Microsoft’s Voice and AI Evangelist Noelle LaCharite Talks About Transforming Healthcare With Applied AI Coding

Nicole LaCharite

[ Music ]

Margaret Flin ter : Welcome to Conversations on Health Care with Mark Masselli and
Margaret Flinter , a show where we speak to the top thought leaders
in health policy, health innovation , and technology , and top thought
leaders who are shaping the hea lth care of the future. This week ,
Mark and Margaret speak with Noelle LaCharite , principal program
manager for applied AI and cognitive services at Microsoft, helping
developers apply their artificial intelligence technology to develop
apps that will adva nce their business and industry. She's also one of
the original voice skills developers for Amazon's Alexa.

Lori Robertson also checks in, the managing editor of FactCheck.org ,
looks at misstatements spoken about health policy in the public
domain , separa ting the fake from the facts .

We end with a bright idea that ' s improving health an d well - being in
everyday lives. If you have comments , please email us at
[email protected] or find us on Facebook or Twitter @chcrad io . We
love hearing from you . Or you can find us on iTunes, SoundCloud , or
ask Alexa to play the program , Conversations on Health Care.

Now , stay tuned for our interview with Noelle LaCharite , principal
program manager for applied artificial intelligence at Microsoft here
on Conversations on Health Care.

[Music]

Mark Masselli: We're speaking today with Noelle LaCharite , principal program
manager for applied AI and cognitive services at Microsoft. She has
spent many years as a coder , developer, system archi tect, an
evangelist for IBM, Red Hat , EMC , and more recently a lead evangelist
and solutions architect at Amazon , where she was one of the first
developers of skills for Alexa , Amazon's voice technology platform.
She's also founder of several of her own en terprises , including Lady
Coders podcast, also Voice Skills Inc. , where she and other developers
have created thousands of skills for Alexa .

Noel le , welcome to Conversations on Health Care.

Noelle LaCharite: Thank you so much . I ' m very excited to be here.

Mark Masselli: As you were thinking about careers , you were early on thinking that
you'd like to be an astronaut , but i t seems like you've forg ed your own
frontier as a coder, developer, systems architect in the computer
space. I think the public saw this dramatic shift from the mainframes
to personal computers. We saw the launch of the iPhone , which really
started a huge revolution in the smartphone space . You say you're
poised again for another dramatic shift with voice as the next major
disrupter in the industry.

Nicole LaCharite

I wonder if you could just talk to our listeners about the current
emergence of voice technology and applied AI in the computing world
and how do you think these technologies will help transform the
user's experience ?

Noelle LaCharite: I have m ultiple aged children in my house . One of the things that I see
them doing today more than I've ever , just clicking on the microphone
button and being able to say the word , bus , and the Wheels on the
Bus come up on like YouTube Kids . I have this vision and really like
proof in the pudding right now of what our life is going to be like 30
years from now. When my kids are working, they're going to have the
same expectation that they're going to be able to click a button or
better yet just be able to talk , and the wo rld's going to mold around
them.

When I was six years old , I was reading like Asimov and Bradbury , and
The Martian C hronicles , and Star Trek, and The Jetsons . It was so
interesting to me to look back now at things . I don't even know at that
time if I thought that was impossible. I now realize that our
technology has caught up to our vision and our dream. Literally today ,
if you have a dream in your heart to build something , it's more than
likely that the technology is there to help you build it . Tha t's why I'm
super - passionate about this space.

Margaret Flinter: I've heard you describe Noelle as an evangelist for artificial
intelligence and voice technology . You've led seminars and
hackathons all over the world , teaching people how to develop voice
r ecognition skills to meld with Alexa and other digital platforms.
You've taken all that collective experience to Microsoft. Share with
our listeners how the marriage of voice tech along with applied AI
technologies are changing the game .

Noelle LaCharite: First , like applied AI is the ability for us to use models that were built
by someone else . The benefit of that is that companies like Microsoft ,
Google , Amazon , they've been building machine learning models for a
very long time . The ability for us now as developers to leverage those
models , and Microsoft , the reason I went there , is , A, of course they
have this incredible social good philosophy , which I'm very passionate
about , but also because they have this elaborate selection of prebuilt
models to choos e from.

