Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:00):
Brought to you by Toyota. Let's go places. Welcome to
Forward Thinking. Hey there, and welcomes the Forward Thinking, the
podcast that looks at the future and says life and
plastic It's fantastic. I'm Jonathan Strickland and I'm Joe McCormick.
(00:22):
In today, we're gonna be talking about one of the
most fun topics we've ever done in a pretty objective sense,
because we're going to be talking about toys. Yeah. Actually
a very specific and potentially creepy toy. Yes, a toy
that has had its share of controversy or controversy over
(00:44):
the course of its entire existence, although the particular one
we're talking about now is a version that has not
yet come out as of the recording of this podcast. Yes,
so I'm very excited to do some thinking about this
particular toy, even though I've learned from one of the
parodies of this toy that thinking gives you wrinkles about
(01:05):
hot wheels. No, we're going to talk about Barbie, Lauren.
This idea came from you. Where where did you come
across this idea? The interwebs. I was on Twitter, i believe,
and people were posting links to this article about this
this new Barbie that is coming out. It was just
debuted at Toy Fair in New York that was at
(01:26):
the end of February. I think that's a place, by
the way, that we absolutely need to send some how
stuff works people next year. If by some how stuff
works people you mean the three of us, that it
would be the best option. I mean, if I have
to give it up, I would do it because I've
been able to go to E three in the past,
and I feel like I shouldn't hug all the childlike
(01:47):
joy out of the office. But I definitely want to
tend a toy fair at some point. Yeah. Yeah, well,
I mean, especially because they are coming out with all
of these wonderful new technologically advanced to ways, and not
just in terms of like what kind of plastics they use,
although that's a whole other cool thing. Sure, um, but
but this this Barbie doll that just debuted is called
(02:08):
Hello Barbie, and and it kind of Hello World or
yeah sure, well, except it's it's a WiFi enabled Barbie
doll that can have conversations with you, Like if you
talk to this doll, it will talk back. That's nothing
new because when I was a kid. People my age
(02:29):
had dolls where you'd pull a string and it would
say something like who interpreted as burned the house, burned
the house? Yeah, I've seen that movie. Sure, sure well
that that kind of thing has existed for a long time.
There's also the kind of tape deck prerecorded talking dolls,
(02:50):
like Teddy Ruxpin. Sure, the ones where you would have
a cassette. You put the cassette in and then the
bear would move its mouth and in time to the
words being being a generous way of explaining what it
was like time such a such a rigid hug. Yeah,
yeah that yeah, that good. That good hard robot hugs
(03:12):
preparing us for the future. Hug your cassette player. It's
pretty much the same thing, although your cassette player never
looked as funny when your older sister put a Slayer
tape in it, so uhair, Teddy Ruxpan's mouth would not
necessarily move in time with the lyrics, but it was
still pretty funny. Does see Teddy Ruxpan rocking out like that? No? Wait, so, Lauren,
(03:34):
you're talking about something a little bit different than the
pull string or the cassette player or even the microchip
style toys. Like when the Star Wars prequels came out,
all the figurines had microchips that would interact with playsets,
and you know, you could get them to they would
stand on a little stand, you push the button and
they would have quotes from that character from the movie.
(03:55):
But if you got them in playsets and you put
different combinations together, they would have full exchanges, dialogue exchanges.
Simpsons also had a set that was like that. But
this is even more complex than that. Yeah, this is
a concept of having a little speaker in Barbie and
a little microphone in Barbie, and and uh and in
(04:16):
using voice recognition so that Barbie knows what you're talking
about and can respond appropriately. Yeah. So so if you
if you you know, and I imagine the early versions
of this, the early stages, a lot of the the
Barbie dolls will be drawing upon a pretty standard array
(04:37):
of various questions and and jokes and stories that sort
of thing that would prompt various responses that then the
algorithm on the back end, on the cloud side could
analyze and send the response that Barbie would give to
the child. So you might have Barbie ask you what's
(04:58):
your favorite color, or if the child were saying something
about color, maybe then Barbie would ask, what's your favorite color?
Because it recognizes the word color and then ends up
pulling from this. But that's just the beginning, right, I mean,
this is one of those things that the plan Mattel
has for the Barbie doll, and we'll get more into
this in a little bit, is for writers to be
(05:21):
able to build in new things that Barbie can say,
get the voice actress in to record those those words
or phrases or whatever, and so the conversation trees can
get more complex and there can be more of them
as the life of the the program continues, right, Because
it's not you're not talking to a single plastic product.
(05:43):
You're talking to the cloud of responses. Here. There's no
onboard memory or negligible onboard memory for for this toy.
It's all being stored and drawn upon through WiFi in
your house to a cloud of responses. You're talking to
the mother brain. Yes, there's there's a cyber Barbie which
(06:06):
has Skynet Barbie, which has all the potential things that
Barbie can say stored in it. And then your WiFi
network network communicates to that web server, which then serves
up the the appropriate response. Okay, so I have never
personally owned a Barbie doll before, but number one, can
I get one of these now? And number two how
(06:27):
much does it cost? But yet? Yeah, not not yet. Um.
