypes of childcare robots. But these are not robust childcare robots like

run into the street. Uh, it's more like a robot that's sort of a fun interaction kind of thing, right, something in between Um, in between that that one and that four which is such a weird space to play in, you know. I I think that that's part of the problem is is that when people are designing stuff like this, they're really hoping for for what they're actually capable of is not quite one. Yeah, it's it's like, maybe it's a little better, better than a computer game, but it's

not nurture. It's it's absolutely not nurture. Yeah. But but so so this thing, this thing is adorable. Uh yeah, it's a well so in the first before I was even reading about paper oh, I saw this patent on this the childcare Robot, which looks a lot like paperrow. Um. It seems to be based on the same kind of design, and it's like, you know, I love you Nanny Mix Snowman.

It looks like a funny little rolling snowman thing. And then they also recently have come out with a new version that I think it's a stationary version, the Papero petite um. And but that's being build more not as a childcare robot, but as an elder companion right right. Um. It's got you know, sensors to to track stuff, microphones, cameras, um, temperature sensors to make sure that there's no fire going on,

stuff like that. It can apparently find people even in complete darkness, which I think is a really terrifically creepy but I mean, but important sensory apparatus. But as we were saying earlier, all of these features are pretty easy to implement or to see implemented, to imagine being implemented as an elder care robot. But it gets a lot more ethically tricksy when you try to apply that to children. Yeah, yeah,

um So. One thing I do think that's interesting about this, though, is the idea of a communication robot, so it's sort of like keeping track of what's going on visually. Um. This does introduce to me an idea that I think might be much more feasible in the short term than an autonomous childcare robot, which is a childcare telepresence robot.

Oh yeah, that could be That could be cool. So telepresence robots, you know, they're they're already showing up in like weird Silicon Valley offices where you basically you stick an iPad on top of a it looks like a scooter and it just rolls around the office with somebody's face who didn't want to fly in, and you know that they ram into your door to let you know that they want to come in and have a meeting with you. And then they can come up to the

water cooler and hang out and talk their shirt. Yeah, we're not making this up. This is a real phenomenon. I'm not sure if it's all that wide spread, but

some people are really doing this. But I can see it having some value in childcare, like if you want to have at least a visual and auditory presence with your child when you're not there, um, a sort of moving telepresence robot that shows your face on the screen, translates your voice, and can move around with the child might be sort of an upgrade over just like Skype

on a static computer screen. Sure, and I could see it being valuable as a kind of advanced baby monitor, like the kind of thing where if you want to go into your office in another room of the house and get some work done and have your robot hanging out with your baby, um, but you know, but still be able to kind of check over on a on a second screen or something like that and see what the baby's up to and talk to it if you

want to Yeah, yeah, yeah, yeah, I can totally see that. Actually, I can also see if we want to get closer to that sci fi vision, let's imagine that the sort of movement and uh and handling of the robots are all the physical capabilities of the robot really outpaced the artificial intelligence aspect. You could maybe get a telepresent safety robot. Like we were saying, it's going to be really hard

to make a robot that autonomously keeps children out of danger. Um, maybe some types of danger, but you know, it's not going to be fool proof. I can imagine a telepresence robot where okay, well you've really gotten movement kind of down, um, but it just doesn't know what to do in every case. Maybe you are looking through a screen and you're controlling this from a distance. Okay, well, but but at that point, are you're really saving time if you have to spend

all your time at the office. Uh yeah, maybe you can get Grandma who lives a few states away to control the robots. I don't know. It's a good point. And telepresence robots are being thought about for for medical purposes a lot, which I think cross over into both elder care and childcare. Uh in terms of, you know, as as developments in the medical field spur more better robots for for those kind of purposes, then maybe we'll

see some of that tracking over into these fields. Yeah. Um, and okay, so now that we've talked about this, I do kind of want to back up and get a little bit of perspective on this. Um. There was one thing I wanted to reference because I thought this was a good source. It was from a few years back.