I started in voice , but even more maybe fundamental things like what
if I could translate my text face app into 60 languages without actually
needing a translation team or a machine learning person , w hat if I
could analyze all the texts coming in from every comment someone
sent on my blog and determine if people are mostly happier or mostly
sad , or maybe I could do text analytics on feedback that I got on social
media and make sure that I guide my podcast down that way.

Nicole LaCharite

These are all things that e ven five years ago I couldn't easily do as a
developer. I really think like anyone can learn to build applications
today . Now , you can actually, as a novice, build elaborate
applications , thing s that can translate and speak , all without really
having to kn ow the underlying technology behind it.

Mark Masselli: There 's some real aspirations for those of us who think about AI . We
saw IBM Watson and others with the promise in this health care
space to really make big changes , but w e seem to have fallen short.
W e're sort of at this infancy stage , it still seems . I ' m wondering if you
could just help us understand the difficulties of applying artificial
intelligence . When I think about that in health care space , I'm thinking
about the critical thinking. You probabl y read too much Isaac Asimov
and these entities have this malicious capability of doing bad things to
us . Give us a sort of the bigger picture.

Noelle LaCharite: My dad would always remind me , even today when he sees the work I
do, he's like, these storie s don't end well in most cases.

Mark Masselli: That ' s right.

Margaret Flinter: Thanks , dad.

Noelle LaCharite: Well, it's funny because he actually encouraged me all the skills that I
built for Alexa are all like mindfulness and kindness based. I was like,
a t least the AI will be nice choice [ PH 00:06:23] about it . One of the
things that I think is really important that the AI we are building, our
goal is not actually for it to do the critical thinking. We actually think
humans are much better at t hat than ma chines will ever be . In my
early days of building weak AIs to m y dad and to my children, they
were like, this is amazing . I can't believe you built this voice skill that
does this.

The reality was is I wrote 200 lines of code that did that work. It was
li terally artificial ly intelligen t because to the user , it seems like
magical . The reality was is that me as an engineer , I wrote all those
lines of code. Being part of Alexa early on , I had that same realization
that artificial intelligence is literally art ificial because there are people
like me with a computer , writing every decision that these computers
will make.

A lot of it is realizing like we are building that , like literally training it
on how to do that work. We call this supervised learning. I thi nk the
more interesting side is the unsupervised side , right, like h ow do we
get to that point where machines are doing more on their own
because there's a recent article, how do we make artificial intelligence
less artificial and more intelligent. Now tha t we're building software
that will evolve and potentially get to the point where it evolves on its
own , you have to be much more deliberate about the things we

Nicole LaCharite

enable , the process es that we train into these models because
unfortunately the world is biased .

It's one of the saddest things. I go into Google and I type in CEO , and
someone like me, a Latin American woman , I'm sure that there are
other colored women that are CEOs , but I have to go to page 3 to find
one. When we think CEO , we think white male an d that's who shows
up for the first three pages. Companies like Microsoft are taking very
specific and deliberate actions to unbias the world's data and make
sure that our models are consciously considering this .

What that means is that we have to do extr a work. It ' s not just letting
the world learn on some because if we let it do that , it would actually
be wrong, which we found already if you do some research on facial
recognition . We realize d , oh , my gosh . The w orld actually thinks
everybody looks a cert ain way. We have to make deliberate efforts of
going around the world and going to places that do not get
photographed often and making sure that the world's data includes
these perspectives.

Margaret Flinter: Well, it reminds me of the work we're doing wi th the n ational All of
Us project to make sure that when we're thinking about precision
medicine or genomics, that we ' re getting the whole world in there.

You've led developer competitions and hackathons around the globe ,
addressing certainly some of the biggest health care challenges like
chronic disease management. You recently presided over a
competition , looking at how to improve the management of diabetes
using voice technology as part of the solution. I'd love for you with
your vision to expand upon the role of this in managing complex
chronic illness and maybe adding on health consequences of
loneliness and lack of social engagement. How are you marrying these
two things together and thinking about the outcomes they might
produce?

Noelle LaCharite: Y eah. This is one of the areas that I'm so passionate about , o ne ,
because we think about using the world's data to [inaudible 00:09:32]
just very dramatic large scale challenges that we have . The reality is
that there is so much that we can do, especially w ith the kind of
applied AI technology that we could do to actually impact people
today right now immediately.