Supposedly it's going to be out in time for the
holiday season. Yes, it's sometime this fall, is what Mattel
has been saying. I don't think they gave a specific
date yet, but they said they're aiming to have it
out by the fall. They're still working on it, so
this is still you know, this is something that is
not yet quite ready for consumer purchase. Uh. And the
(06:51):
suggested retail price according to an article I saw. I
did not see this from the tell however, but according
to an article, the suggested retail price iss. So it's
not cheap, it's it's a it's a fairly expensive toy. Sure, sure,
you know some some toys, A lot of toys with
digital components do range up into that price, though most
(07:13):
you know, non WiFi connected run of the mill Barbies
are something like ten. Presumably that price also builds into
it the idea that this this doll is perpetually going
to be pulling from a service that's going to need
its own maintenance, right, you know that that's an ongoing
cost for Mattel. And as far as I know, again,
(07:36):
I haven't seen anything, but as far as I know,
there's no sort of subscription service, like you don't have
to continue, like like you're you're you don't have to
worry about Mattel holding you hostage and that your child
is going to come to you and say Barbie doesn't
talk anymore. Yeah, that's why I started cracking up the
idea of Barbie being like, oh honey, I'd love to
(07:56):
continue this conversation. Ask one of your parents to get
there at a card. It's just a low monthly fee.
You can do anything. Yeah, that's just the beginning of
no of our more cynical and dark commentary we will
have on this podcast. What I imagined, sorry, was that
(08:16):
your conversation with Barbie will continue after a brief message
from our sponsor. Again, and I'm not I'm not criticizing us.
I think we have to take a very critical eye
to this approach because it's it's not that the approach
itself is necessarily right or wrong, but there are ways
to implement it that are clearly going to be better
(08:38):
than other ways. Right, there are things that we absolutely
must be cognizant of in order for something like this,
which I think the idea of the toy is phenomenal.
I think I think about when I was a kid
and I would play with toys that having this sort
of capability to really kind of work with my imagination
and and to feed off of it with out completely
(09:01):
supplanting it. So it's not like it's not like you're
suddenly a passive audience member watching something, you are interacting
with something. I love this idea, but I also realize
that with this idea comes some concerns that we have
to we have to be aware of, and we have
to make sure are addressed if we want to make
sure this is done in a responsible way. Absolutely, and
(09:23):
I do want to pose it right here at the
outset that that as goofy and kind of cynical is
we're all probably going to get about this. I I, personally,
I don't know about you guys, do think that Mattel
is making a good faith effort to make this a
good product that is not scary or so I agree
or anything like that. Every everything I've read from them
is pretty cool. Yeah. I think I think, uh, if
(09:46):
we take the messages that we've seen from Mattel and
we we give them the benefit of the doubts saying
that this is in fact the direction they want to
go in, they are they are doing this in a
way that is uh much more appropriate than how I've
seen it handled in other recent cases. Then I'll talk
(10:07):
about a little bit later in this episode. Yeah, I
certainly do at least like that. In in the statements
I've seen from Mattel and the people behind this product,
they seem to be talking about how they want Barbie
to be having conversations with the children about like their
feelings and about career aspirations and things, rather than let's
buy makeup so the boys will like us. Yeah. Yeah,
(10:27):
I mean this could definitely be And I think I
think the other cool thing is that it's a It's
a toy that will continue to evolve as far as
its capabilities in communicating because of the way that Mattel
is trying to continue to write for this doll. Uh.
I think it's also cool that that the largely the
child is going to be the one to drive how
(10:50):
the toy interacts with the child, right, the child. The
child's interactions are going to be what guide those kind
of conversations. So if so, if you have insipid consumerist
conversations with your Barbie doll, it's your own fault. I'm
just saying that if that's if, if the child is
interested in that sort of stuff, there's nothing. I don't
see anything wrong with the doll having a conversation about it,
(11:12):
you know, sure, sure, And okay, uh, well, let's let's
get into a little bit here. Be sure we pontificate
too much more about who's who exactly is behind all
of this, because of course it's Mattel, which is the
owner of Barbie. Uh. And they're they're doing this in
affiliation with a company called toy Talk, which has made
(11:32):
some interactive speech recognition involving apps like like app games before. Um,
but this is the first time they've moved into the
toy market. And the CEO is a guy by the
name of or In Jacob, who is the former CTO
of Pixar. So what he thought was, Hey, those toy
story movies, let's make it real. Well, I mean, he
(11:54):
worked with Pixar for like twenty years and and everything
that I've read from him about this product is just
nothing but excitement about, like like the kind of market
research that Mattel has done and saying that the number
one thing that they've always heard from kids, and I
do love this is a thing that I am so
excited about. In all of the press releases about this,
(12:15):
they use kids or boys and girls. It's never specified
that this is a toy for girls. Just exactly why
I say child as well. Yeah, yeah, yeah, no, it
just it makes me so happy anyway, So thank you guys,
if anyone is listening that was responsible for that, Like
you are just making my heart grow three sizes. It's
it's beautiful. Um So yeah. Yeah, they're talking about children
(12:37):
asking for Barbie to be able to talk back to them,
and that this is them fulfilling this this genuine and
interesting desire of their of their customer base, right, and
and doing so in a way that's not limited to
a vocabulary of fifty phrases or less. Right. That's yeah,
that's the kind of stuff we've seen in the past. Yeah,
(12:57):
but Barbie has talked before, Yeah, yeah, she has. Barbie
Apparently Barbie has not really listened before. No, and that's
really the thing. Yeah, to make this an actual conversation
as opposed to you are listening to a series of
very short model hous. So I'm curious how this is
(13:18):
gonna work? What what? How are they going to um
build this sort of like answer like question response or
interaction database. How's that gonna actually function between the cloud
and the doll and the child interacting with it? Okay, well,
I mean we can talk about both the technology and
the the sort of the writing process side of it,
(13:39):
because it's all part and parcels. So just from a
pure technology standpoint, we're talking about a doll that has
some form of WiFi receiver in it that can transmitter. Actually,
because it can transmit and receive transceiver, I guess I
should say, um, so it connects to your WiFi network.