It was in two thousand eight. There was an tutorial and science by the roboticist Noel Sharky called the Ethical Frontiers of Robotics, And in this piece, Sharky sort of warns about, hey, let's not necessarily rush into this whole thing with robots taking these social roles like say, entering the battlefield or doing uh, you know, childcare. Yeah. He

he was a little bit harsh about it. Um, there have been some studies in the in this senior care sector about reactions to these companion robots, and they've generally found those interactions to be favorable. But my my favorite quote in this piece by Sharky was, um, they meaning robot companions for the elderly are being touted as a

solution to the contact problem. But these are still toys that do not alleviate elder isolation, even if they may relieve some of the guilt felt by relatives or society in general about this problem. The success of these robots may stem from people being systematically deluded about the real nature of their relationship to these devices. That's interesting but harsh. Of course. Then again, you could also say that we enjoy our relationships with our pets because we're systematically deluded

about what our pets think of us. I think that's kind of the point. If we were small enough, our pets would eat us. Of course they would. I mean, we look delicious. But and and just because they're grinning doesn't mean that they're smiling. It means that they're panting. But but yeah, but at any rate, I mean, we still get a positive interaction out of that, whether or not it's based on delusion. Yeah. He He also warned against um the possible effects of having a child get

too much of its parenting from a robot. One analogy he drew was too studies of the development of young monkeys that actually he referred to what I think he was referring to as a controversial study about what happens when you take away real monkey mothers from the monkey babies and and put in a wooden and wire monkey that that does have that does contain milk, I think, But I think the some of them had milk and

some didn't, and some were covered in cloth and some weren't. Um, this is a controversial experiment, and it's now considered to have been below ethical standards for how you should treat animals in experiments. But um, it did show that monkeys don't develop right if if they don't have the right actual parenting presences. Yeah, the wooden monkey babies were pretty we're we're antisocial and just not not responsive in the

correct ways. Well, the way he puts it is, studies of early development in monkeys have shown the severe social dysfunction occurs in infant animals allowed to develop attachments only to inanimate surrogates. Now I would, on one hand, I can see that. On the other hand, and here we're talking about kind of crude inanimate approximations of a monkey mother shape un puppet, it might be a different thing

entirely if you're talking about a fairly convincing robot. Right, anything that responds to you is obviously going to be more interactive and more socializing than something that apps that has zero response. It's a whole I love you, cold and feeling robot arm kind of kind of thing. Invader some fans, anyone, Okay, Well, in the end, I think we really just don't know. And obviously it's not going

to be ethical to just experiment on children. Yeah, I mean, no one is proposing, well, no one that I've read about is proposing just taking babies and giving them to robots and seeing what happens. Now, well, I mean I'm all for research to see in a very controlled and ethical setting. Obviously, you know what what children can get

out of certain interactions with robots. I guess I mean the human equivalent, Like we shouldn't be like stealing babies away from your parents and and and giving them to crazy robots. Yeah, the goblin king styled robots. None, None of them. I don't know if if a robot saying saying David Bowie to my baby, I would be pretty down for that. That would be a great robot babysitter.

That's maybe this is why I don't have babies. Ka. Well, Brian, we have now talked about robot childcare and your basic answer is interesting question, but um, we're just gonna have to wait and see about that one. Yes, and thank you so much for asking that question. If you have any questions that you would like to ask us, you can find everything that you want at our website all of the ways of contacting us. The website is at

www dot fw thinking dot com. You can also find us on Facebook and Twitter and Google Plus where we are FW thinking. You can furthermore, and this is relatively new exciting technology. You can email us. We are finally receiving email. It only took like a year. Is that FW Thinking at discovery dot com? That is correct, Joe,

Thank you so yeah. We we hope to hear from you about anything that you want to hear about or hey, if you've got a response for us, we're going to start trying to do some some listener letters on air, so so send those in and either way, we will talk to you guys again really soon. For more on this topic and the future of technology, visit forward thinking dot com. Brought to you by Toyota. Let's go places