introduction I can’t tell you how
excited I am to be here today with you
all talking about a topic that I’m
really passionate about and that’s how
data science combined with artificial
intelligence has us on the precipice of
unprecedented change of the course of
human history and if you think real
quick if you look back over some really
major changes that have occurred in
human history of course you’ve had the
Industrial Revolution the Industrial
Revolution as you know was the first
time which machines began to perform the
functions that humans previously did in
this case manual functions and at the
time there’s a lot of concern as to how
humans have we put out of jobs you know
the rapid unemployment but none of that
happened when the Industrial Revolution
unfolded of course there’s some
displacement but what we saw was a huge
explosion in new jobs that didn’t exist
before we saw people moving from farms
into cities we saw a huge increase in
the quality of life in life expectancy
and that unfolded over about a 90 year
period and then fast forward to about
1950 with a start of the computer
revolution and now we have machines not
performing manual functions that humans
used to perform but actually now
providing low-level cognitive functions
that which freed up people to take on
bigger more challenging problems than
than they had in the past now fast
forward to today we have the computer
revolution we have the access to
information anytime anywhere it’s at our
fingertips and we have this explosion of
new skillsets like data science this is
all coming together to create a scenario
where we’re about to take significant
leaps forward in what we’re capable of
as humankind so it’s really I think
important to think about the differences
between how humans learn and how
machines learn before we get into what
the future could very much look like
humans basically learn through evolution
experiences in culture and over the
course of human history there was a
pretty pretty steady trajectory going
forward then with the advent of
computers all of a sudden you saw that
he
knowledge increasing at phenomenal rates
most people don’t realize but 98% of the
data that exists in the world today was
created in the last two years that’s a
stunning factor so you just think about
what that means for our knowledge of the
world around us
now machines on the other hand we built
machines to learn a little bit
differently than humans we built them to
learn through simulations through
anomaly detection and by trying to model
the human brain and trying to emulate
the human break we’re not really far
along in a slaughter point but that will
be something that will certainly unfold
in the years to come so machines can do
some amazing things but there’s also a
paradox here there’s some simple things
that they can’t do so as you know you
can interact with technology now you can
interact with Siri you can interact with
Alexa and Syria or Alexa will perform
simple cognitive skills or functions for
you but you really can’t have a
conversation with Alexa if you’ve ever
tried having a conversation with any of
the digital assistants it’s really
really frustrating there’s a lot of work
to be done there everybody’s heard about
driverless vehicles we that’s certainly
the future but there’s still great
limitations that exist that machines
have yet to be able to overcome
for instance you can’t operate
driverless vehicles in a snowstorm in
fact you can’t really operate it in a
heavy rainstorm so there’s a lot of work
that needs to be done there as well now
there’s a couple things on this list
here that when we built this these slide
six months ago machines couldn’t do know
the ones that are crossed out six months
ago machines weren’t able to create new
cures for diseases or vaccines for
diseases and I’ll talk about that in a
moment
and although machines are able to create
art they weren’t able to create art that
was visually appealing to humans now we
can so again we’re coming of a very very
long way the other thing that machines
can do is they can really whip humans in
anything that’s a mathematical base or
strategy based game so those are things
that you know we’re going to rely on
future on machines for in the future but
what machines can’t do is they can’t be
inquisitive they can’t take lessons
learned in one field and apply them to
another and that’s where humans come in
and it’s very powerful when you think
about the marriage of these increasingly
powerful machine
with heat with the human curiosity so
again to the paradox between what what
machines could do in what humans can do
a three-year-old can ace the test here
and pick out which one are the dogs in
which one of the blueberry muffins
computers are terrible at this they
don’t even get it right 50% of the time
which really goes to the complexity of
the human brain and what we’re capable
of and what machines are capable now it
doesn’t always have to be about food in
animals but another picture here –
between doughnuts
I’m sorry bagels and dogs computers
can’t tell the difference yet everyone
sitting in this room could easily tell
that there’s huge differences here so as
Mark Zuckerberg said we’ve never been
closer or further away from the future
of AI now in terms of some of the really
awesome things the machines could do
normally we want to show this slide to
folks I asked folks to say okay can you
pick out Eve in this picture kind of
like Where’s Waldo and everybody picks
the scientist peering through the
microscope on the right there that’s not
Eve
that’s a scientist who is working with
EE – Eve is the machine that you see
here Eve is a really remarkable machine
as I mentioned before 6 months ago a
year ago we did not have machines coming
up with cures independently of humans
now we have a very very promising
malaria drug and you think about the
years that have been spent trying to
come up with a vaccine for malaria and
here you have Eve doing it in the course
of six months to a year and that’s just
the start
tremendous possibilities there other
things that machines can do better than