Presumably I would hope it has some form of password
(13:59):
security that s included, because you want that to be
a secure point. You don't want it to be an
entrigue point into your WiFi network. But whatever, so she
is able to connect into the WiFi network, and then
you have a button on the belt buckle of the
Barbie doll that you can press that activates the microphone.
(14:19):
At that point, you can speak to the microphone which
is in Barbie's necklace in this doll, and then it
will send that message to the cloud based server which
will use speech and voice recognition algorithms to analyze that,
and Barbie will respond and will try to respond in
(14:39):
a way that matches whatever was said to Barbie. So
if you said again something that mentioned color, she might
ask you what your favorite color is ye, Or if
you say, uh, you know, I really want to be
a veterinarian when I grow up, then maybe she'll start
talking to you about how to be a veterinarian. Maybe
if you're saying it's time to go to bed now,
she said, would you me to tell you a story?
(15:01):
You know, it could be that sort of stuff and
let's defy mom and stay up rebel against the system.
Uh So, Originally this is going to be based upon
whatever Mattel has kind of accumulated upon the launch of
the product, where they've had a voice actress come in
(15:21):
record the lines for Barbie, and they have this this
range of things to draw upon. What they will also
be able to do is analyze the incoming messages sent
by children who have talked into the microphone and got
sent up to the cloud based server to see the
sort of things that kids are talking about and then
(15:41):
be able to address those things more directly. Have writers
look at that and say, there are a lot of
kids who were talking about this one thing that we
didn't even we didn't even consider when we were first
writing all the potential dialogue here. So let's write some
new things. Get the voice actress back in where record
some stuff. As it turns out, rocks are really popular,
so let's get someone to talk about geology. Yeah, it
(16:04):
could very well be something like that, or maybe there's
some story that ends up having everyone suddenly really interested
in marine biology, and it might be let's let's have
some some interesting stuff around that with exactly, And the
idea being that this in this way, you could have
a continuously evolving toy, one that would be able to
(16:26):
have new interactions and you would not have that case
of hearing the same thing said four thousand times and
your your child gets bored with this toy. You know toy, uh,
so you know it has it has a promise of
continuing and perhaps even you could argue increasing value because
(16:48):
of the different interactions that can happen. So that's the
the idea, and really, ultimately it's not that that complicated,
you know. It's it's a pretty simple system, apart from
the fact that you do have to consider it's a
very sophisticated algorithm to to recognize speech and voice so
that you can actually send appropriate responses back down. But
(17:11):
everything else is you know, fairly simple. A transceiver and
a microphones really and in a battery source, some power
sources all you need for the doll and uh. That
is is kind of the basics for how this works
and how it's supposed to continue to operate over you know,
however long Mattel supports it. I can imagine one of
(17:34):
the reasons Mattel might be looking at things like this,
apart from just you know, admirable ambition and pushing the
boundaries of what toys can do, is that apparently Barbie
sales have been down. One of you all put a
note in the outline about this, and I was reading
about it. The sales have been down recently, and I
can sort of understand why, because people have been kind
of getting real about Barbie for a long time now.
(17:55):
Think about her unrealistic body proportions, and that might not
be a healthy thing to be, you know, pounding into
kids brains that you should somehow be looking like a
neptune in our glass monster, or presenting this kind of
narrowly constrained, old fashioned face of femininity. Um. And I
mean people have sort of rightly been asking the question
(18:17):
for a long time, like where's Barbie the bomb disposal officer?
And where's Barbie the astrophysicist? And it's gotten Yeah, but
that's a real short punt. So yeah, so I'm not
surprised a lot of parents have been less inclined to
push Barbie dolls on their kids these days. Even if
Barbie has been improving lately, there's sort of a lot
(18:37):
of catching up to do in terms of I don't
know what you might call the sort of like progressive
credentials of Barbie. Um. But you can also see for
this reason why they'd want to take Barbie in a
new direction, have Barbie do something that Barbie has never
done before, right, and so so it makes sense to me,
and and frankly, I'm interested to see what a toy
like this could do. I mean, I don't know if
(18:59):
I has been the seventy whatever dollar price tag, but
I was kind of thinking, oh man, I'd almost buy
one of these that just to see what it's capable of.
I don't I don't mean to throw a suggestion into
your heads or anything, but we do work for a
company called how Stuff Works, and this is something worthwhile
of research, I would say, so technically we could probably
(19:23):
see about buying one of these on the company dime.
