Artificial Intelligence Winners and Losers.
artificial intelligence
winners and losers in the brave new
world
an important talk like this involves a
lot of preparation
there are a lot of critical decisions to
be made and in a pandemic
there is a big dilemma the real social
dilemma
[Music]
do you go outside and get a haircut or
do you get a home haircut
and look funny like this
artificial intelligence is a very
important topic but it’s also a complex
topic
it means different things to different
people and each of us
has our own perception of what ai is and
what ai’s impact will be
we’re just like those blind men in the
old fable trying to make sense of the
elephant in the room called artificial
intelligence
a lot of anxiety about ai is because of
these differing perceptions
let us try and clear the confusion if
you want to learn about a topic
your first stop usually is google so
let’s go to google and search for
artificial intelligence
you can try this out now but there are
more than 130 million results for
artificial intelligence
too much information overwhelming how
about we pick a book
so let’s go to amazon and search for
books on artificial intelligence
40 000 books wow information overload
again
or as hell 9000 would say
i’m afraid i can’t do that
have you wondered how we form our
opinions
about ai who teaches the world about ai
i know the geeks and nerds among you are
thinking about professor andrew ying
note that i’ve had this haircut even
before zuckerberg
professor andrew’s courses on ai have
about 3.4 million enrollments
but everyone who signs up for an online
course knows
we excel in what we excellent dropping
out
only about three percent complete the
courses
i’m sorry andrew we can’t do that
then how does the world learn about ai
from where
do we get our knowledge there are two
main sources
the first one is hollywood
yes from movies professor arnold has
taught more folks about ai than
professor andrew
these movies are all about humans versus
evil robots
robots taking over the world world
domination
apocalyptic scenarios robots that fight
their creators
sorry i goofed up here my bad my
apologies to all the rajini fans
this definitely super intelligence
movies are also about love stories
they teach us how humans fall in love
with ai and androids
our relationship with ai it’s
complicated
these movies are great fun but they’re
far from reality
i tell you our professors are misleading
us
so we don’t have machines or robots that
have consciousness
free will feelings and emotions
or what we call artificial general
intelligence
our next main source of misinformation
i mean information
the news media
the media they try hard but that
business
it’s all about being sensational if
hollywood
is about the end of the world media is
about the end of jobs
so the media is preoccupied with how ai
is overtaking humans
how ai will replace humans how it will
lead to massive job losses
again the reality is very different yes
ai is definitely getting better but our
fears of massive job losses to ai
are overblown in reality media hypes up
the current capabilities of ai
the media narrative is that ai is a
solution to practically everything under
the sun
from tapping renewable energy to curing
cancer
from improving education to saving the
planet from solving for climate change
to ending world hunger
don’t believe me just read the newspaper
journalists also don’t question whether
ai is indeed the right solution to such
problems
media also does not acknowledge the
debates about ai
its side effects such as biases and
unintended consequences
to be sure ai is useful in many of these
areas
but is it really the hero or the
sidekick
to sum up for hollywood ai is superhuman
for media ai is a superpower but these
two narratives
don’t give us the real picture of ai at
this point in time to give you a clear
picture
i will discuss a few questions that are
on the top of your mind
things like what actually is ai what are
the skills of the future
what’s the advice for students what’s
the advice for professionals
how should countries prepare as
artificial intelligence becomes more
widespread
very tough questions how i wish there
was a well-read ai
that can give us good answers to these
questions
surprise ladies and gentlemen the next
part of this tedx talk
will be given by ai yes that’s right
[Applause]
media people watching that’s a good
headline ai gives a ted talk
ai that is very well read yes it has
read all those millions of documents
the ai answers these questions based on
what it has seen in
the millions of documents it has read
yes
it’s our high tech version of monkey c
monkey do
because the answers it gives range from
the excellent to the unmentionable
humans are needed to separate the pearls
of wisdom
and the bs
the questions are mine but the responses
are from this ai software
called gbt3 a shout out to tech
entrepreneur
bashar khattar and the team at
philosopherai.com
some of whose content i will be using
today so let’s go
hello ai let’s start with the basics
tell us a bit more about yourself
what is ai
[Music]
artificial intelligence is the science
of making machines
do things that would require
intelligence if done by people
intelligence is a simulation of human
intelligence processes by machines
ai is the ability of computers to digest
large amounts of data
which is impossible for humans at that
scale and get to accurate results and
decisions in a short period of time
the central tasks of ai include
understanding natural language
recognizing objects and scenes in images
playing games
and strategies
hope you like my robot voice
we’re off to a good start so let’s
continue we all want to know how the