humans were at the point now with you
know based on your genetic code machines
can analyze and determine the best
treatments for diseases in afflictions
that you may have way better than then
doctors can in many cases as I mentioned
before we have driverless vehicles we
have driverless not just cars and trucks
but we’ve driverless drones airplane
submarine ships and a lot of these
repetitive tasks again will be done by
machines in the future so limitations
but we are making lots of progress
I mentioned the example of
about art and although machines are
producing art that’s visually appealing
we machines really do struggle still
with a written word it’s got something
to do with the complexity of language
and so machines aren’t able to write
poems that are appealing to humans or
write songs or write books or novels but
maybe someday the other thing that
machines I Ben should just really whip
humans at is any game of strategy
you know the game go here or chess or
any type of complex strategy game we
just can’t beat machines anymore
now I’m also really excited about where
we’re going in the future so Booz Allen
as you as Zoe mentioned we’re consulting
in technology firm been around for about
a hundred years and we’ve really pivoted
it of last four or five years to really
focus on this combination of Technology
and human curiosity and what that can
create to make the world a better place
and one of the things I’m super excited
about is we have a concept that we’re
leading up called super agent and it’s
the idea is that you’re going to help
law enforcement officials be infinitely
more capable than they are today and
it’s not about arming and with more
weapons it’s about arming them with data
and so what the super agent concept is
think about Robocop without all the
negative aspects of it it’s a series of
sensors that a law enforcement official
wears as a inter-row and it
automatically scans that room in
determines if there’s guns knives or
anything threatening that that officer
should be aware of it’s also scans the
room for all the individuals that are in
it and does a biometric scan to see if
any of those are people of interest or
people who have dangerous backgrounds
and the last thing I’m really excited
about on this super aging concept is
we’re going heavy into this concept of
micro-expressions and if you’ve ever
studied the field of micro-expressions
it’s fascinating really aberrant
behavior is often characterized by very
strange combinations of micro
expressions so our theory is is that if
we can identify what those micro
expressions are that causes somebody to
flee order to causes somebody to act in
a violent way we could certainly make
law enforcement much much more effective
then than they are today the second area
that I should highlight that’s really
cool we’re working with FEMA the Federal
Emergency Management Agency to really
change the way we respond to disasters
disaster response today is inherently
inefficient ineffective we often spend
40 10 20 times what it would cost in a
non disaster type environment and we’re
trying to leverage crypto currencies
we’re trying to leverage the digital
economy like uber and airbnb to make the
process of responding in preparing for
disasters way more effective in
efficient that’s something we could have
never done even three or four years you
know the last area I’d highlight that
were excited about we launched if you’ve
ever heard about it the data Science
Bowl it’s basically data science for
social good and we were able to create
last year a set of the algorithms that
allow ready
radiological technicians to scan x-rays
more effectively than humans we focused
on lung cancer and could you detect lung
cancer quicker in earlier with humans
with machines and you can’t humans and
I’m very proud to say that you can and
so we’re excited to see where that goes
and other types of imagery related
medical applications and so yeah as we
go forward in the future there’s four
things you can absolutely be confident
of machines will be doing more and more
repetitive higher order cognitive tasks
we’re freeing up the human intellect to
take on more and more challenging
problems you will have machines as
co-workers which is going to be weird
but you will and in addition to having
machines s co-workers
you’re not going to interface with
technology like we do today where you
sit down in front of a keyboard you type
stuff in you’re going to interface with
technology like you do with other human
beings and I think that’s exciting but
really when you get right down to what’s
the most exciting part about this is
that unprecedented change we have the
ability in front of us to fix problems
that we never dreamed of fixing before
we have the ability to create increased
life expectancies by orders of magnitude
to clean up the oceans to clean up the
environment
and really do some amazing things and
doing it cheaper faster and better than
we ever did before so my advice to you
unlike the first speaker it’s been a
little bit longer for me since I was
sitting in your shoes but I do vaguely
remember those those years think about
the future imagine 10 years from now
imagine your field of study imagine
you’re drinking job you’re working your
dream job
think about those repetitive cognitive
functions that you won’t have to do
anymore
and instead what will you focus your
time on what kind of bigger problems can
you can you take on that you wouldn’t
have been able to do if it wasn’t for
the advent of artificial intelligence
and also think about what other skills
you need to be able to take on those
those projects and I think if you do
that you can have a really really
successful future now if you want to
learn more about AI and where it’s
headed to my colleagues just Sullivan
and Angelucci tavern wrote this book the
mathematical corporation brought several
copies by and to have it in the library
encourage you to check it out or drop me
I know I’ll I’ll make sure we get that
post so don’t make sure we get you guys
copy the book thank you so much [Applause]