And you cannot wait to disassemble a Hello Barbie. Also
like we've got extra mix guys, we could have guest
star Barbie. That's true on this podcast. I'm totally in
favor of this. I'm not even joking, like, alright, alright
at any rate? Okay, Well, to talk about the implications
(19:47):
of this, Barbie, I think we could just do a
really quick refresher on sort of the history of speech recognition,
like where are we now in in speech recognition and
how do we get here? Right? Well, if if you
look back at the history the timeline of speech recognition
and voice recognition, a lot of them also include things like, uh,
the representation of voice, like the syntheiz synthesizing voice. I'm
(20:08):
not going to bother with any of that that that
dates all the way back to the nineteen thirties, But
if you're getting into actual speech and voice recognition, in
nineteen sixty nine you had John Pierce of Bell Labs
essentially state that it was going to be decades away
because it was going to require artificial intelligence. As it
turns out, it didn't take that long for at least
the early advances in speech recognition and voice recognition to
(20:30):
to really kind of get get the ball rolling. In
the nineteen seventies, you had a guy named Linny Baum
from from Princeton University who used the hidden Markov modeling
approach to speech and voice recognition. That's largely a pattern
recognition methodology where you start to be able to to
narrow down categories of stuff based upon input uh, and
(20:51):
that ended up being the basis for a lot of
speech and voice recognition over the next few decades. DARPA
got involved in the early seventies and created these SPEE
Each Understanding Research Program, which involved a ton of different
research facilities and universities, had some decent funding for five
years and really got things off the ground. You have
the founding of Speech Works, which is a company that
(21:13):
creates over the telephone automated speech recognition technology. By nine
you started getting dictation speech recognition software, although it was
somewhat limited. You had to speak each word very distinctly
from every other word. You couldn't just speak without pauses,
(21:36):
a problem for check off very likely, UH. But several
different UH entities came out with speech recognition software that year,
including Kurtzwile. We'll talk about Kurtswile a little bit later
as well, or at least I'll mention him. By a
company called Dragon Systems released a continuous speech dictation software,
(21:56):
so you could just speak the way I am now
and it could continue a leading to natural language recognition, yeah,
at least to the point where it could transcribe it,
if not understand it. By the late nine Microsoft really
began to invest heavily in voice and speech recognition companies
like Kurtzwil and in Tropic. By two thousand seven, Sirie
(22:17):
Incorporated was founded UH it would be incorporated into Apple
iPhones by two thousand eleven. Two thousand eight, Google would
introduce Google Voice Search to iPhones. So this was the
point where we were getting into more of the natural
language recognition, not just being able to transcribe, but to
actually parse it and and figure out what a person
(22:37):
was asking for. Uh. And we're still on that journey, right,
We're not at the destination yet. So right now, we've
got voice commands being incorporated into all sorts of technology, smartphones,
smart television's, uh, set top boxes, video game consoles, security
systems in some cases. Uh. And we're still, like I said,
(22:58):
trying to get to that point where natural language is
no longer a barrier, where I can speak to a
machine the way I would speak to a person and
the machine would quote unquote understand what I'm saying and
respond appropriately. So that's kind of where we're at. So
I mean, so really, whatever you think about this Barbie doll,
it is awesome technologically speaking that that this is this
(23:22):
kind of technology that you know that DARPA and Great
Hurts While have worked on is coming to a Barbie doll. Right,
That's it's gonna be something that a child can interact with.
But obviously there is some controversy, oh mega controversy in
the case of this, I mean, and and it's understandable.
I mean, we're talking about a toy that children are
(23:44):
going to interact with regularly, and it's one that can
pick up and transmit voices to a separate server and
get responses back. But beyond that, if writers are going
to write new content based upon the things that children
are saying to this doll, that means people are looking
(24:05):
at what your kids are saying. And it could very
well be a case of is this technological advantage worth
the fact that we are going to have to give
up some privacy in the process, and that amount of
privacy may depend upon the way your kid interacts with
the doll, or whether or not the kid carries the
(24:26):
doll everywhere. And perhaps you maybe you work for the
Department of Defense and you can't have a doll in
the house that could potentially record something you say and
upload it. That seems like an extreme example, but there
are a lot of people who work in very sensitive
jobs who could not have something like this in their
homes and be completely free of fear that it was
(24:49):
going it was never going to come back and haunt them.
Oh sure, well, I mean, and you know, furthermore, this
is a toy being marketed two kids by a giant corporation,
and how much how much can we trust this giant corporation, Like,
are are they genuinely doing this because they want to
provide an awesome thing to kids? Or are they doing it?
(25:11):
Are they doing it to make money? Are they doing
it to hear your child's innermost thoughts? You know? Yeah? Yeah,
the innermost thoughts thing is the one I think is
probably the least likely, simply because I've listened to kids
talk and I don't I couldn't really care less. I'm
definitely being grew up the old man in that case.
But but seriously, I don't. I don't think that's as
(25:33):
big a concern as the fact that I mean, clearly,
children may end up speaking to a doll and say
things that they wouldn't say to perhaps other people, because
it's a doll that they're talking to, not a person,
And that could end up doing things like what if
the kid is is expressing uh an opinion about a
(25:54):
parent in a way that could be misinterpreted. Perhaps everything
in the home is fine, but it could be miss interpreted.
Or let's say that the kid is expressing an opinion
that isn't being misinterpreted. These are some serious ethical questions.
If you've got people who are reviewing this stuff and
they're seeing some statements that are raising eyebrows. What's their
(26:15):
role in that? Do they report that? Yeah? I mean,
is it's hello, Barbie going to be giving testimony at
divorce proceedings. This is a serious question. It's I mean,
it's you know, it's not the intent, you know, I mean,
it could never be the intent because that would bring
Mattel into liability issues that no company wants to get into.
(26:35):
But it's one of those things that you really once
you think beyond the the original scope and purpose of
the toy, you know that there are a lot of problems,
whether it's a privacy issue or ethics and morality however
you want to define it. There are a lot of
questions that this raises. And the thing is that we
all have this technology, or maybe not all of us,
(26:58):
but many of us have this kind of technology and
corporated into a lot of the stuff we have right now.
But we're adults, and it's a little different when you're
an adult and you can make that decision and as
opposed to and presumably you could be cognizant enough to
turn things off if you needed to, as opposed to
being a kid. Yeah, yeah, I'm just picturing, like what
kind of conversation you have to have with your child,
(27:20):
like like, so, honey, this is Barbie. She'll talk back
to you now. Also by she, I mean a team
of script writing adults in California somewhere who are strangers. So, baby,
don't tell Barbie anything like like, how do you you know?
And it's six and eight. Don't press Barbie's belt buckle.