nature of work will change
as ai is more widely adopted and how we
should repair
so my next question is what should
employees
do to succeed as jobs are increasingly
being automated
[Music]
the future of work will be a combination
of automation ai
and human skills the keys to success
will be to get the right balance between
the three
as well as to acquire the skills that
machines and ai cannot do easily
[Music]
smaller monkey by the way is gbt2 the
previous generation of this ai software
a lot of our friends listening today are
students
so this question is on their behalf what
courses
should students study to be successful
in the future
[Music]
the future of education is not about
acquiring siloed skills
but about acquiring skills that are
relevant and transferable across a range
of disciplines
this means that those who can operate in
an interdisciplinary way
using cross-disciplinary skills are the
ones who will be more successful in the
future
[Music]
that’s a very glib response i know where
it is coming from
i think this ai is a fresh mba grant
it has all the right words it knows
everything yet it knows nothing
by the way this gyan is not free
it costs about five rupees for each
answer
my mba was so much more expensive
isb we need to have a chat
here is the all important question which
skills will be in greater demand in the
next five to ten years
[Music]
in order to understand what skills can
be in demand in the next five to ten
years
we need to know what kind of tasks will
be done by machines
we can figure this out by looking at the
components of the automation technology
we can think of automation as a three
level pyramid
at the top are super intelligent
machines that are able to replace humans
in almost
all tasks in the middle of the pyramid
are machines that are able to replace
humans in a specific type of task
finally at the bottom of the perimeter
machines
that only replace humans in a very
narrow category of tasks
[Music]
wow that’s a good response i see what
you’re trying to do here
you’re giving us a framework to think
through
that’s what consultants do
there is even the rule of three
hey vacancy and other consultants you
better watch out
is coming after your work
so here is a serious topic we hear a lot
about
bias and discrimination because of ai
software so how can we reduce the bias
in ai software
[Music]
bias can be found in three basic ways it
can either be built into the data that
is used to train the algorithms
it could be the result of the algorithms
themselves or it could be in the people
who train the algorithms
algorithms are trained to make sense of
the world using examples
if your training examples are biased
then so will be your algorithm
we should do everything we can to ensure
that our training sets
are as diverse as possible we can reduce
bias
in machine learning by being more aware
of the limitations and biases in our own
thinking
[Music]
wow that’s a great response i don’t
think i could have done a better job
articulating that
so we’ve talked about the individuals
students
employees so let’s talk at the countries
i don’t know
what should countries do to benefit
fully from artificial intelligence
there are many things that countries
should do to benefit fully from
artificial intelligence
first they should invest in ai it is
important to realize that ai is not just
a technology it’s a new way of thinking
about problems and solving them
it’s a new way of organizing people and
machines to work together
it’s a new way of doing business second
they can also invest in education
it is important that people learn to
work with machines it is also important
that they learn to work with other
people
third they can invest in research ai is
a fast-moving field
it is important that we understand what
is happening why it is happening
and what it means we should also invest
in
ai for the public good we should invest
in ai for the environment
ai for health care and ai for the arts
[Music]
wow that’s like ai monkey bath
niti ayog and other think thanks i hope
you are taking good notes it’s been nice
chatting with you but it’s time to wrap
up this interview
i had a lot of fun but one last question
it’s about the future
your future where do you see yourself
five years down the road
okay what does the future of ai look
like
the future of ai will not look like
skynet from the terminator films
it’ll be quite benign as long as we make
conscious choices about the future of ai
we need to make sure that the benefits
of ai are broadly shared
and that we use these technologies to
support the need for care
compassion and creativity we want to
have a future with more jobs
and fewer people being left behind
that concludes the first ever ted talk
on ai by
one more thing
as humanity now we are at that crucial
juncture
that tipping point at the creation of ai
humans are able to show empathy
compassion and morality
these are not traits which ai possesses
humans are able to form
emotional relationships with other
humans this is something that ai will
never be able to do
humans also have the power of
imagination which enables us to create
new concepts and ideas
artificial intelligence is a great tool
that can help us in many ways
but if we fixate on a vision of ai that
is not here
or may never be here or if we
uncritically cheerlead or overstate its
capabilities
we will miss a golden chance to make a
better ai
ai is not a superhuman ai is not a
superpower
just like electricity it is a great tool
that we can
harness effectively if we use ai to
augment our capabilities
if we make ai more inclusive if
we use ai responsibly humanity will be
the ultimate winner
thank you