One of those is voice technology . Oftentimes , getting real - time data is
difficult , and so b eing able to provide a conversational interface. The
n ice thing is you could use Alexa and ask questions kind of on a daily
basis to make better choices. Then imagine if you don't need an Alexa .
This could be on a phone app . That phone app could send a text
message twice a day that just says, hey, want to hav e a quick
conversation ? Click this button and talk to me . We can literally enable

Nicole LaCharite

chat on any device.

The combination of Alexa and Cortana allows us now to use Alexa on
any device that has Windows 10. Now you think about like if you're
even a doctor and y ou want to be able to do certain types of voice
enablement . Not everything is a super - tough problem. I will tell you
one of my sad stories . I go into a dentist office . I have four children
moving into a new city and t hey literal ly hand me four stacks of pa per.
Now I was like, can I please buy you an i Pad? How can I help you?

Today with things like chat bots, you could get my information . I could
be sitting at home , or I could even fill it out at home , and you take a
picture of it . You could use OCR to pull that into your system . You
don't need any special data science skills to do this. This is just
standard web application develop ment that you're already doing.

I want to make sure that everyone in health care realizes that , yes,
there are big problems we'r e trying to solve, but there's also so many
kind of things that we could do to change the lives of people like r ight
now . The value of having voice involved creates this kind of level of
comfort with our end users and with our patients.

I'm very passionat e about custom speech. Wouldn't it be interesting if
we could create a voice that provide s that level of soothing? When
somebody is like , hey , I have a stuffy nose, you have a voice that
represents your brand as a physician telling them, oh, don't worry .
T hat's probably only rhinitis, but it could be something more . Why
don't you come in . It's called [inaudible 00:11:44] , j ust super naturally
sounding speech as opposed to the robotic stuff we hear kind of from
banks and airlines today.

It's conversational . Oh . this is my friend. This is someone who cares
about me. When it starts to sound more natural, we feel more
comforted by it. This is something that we've done studies on just
how easily we can delight our customers and help them by just
creating an inter face that's more natural.

Mark Masselli: We're speaking today with Noelle LaCharite , principal program
manager for applied A I and cognitive services at Microsoft , where her
focus is on helping developers apply Microsoft's artificial intelligence
technology to create conversation al apps .

Noelle , you seem to be all about empowerment , and , Margaret , I
think we've had a couple of fascinating people when you have that
same energy on the radio show . Todd Park liberated data , a great
friend . You're really a part of that vanguard of let's just open this thing
up for people. You've created tutorials for Alexa skill development for
anyone to access it. I'm wondering if you could just talk to our
listeners about some of the tools that are out there for people who

Nicole LaCharite

want to develop their own voice technology apps. Tell us more about
that.

Noelle LaCharite: Yeah, absolutely. The nice thing is you go to Alexa today and there are
dozens of blueprints. I'm a huge fan of learning by doing . These code
snippets are fully functio nal right there. Granted , one of them is like
about space facts . They're not about the content , but all the code is
there. I always encourage people like just go . You'd be shocked with
zero technical skills , i f you know how to follow instructions , you coul d
build a skill for the first time, publish it to 10 million Alexa users , and
have never built code before. I've many, many people that I work
with , they do this every day that are like, oh my gosh, I can't believe I
just built a production app . Even Alexa today will even host your app .
You don't have to worry about infrastructure or how to do it.

The most important thing today is thinking about how can you create
content and create an experience on Alexa that your users will want.
The nice thing of course is that today , people are more tolerant than
ever because voice is just still very cool . Another place to see speech
and voice integrated with other technology is in the AI School . There's
this free online school , called AI School at microsoft.com. We hav e
skill development , a ll of these technologies kind of com ing together to
create scenario - based applications . There are blueprint s you can do in
an hour . These are more like eight - hour projects , but you end up with
a portfolio of work that you can literall y claim that you've built.

It helps you refine your résumé , add these skills to your toolbox , and
just get super - excited. Of course , I'm always trying to share all of it.
There's a Channel 9 on YouTube with the AI Show . It's on YouTube .
Anything you could do to immerse yourself in this. If you have a desire
to build , the technology is definitely available and ready for you to do
it.