(27:41):
And it's not that kids are incapable of of parsing
that kind of information from you. It's just that I'm
I'm concerned about whether or not parents would be having
those kind of conversations with your kids. And I'm concerned,
you know, I'm inherently squeaked out by the idea of
stranger listening to anyone's speech, Like I find that very creepy. Now, personally,
(28:07):
we don't know to what extent any child's interaction with
this doll would be identifiable to a specific person. We
do know that they are supposed to be customized so
that the doll learns the child's name over time and
can actually say the child's name, But again, we don't
know necessarily that that is associated with a particular I
(28:29):
d I mean, it could be that on the mattel side,
the server side, that every doll essentially is given the
unique identifier. That doesn't that is not directly connected to
an actual human being's personal identity. Sure, sure there's no
profile that's saying like, hey, this is Susie Johnson. Right, yeah,
(28:49):
it could be. But but as we all know, it
does not take very much information for any system of
sufficient depth to identify a person based on just a
few pieces of data, right, Like we have discussed that
if if you give out your your zip code, that alone,
(29:10):
like your zip code and maybe the the uh, the
make and model of your car, can be enough to
identify a specific person. Like that, it gets to this
thing where there doesn't need to be that much data
out there for a person to be directly identifiable. And again,
unless you're monitoring everything the child is saying to the doll,
(29:31):
maybe yeah, there are other ways of course the child
the child might end up saying things that just that
just identify who the kid is based upon the things
that are being said. Um, there are other questions as well,
For example, uh, is the information encrypted? It better be, like,
because if it's not encrypted, then that then someone listening
(29:52):
in could hear this stuff? Yeah, and that's the concern
about anything that's going on over any kind of internet,
but especially WiFi. Yeah, there was one of the examples.
I was going to bring up the Samsung Smart TV.
Did you guys hear about the brujah about that? Yeah?
So smart TV that has voice interactivetivity with it voice
(30:12):
commands and uh, this was one of those things that
the media really ran with and some people object to
the fact that they that Samsung was kind of singled
out in this way. But here's the story. So the
Daily Beast did this look at the policies of Samsung
accord that relate directly to the smart TV, and it
mentioned that the the stuff you said to the television
(30:36):
would get recorded and sent to a third party, which
raised a lot of eyebrows. They're like, well, I don't
want I don't want to television spying on me and
sending this stuff to a third party that's going to
end up using it for marketing, research or whatever, you know,
whatever purposes. And Samsung I was saying, no, no, no.
The third party is there to work on the voice
(30:57):
recognition to make sure that the system gets more robust
and more usable over time. And uh, and this is
a system that you initiate by using a specific phrase
in this case would be high TV and then whatever
command you wanted to say, just as if you were
using an Xbox one and you wanted to use the connect,
you would say xbox and then the command. So it's
(31:20):
an activation phrase. It's not recording everything all the time,
always and sending it off someplace. That was more or less.
The story of the media kind of went with was
that you essentially have an open mic to the Internet.
But one of the things that Samsung really didn't do
very well was that they had said the communication was
(31:41):
encrypted so that when you use your television, added an
encryption layer so that anyone who intercepted that message would
not be able to see what it actually said. Turns
out it was unencrypted. It was unencrypted, and then it
was sending a transcript of what you said Matt to
you or TV, and the transcript was also unencrypted, So
(32:03):
two chances for someone listening in to see what you
had said. Now, granted, you were probably saying something like
volume up or volume down, but there's the potential that
you could have also had another conversation going on that
got picked up or other elements that could have been
picked up, and it raised the question of well, what
if someone were to get access to the hardware remotely
(32:23):
or otherwise and activated so that the MIC is always
on and unencrypted, then you just have a bug. You
You have got an electronic bug that's also an appliance
in your house share And at that point, you know,
like like pick your paranoia poison. You know, are you
worried about a hacker getting into your system and uh,
(32:44):
I don't know figuring out when you're home and when
you're not home and when it's safe for someone to
go break in. Are you worried about the n s
A listening to what you say and you know, the
New American Gestapo eventually coming and busting down your door? Right? Yeah.
The idea being that even if no one is actively
listening to you, we know that there are organizations like
(33:06):
the n s A that are fully capable of building
out this enormous database of recorded material and then being
able to search for it again using things like speech
and voice recognition and transcription services where using a search
algorithm they can pull up stuff just using keywords. So
even if you're saying, well, no one is directly monitoring
(33:26):
what we're talking about. That stuff is being put in
a is potentially being put in a place where it
could be reviewed at a later date, and it's still there.
And whether you feel like, well, I don't do anything wrong,
that's not the point. It's a privacy issue, not necessarily
you know, someone doing something right or wrong. Yeah, I've
(33:46):
said this before on the podcast, I believe. But anybody
who ever makes the argument that well, if you're not
doing anything wrong, you've got nothing to hide, right, just
say okay, can I read all your email? Yeah? I
mean it's it's ridiculous. You know. Anybody who says that
just hasn't thought about the issue. I mean, you just
(34:06):
people want certain things private. Yeah, And there's nothing wrong
with that necessarily. So the you know, there are a
lot of a lot of potential concerns with this this doll.
Now Mattel's part has they they've said that the things
that are going to be recorded are meant specifically for
building out that database of of potential things that Barbie
(34:28):
could talk about to improve the doll, to improve the doll,
to improve speech and voice recognition as well, which is
an element of AI, so that is also an admirable
thing to do. And they've said that they will not
use it for marketing purposes. Now I don't know if
that means they won't. They might still use it for
marketing research like a focus group. So for example, of
(34:50):
kids say that, you know, they talked to Barbie and
say that they really wish they had such and such
or whatever, you know, like some idea, whatever access or
or place to place exactly, like something that's something specific
than then maybe Mattel would say, we should do another
run of that, you know, or we should build out
this thing and market it because that's a good idea.
(35:12):
That's why don't we have geologist Barbie. Let's do it. Yeah.
Now that's a lot different from having Barbie say to
the kid, uh, you know you like shoes, you should
really look at this shoe accessory set that's available in
stores right now. That would be totally different. Were like, oh, man,
I really want a pony. I'm the saddest Barbie ever
(35:35):
because I don't have a pony, but if only you
got you know, my pony Princess from Mattel or well,
I mean the other thing would be like, Hey, it's
good to see you again. Barbie's got a question for you,
what features do your parents really want in a luxury
city that that would also be inappropriate. And Mattel, for
(35:56):
its part, is said that that's not gonna happen. So
and and like we said, Mantel has been very good
about responding to this sort of thing. It's clear that
they have paid attention to some of the other stories
that have come out over the past and and and
I don't think necessarily that they need to abandon this idea,
although there are petitions for that, right. Yeah, there's there's
(36:17):
a couple of petitions out there that have a few
thousand signatures each. I I don't think that I personally,
Although I I approve everyone's right to petition for things
that they like or don't like, I don't think that
that is going to be enough to stop production on
the toy or anything. And I think ultimately the best
way of not uh supporting it is not to buy
(36:38):
the toys, you know. I mean, maybe if you're afraid
that people you meet on the street are carrying a
Hello Barbie surreptitiously on their person and pressing the belt
buckle repeatedly. In that case, you might need to get
new friends. I mean, it seems like boycotts or boycotts
and petitions for banning things are rarely effective for new
(37:00):
like new products and new works of art, where they
typically just bring more publicity. What I think what I
think it might do, however, is really convinced Mattel that
they need to make sure there are they are doing
their due diligence in presenting this in the most responsible way,
so making sure things like that encryption is in place,
making sure that there is the capability to password protect
(37:22):
any sort of WiFi connectivity so that Barbie does not
become an entry point to hackers getting access to your
WiFi network. Because we we talked about this on I
think an episode of tech Stuff maybe, But there was
the story about the baby monitors that had here. It
wasn't about where where there were internet connected baby monitors
(37:43):
that got hacked by somebody who has started yelling at
babies to tell them to wake up well and and
hacked in the sense that if if it doesn't have
password protection, or it has password protection but there is
a default password that no one ever bothers to change,
then you don't. There's not really any hacking necessary. It's
(38:03):
essentially an open connection for you to to access. Uh. So,
so if you know it needs to have something that
would prevent that from happening, you don't. Again, you don't
want someone to hack into the toy and have Barbie
say inappropriate things to your kid. That would be that
would be all I would be upset. You know. I
have another question that may or may not be interesting.
(38:25):
I'll see what you'll think before we get into the
reasons that I think that this is actually really cool
and could lead to good advances. Uh. The question would be,
is this necessarily a good thing in how children's speech develops?
I mean, I don't know if we've ever had a
situation before in history where children are going to be
(38:48):
interacting back and forth continuously with a with a truly
reactive electronic agent, And I wonder how much that would
affect the way a child learns to speak. If this
is a toy that you're giving to a very young
child who's you know, still developing their own speech patterns,
do you know what I mean? Yeah? Yeah, Well, there
are a lot of um talking robotic books and apps
(39:14):
and computer programs that for certainly the past decade, maybe
maybe fifteen or twenty years, even children have been using
to learn to read or to uh increase the vocabulary
stuff like that, sure, or even just you know, in
playing games. That's something that I think that more kids
(39:34):
of this up and coming generation are more used to.
So I haven't personally read research about it, although you know,
I would say that since the things that Barbie says
are completely controlled by the the by Mattel, they can
make certain that Barbie is speaking properly, is using grammar properly.
You could even have Barbie be able to answer questions
(39:56):
about that sort of thing if that ended up being
uh of interest. Now, I doubt that we would ever
get to a point where we have Barbie correcting people's grammar,
because I think that would I think that would just
be obnoxious. Barbie is like, ain't ain't a word stupid?
Or or I want that toy. They're right, like if
(40:20):
the if if if the if the child says I
did good on my test, and just know you did
well on your test, which is odd because you don't
know how to say well versus good. That that would
have been you know, yeah, yeah, but I think that
you know the fact that you could you The nice
thing is that Barbie is not just randomly constructing sentences
(40:41):
from a huge group of potential words where she could
start speaking like Yoda, She's going to be speaking in
a way yeah scripted. I think you make a good point, Lauren.
I I maybe I was overestimating the degree to which
this is different than things that came before. I'm just
thinking because because it seems to be offering something that's
(41:02):
a sort of heightened level of interactivity that actually learns
from the child. But yeah, children already do interact with
a lot of like sort of speech driven electronic media
things that talk at them, just maybe not quite in
the same way as we're talking here, So maybe maybe
it wouldn't make that big a difference. It also it
kind of I mean, I understand why some people are
(41:23):
finding this sort of creepy because it reminds me of
what my initial reaction was when Google Now came out.
And Google Now is like I I get these little
little notifications on my phone, and a lot of them
are prompted by my browsing habits using Google Chrome, and
all of this is because I have, you know, my
(41:45):
Android phone, connects through a Google account, so it's all
in that same sort of playhouse. And if I pull
up those those things, if I've done something like searched
for flights to a different city, uh, then it might
give me information about that city, like things that are
happy thing at the time when I was doing the search, saying, Oh,
this weekend at that city you were interested in? Is
(42:05):
this thing? So if you're gonna go there, you might
want to check this out. At first, it starts seeming
creepy because it's like, how does this thing know so much?