Margaret Flinter: I can sense your mission driven ness . We're very mission driven and
our work as well. I've picked up on in w hat's been written about you is
your drive to empower everybody , but especially women through
your Lady Coder s podcast where you see to encourage women
technologists to embrace their tech skills. We often hear that women
have been under - recognized in the t ech world . That's really starting to
change. Share with our listeners your quest to empower more women
tech developers.

Noelle LaCharite: Luckily , just recently , the Gates Foundation did this huge study and
released its report , called the Reboot Representa tion . It was good and
bad . The gist of it was we've got a great new thing with girls who code
and black girls code and all these things that are opening the funnel of
women or girls really moving into this space of computer science.
What then happens is th at within your first two projects, you kind of
start to get this friction. It ' s part of this kind of unconscious bias . The

Nicole LaCharite

people that are providing this friction don't even know they do it ,
which is the worst part about it because it's very hard to fix so mething
that people don't know is broken. After your first or second project ,
you're kind of like, do I really want to live in this like inherent friction?
People end up dropping out.

We see a massive amount of dropout rates across newly hired and
newly g raduated computer science women . They go into things like
psychology and teaching . It's all because they just didn't have a better
tribe to be like, no, no, no, you're fine. In my whole career , whenever
something bad would happen , I'd always be like, I am so alone in this .
Only now when I'm more senior, like I feel more confident to reach
out to people and be like, does this ever happen to you? Almost every
woman I talk to is like, absolutely, it happens all the time.

I built Lady Coders to do that just to tell the stories , so people know
they're not alone . We have an event that we do every year , called
Harmony Conference , where we bring women together and we do
tech sessions, leadership sessions, career building sessions because
sometimes you're going to b e all alone and you may be the only
woman on the team , so how do you not feel that way.

We're doing a great job getting women into the field. We're not doing
a great job of really nurturing them and keeping them in the field . As
a Latina , if I get a situa tion where I'm not super - happy or if I feel not
appreciated, I'll just kind of wipe my hands and go to the next thing.
The Latin culture is very much like, just walk away . I don't want that
for the other Latin women . We went from 4% , which is dismal as it
sounds , but we went to 1% over the last three years. Like I said, I
recognize why that happened, but I would really like to help and
nurtur e and en courage, not just in my specific demographic, but all
women in technology.

Mark Masselli: Margaret mentioned earlier that we're engaged in this All of Us
project th at NIH has set out for a million lives . When you were talking
about the CEOs , they don't look like us. That's exactly what Francis
Collins was saying w e need to do . From our population , we ' re really
tr ying to figure out how to reach out to all people , so that they can be
part of this opportunity for the new sciences that are being
developed. As I think about precision medicine , I think about maybe
the intersection that might occur with the work that you 're doing.

There's a paucity of genetic counselors in the country. There are like
4,000 of them . We need ten fold on that . I'm wondering if you see any
intersection of the work that you're doing really being supportive to
this new field that's growing and where you have so many people
who have questions to ask , and there's nobody really to answer them.

Noelle LaCharite: It's very easy for me. I talk on like global stages and I'm talking to

Nicole LaCharite

multiple entities at the same time . It's so much more powerful when
you can be use case driven . That's something , like at Microsoft
specifically and also at Amazon , we 're very much focused on let's
identify a specific user that we are targeting and make sure that we
are building technology that enables them.

For example , you're talking about geneticists. I have a son who has
D own 's syndrome . I had no idea when we first got that diagnosis what
was going on , but worse I was assigned a geneticist and she was like
75 and done . I think she spoke 10 words to me. I had no idea .

There 's so much the geneticists that are in house, like in seat [ PH
00:19:03] right now, could do to kind of expand the reach of their
scope. They can build applications . They can build a conversational
aspect, let users talk to what's in their head withou t actually having to
talk to them . It could be the far reaching benefit of all of these
parents and children , and families that would benefit from that. It's
totally worth it . I think the precision medicine space is timed to do
just that kind of work.

Marg aret Flinter: We've been speaking today with Noelle LaCharite , principal program
manager for applied AI at Microsoft . You can learn more about her
work by going to noellelacharite.com or you can follow her on Twitter
@NoelleLaCharite.

Noelle , we want to t hank you for your enthusiasm, your innovations ,
your absolutely groundbreaking work , and good humor , and for
joining us on Conversations on Health Care today.

Noelle LaCharite: Absolutely . It's my pleasure.