And then you realize, oh, it's just taking the information
that I'm using in all of my activities and kind
of cohalating it and curating it in an easy, easily
digestible way. So it seems creepy, you know, in you know,
a first glance, but then it just becomes really useful.
(42:28):
I could see the same sort of reaction with this,
the idea being that it's a doll that gradually gets
to know more and more about the child and how
to interact with that child based upon the things that
the child seems interested in. Yeah. So, on the other hand,
I think we should look at some of the cool
things that this could actually do for the future, because
(42:49):
it seems like a toy like this is more advantageous
than just that it's fun for kids, Like we could
actually learn things from it. Sure, for example, I'm in
how to keep developing a speech recognition for one thing. Yeah,
this is one of those things that we've seen lots
of development over the last several decades, but it's kind
(43:09):
of slowed down, Like this is one of the reasons
actually why I think it's amazing that kurtzwil is a
futurist who believes in the singularity, because he has this
very strong belief that singularity is just a couple of
decades away, very very close now. Uh. And yet he
also worked heavily in speech and voice recognition, where it's
(43:31):
clear that technological progression does not does not go at
the same rate for every kind of technology. Oh yeah, yeah,
And and and things tend to plateau during certain cultural or
economic or technology technological challenges. You'll have something happening you're
just like, oh, well, we kind of hit a roadblock.
(43:53):
And it's not that we're never going to get around it,
it's just it slowed us down. Right, We're not We're
not progressing at that same uh you know geouh geometric
rate that we were earlier and uh so, but it
does mean that we can continue to improve on that,
which is part of artificial intelligence. So uh it means
(44:13):
that it's helping push the discipline of artificial intelligence further,
which is exciting. Yeah, the mad scientist in me thinks,
like man with proper consent, you could use this to
learn fascinating things about the development of language and humans.
But then again, I'm sure there are tons of questions
about how it would be difficult to wrangle that kind
(44:33):
of experimentation ethically well, especially like you. You'd also have
to factor in so many other different variables, things like
the region that you're talking about the Now. Granted, you
also have to keep in mind this is a very
expensive toy and it also requires that you have that
WiFi connection. Now. Granted, because the toy is so expensive,
it's probably pretty safe to bet that the people who
(44:55):
are buying it are also people who have WiFi. But
it means it means that if you were doing any
kind of social study on that, you're looking at a
particular democratic Yeah, if you were trying to base it
on who who just bought them naturally, as opposed to
like where you you know, gave them to control. Yeah, yeah,
(45:17):
if you went into the wild and yeah, so that
they could actually use it, because I mean, well, yeah,
I mean you know why, I guess you're for research.
You brought kids in then sure, but I think you
could do that in a way that would be that
would not require, uh, the the actual doll, Like in
that sense you would you could just set up a
(45:40):
camera and microphone in a in a room and have
regular toys for the kid to play with. Because yeah,
I guess you could have the interactive doll and that
would help you just listen to the kid's responses. But
you don't need the interactive doll, Like there's no you
wouldn't be looking at the back end at all. For
the Mainly, I'm just saying we need to learn more
(46:01):
about the evil heart of man from what children say
to their dolls, because a lot of what they say
to their dolls is quite disturbing. Also, really, I don't
remember ever having disturbing conversations with my action. This is
the thing. People abuse their dolls, don't they, especially especially Yeah,
well that's a thing I encountered when I was a child.
(46:22):
I remember this anyway one feeling bad now. I never
abused my action figure. I went on adventures. But so
back to happy Land. I want to see hello Barbie
based art tech projects. I had an idea just earlier
as we were talking. So you've seen um Twitter bots
(46:46):
play spam tennis right where, Like we talked about that
in a podcast not too long ago, the one about
what happens when a robot breaks the law? You can
get bots talking to each other. They're they're designed to
respond when people talk to them. What have If a
bot talks to a bot, they just keep talking back
and forth, garbage. What happens if you have Barbie talk
(47:07):
to Barbie? I wonder, I wonder if she if she's
good enough to to hear that it's another Barbie talking
and to be like, hey, fellow self, how are you doing? Oh?
I have great, fellow self. I don't know like that
or or what I imagine is they could actually do
the thing that Doc Brown talks about Back to the
(47:28):
Future too, where if you see yourself from the future,
you could destroy the entire universe. Yeah, yeah, I just
imagine a g I Joe place that just starting to
smolder in the background. It's just affecting like toys. Um.
I I suspect that while we I don't know exactly
who initiates conversations using this toy, right I don't, I
(47:52):
assume that it's the child. So I assume that you
hold down the belt buckle that activates the microphone and
the child ends up saying something or the owner whomever
says something into the microphone that prompts the beginning. So uh,
you would first have to be able to time that
out just right right, so that you get the first
(48:14):
response from Barbie number one as you're holding down the
belt buckle for Barbie number two, so that the response
from Barbie one goes into the microphone of Barbie two.
Then you'd have to time it out so that you
could do the same thing with Barbie number two starts
speaking to Barbie water. Somebody can come up with like
some kind of little large we know setup that that
has them go back and forth. Use your fingers too,
(48:34):
I mean it, I'm just thinking you hear about the
TVs that you want you have to be able to
say things to them rather than pressing buttons with your fingers. Look,
I love a Rube Goldberg device as much as the
next guy. I'm just saying it's not always necessary to
do something simple, but uh, I mean I would also
be interested to see what would happen, Like how long
(48:54):
would that exchange go before it would break down? Yeah? Yeah?