[Music]

Mark Masselli: At Conversations on Health Care , we want our audience to be truly in
the know when it comes to the facts about health care reform and
pol icy. Lori Robertson is an award - winning journalist and managing
editor of FactCheck.org , a nonpartisan , nonprofit consumer advocate
for voters th at aim to reduce the level of deception in U . S . politics.

Lori, what have you got for us this week?

Lori Robertson: In response to a question about climate change , President Donald
Trump falsely claimed that the United States was " the cleanest we've
ever been " . Data from a variety of sources show that while the
country has made strides in the last decade, its carbon dioxide
emissions are far from their lowest point. The President's comments
came in late November after he was asked by a journalist about the
latest national climate assessment. The report , written by more than
300 government and non - government experts , describe d in detail
how climate change has already begun to affect the nation. It also

Nicole LaCharite

warns of future harm to people, ecosystems , and the econ omy if
more action isn't taken soon.

We've previously debunked similar claims by Trump that the U . S . has
the cleanest water and air. It doesn't , b ased on the 2018
Environmental Performance Index. The relevant metrics are
greenhous e gas es , particularly CO2 . Data from the Energy Information
Administration shows that energy - related CO2 emissions in 2017
were the lowest they had been since 1992 , but e missions were lower
still in every year from 19 73 through 1992. Also , EIA projected in
October that CO2 emissio ns will go up in 2018 by 2.2%. That would be
the first rise in carbon emissions since 2014.

The President also compared the U . S . to other countries . While the
U . S . has made progress on reducing its greenhouse gas emission s, it is
not the lowest emitter. D ata from the International Energy Agency for
2016 show that the U . S . is actually second to China in its total CO2
output , followed by India, Russia , and Japan . The U . S . is higher than all
of these countries in its per capita value s .

That's my fact check f or this week. I'm Lori Robertson, managing
editor of FactCheck.org.

Margaret Flinter : FactCheck.org is committed to factual accuracy from the country's
major political players and is a project of the Annenberg Public Policy
Center at the University of Penn sylvania. If you have a fact that you'd
like checked, e mail us at chcradio.com . We 'll have FactCheck.org ' s
Lori Robertson check it out for you here on Conversations on Health
Care.

[Music]

Mark Masselli: Each week , C onversations highlights a bright idea about how to make
wellness a part of our communities and everyday lives. When Leanne
Brown was a graduate student in nutritional science at NYU , she came
to understand the enormous challenges of eating healthy foods , w hile
doing field work with some of New York City's more impoverished
populations. She thought why not write a cookbook of health recipes
aimed at the millions of Americans , living on food stamps or SNAP
stipends of $4 a day .

Leanne Brown: $4 a day is sort of the general rule of thumb for around how much a
person who is qualifying for food stamps would have to eat per day
sort of explained like a budget rate.

Mark Masselli: She conducted extensive research on shopping and healthy cooking
techniques , and craf ted Good and Cheap , a cookbook aimed at not
only teaching these families how to shop for affordable produce and
healthy foods , but how to get an entire family to cook in a more
healthy way.

Nicole LaCharite

Leanne Brown: I really wanted to arm people with the ability to wa lk into a grocery
store and say like, okay, this is on sale . I can totally make something
delicious out of that, find the deals and find the value in order to
really get b uy - on on such a small amount of money .

Mark Masselli: She launched a Kickstarter camp aign to make the book available at
soup kitchens , women shelters , and community health centers .

Leanne Brown: The idea is that this is a cookbook that needs to be put into the hands
of someone who really can't afford the cookbook . That's where the
idea of doing a buy one get one like TOMS Shoes came from .

Mark Masselli: Good and Cheap, a cookbook aimed at anyone on fixed income
helping to positively influence their diets and well - being , n ow that's a
bright idea.

[Music]

Mark Masselli: You've been listening to Conversations on Health Care. I'm Mark
Masselli .

Margaret Flinter: I ' m Margaret Flinter.

Mark Masselli: Peace and health .

Margaret Flinter : Conversations on Health Care is recorded at WESU at Wesleyan
University , streaming live at chcradio.com , iTunes, or wherever you
listen to podcasts. If you have comments, please e mail us at
[email protected] , or find us on Facebook , or Twitter. We love
hearing from you. This show is br ought to you by the Community
Health Center.

[Music]