And just how how can Vin sing? Is this actually
going to be to to a kid? Like? Yeah? And
you know, because the age range for for toys like
this is pretty enormous in terms of childhood. I mean,
you know, at anything from you know, like older than
the age of might eat Barbie's hand and die up
(49:17):
through up through puberty. You know, at least is within
the marketing range of this kind of toy. So is
it going to be the thing like when we were
kids and talking toys came out and we were just like,
this is the lamest thing I've ever seen in my life.
I went on adventures with maybe again, you sound like
(49:41):
you were much better adjusted than I was, Jonathan, But
I don't think. I think this is a pretty cool idea,
and I'm I'm excited to see where speech recognizing toys
go in the future. It is way cooler than what
they had when I was a kid, which was just
like you know the polsterring talking toys, Well, there's a
snake in my boots, or the I think they had
(50:05):
some diaper wedding baby dolls when I was a kid. Yeah. Well,
and and perennially talking about career aspirations is just cooler
than wedding diapers. Well, and this technology, this this ability
to UH for for speech and voice recognition and responses
and using voice control as a user interface, is something
we're going to continue to see creep into more and
(50:27):
more technology around us. And I think the objections to
Hello Barbie do bring up some points that we need
to be aware of just moving forward in general, especially
like we've talked about the Internet of Things a lot
on this show, and I love the idea of the
Internet of things, but even I as someone who is
truly enthusiastic about that technology, I recognize that with the
(50:50):
wrong implementation or the wrong level of awareness, it could
end up being problematic at best because the fact that
there are some really serious privacy concerns and some security
issues that you have to think about make sure you
design everything properly. There was a recent story about people
hacking into WiFi networks using a nest thermostat as an
(51:15):
entry point. So that's the sort of stuff that you
have to be aware of. If you're going to have
this interconnectivity with all these different devices, you need to
make certain that you keep security at a level that
is going to be manageable but also effective enough to
discourage most attacks. Now, keep in mind, most of us
are pretty safe from the the the leite hackers out there.
(51:38):
It's it's the folks that have access to tools but
are not necessarily uh willing to go and put forth
a lot of effort to break into a security system
that we need to really be cognizant of most of us. Obviously,
if you hold an important position somewhere or you are
in the public eye, then it may be different for you,
(51:59):
But for the most of us, we just have to
worry about that lower level of security. But still means
that we need to keep that in mind, and I
think that's just gonna be something that we have to
look for, both for speech and voice recognition type of
of technologies and just general technologies that gather information about
the area and send it off somewhere in order for
there to be a reaction in your environment. Because that's
(52:22):
always going to pose a potential privacy issue, and if
we're aware of it, we can implement this technology properly
in a way that benefits us without putting us to
great risk. At least that's that's my kind of opinion
about it. I know some people are a little more
more cautious than I am, and I don't blame them
one bit in that case, Uh, put put electrical tape
(52:47):
over your webcam viewers, and don't buy your kid at
Barbie Dollar. I know, I know people people who are
just as enthusiastic about technology as I am, who regularly
put a post it note over a web am. I
know folks who are you know, they're not, and they're
not paranoid that they just they recognize the potential that
(53:07):
this technology has means of both benefiting us and possibly
being a security or privacy issue, and and acknowledging that
is the responsible thing to do. Acknowledging it and making
sure that we are as responsible as possible moving forward
means that we can ensure that we have the cool,
awesome future and not the creepy orwellian future. Joe, what
(53:32):
do you think you've been kind of quiet? I'm just listening.
You're just recording everything for the future, despite the fact
that we've literally recorded this and you can literally go
back and listen to it whatever you want. Well, yeah, later,
I'm going to analyze the transcript and then figure out
how to sell you a car that will be a
heck of us doll. Doll, you have a much better
(53:57):
shot at Jonathan. I know you don't dry, but hey,
if you take a bunch of dolls together, you can
use them as a sled to get down the street.
I used to go on sledding adventures with my action figures.
Well that about wraps up this episode forward thinking, Do
you guys have any other closing thoughts you want before
(54:17):
we we say goodbye to our beloved listeners. I think
I'm more creeped out by a sled made of barbies
than I am by like random strangers listening to things
that my hypothetical children say into a microphone. So good job, Joe,
you found something that totally creeps me out. Maybe I'd
be more creeped out about this if I had kids.
I don't know. It just seems pretty interesting to me. Yeah,
(54:40):
and that's to also be fair, none of us in
this room have kids, So we're we're also coming from
this from the standpoint of people who are interested in
technology but also aware of the potential pitfalls. But we
don't have a personal steak in this efficient paranoia. Yeah.
But but so if you, guys, if if you are
(55:00):
in the market, either your yourself, if you are in
this this demographic that is interested in owning this doll
or or if you do have children and have an
opinion about all of this, let us know. How can
they do that? Jonathan, Well, they can write to us,
And I would recommend writing to us. If you're going
to express an opinion that's going to take more than say,
a hundred forty characters, send us an email. That email
(55:21):
addresses f W Thinking at how Stuff Works dot com.
Or you can drop us a line on Twitter, on
Google Plus. On Facebook and Twitter and Google Plus, we
are f W thinking. Go to Facebook search FW thinking
will pop right up. You can leave us a message
and uh, if you have suggestions for future topics, definitely
let us know. We've got a list of those that
we're going through. And we love hearing from you guys,
(55:44):
and and your ideas are fantastic. Often you suggest things
that just hadn't occurred to us, and they end up
being some of our strongest and most interesting episodes. So
please continue to let us know what you think and
we will talk to you again really soon. For more
on this topic and the future of technology, visit forward
(56:05):
Thinking dot Com Problem brought to you by Toyota. Let's
Go Places,