artificial intelligence or ya for
the intimates
when I started to get interested in this
research area I remember
it was winter long ago it was
not very popular researchers in ia
it’s a little crazy people locked up in
their labs working on
theories that had no chance
lead
by the way when I told my grand
mother I was going to do a thesis in
artificial intelligence she taught me
looked at the eyes full of love and she
told me I congratulate you my little darling
but frankly you should want to do what
like real job after his voice the granny
I want to tell you in 2017 to make
iia it’s like being geek it’s become
cool super fun for the researchers
in artificial intelligence because
that’s it finally the company is interested in
what we do is beautiful satisfaction
and at the same time it’s very annoying very
annoying to hear all the nonsense
that we tell about it is
besides annoyed that’s what I was
when I saw this new blanket
yorker that will appear
I assure you she will appear or not
that’s so annoyed that’s what I was
when I saw this new blanket
yorker that was directed by an artist
his name is kikuo johnson and I got
said that’s it’s gone we’re gonna
tell a lot of nonsense about
artificial intelligence
then I thought about the cover
inspired I said to myself but do we
could get here
could we come up with a
robotic society in which the
humans would be left behind or
worse parked in waters with a
sign species endangered
around the neck
the answer is simple it’s yes we
could get here and if you are
skeptical about the possibilities
of technology evolution I you
invite to ask you a question they
imagined that I put you in a
space time capsule and I you
send see aristotle you arrive
take your phone out of your pocket and
you tell him and look with this object
you can talk live to a guy who lives
in orbit around the planet where if you
prefer we can watch videos of
cats in your opinion they would be they
would be skeptical aristotle
I think it would be like we
is in relation to this
you evolve good then reassure you I
I do not want this scenario to happen
produce on the other hand I do not have no
no longer want to stop research in
artificial intelligence I want
especially not forbid me to work
on systems that could
improve our daily lives
social lives of our elders and why
not even the state of our planet
so the real question that I ask myself
that’s how we do to avoid
get there and this question I’m not
the first to ask me and take
asimov for example brilliant writer of
science fiction asimov he imagined a
world in which she is in which
men and robots can
to live together and have fun creating
three laws the famous three laws of the
robotics to prevent things from going wrong
and actually this way of creating laws
this is what we are observing
in the scientific community everyone there
goes for its grand prices these big
principles to ensure that the ia is
beneficial to humanity but I
do not think it can work from
put in place principles in my opinion
rather than imagine laws what he
we should do is we
devote all our energy to
transfer to give humans the
knowledge that will enable them to
understand what it is
artificial intelligence
why because more intelligence
artificial progress and I assure you
she progresses more she is difficult to
understand more she is difficult to
understand
and the more difficult it will be to master
see for example at the beginning when we
was making smart systems
used rules to years appeared
like that, for example, that we made
systems that are capable of us
to beat chess is systems
simple enough to explain he had a
specific task to be performed playing
failures one could explain how the
rules had been constructed only after
we count on the computing power of
the machine to do the job we
humans we are not always able to
do these tests and practice it was
understandable by cons the disadvantage
it was necessary to think of everyone by
example if you forget to implement the
rule if there is fire in the room
stop playing chess and not flee
had to hope that there is no fire
in the room today fashion is
to create artificial intelligence
that leads to problem solving
specific like for example
recognize cats so
they are presented with thousands of examples
even millions of copies of
things that need to be recognized and by
recognize we do intensive computing
we train them for hours and
then we hope that in the end our
artificial intelligence will be able
to recognize nine out of 10 cats for
give you an example it’s the
techniques that are used in alpha
go this famous program that beat the
world champion 2 go there is no if
a long time I gave you two figures
alpha go he played five million
parties in three days sealys idol so
the world champion who was beaten by
alpha go had wanted to do the same thing
they would have had to live 1370 years
good then it’s true that now we have
artificial intelligences that are
particularly effective they
solve very specific problems
that beats us in full
domains
what we lost, however, is
that they are much more difficult to
explain we understand less what is
going on inside there is still no
solved either the problem of 6 to the
fire you stop playing you flee the same
main problem with the hyades
today is that we need them
train and that it takes time that
it takes a lot of computing power
and every time we go
made 3 cm funds to miss so
the goal of iia for the future is
rdi ya who will no longer need to be
trained who will be directly capable
to learn for oneself
knowledge they need to
evolve in the environment
so for sure they are
are not ready to fight or to
chess nor ogoni to other games but by
against it is hoped that they will capa-l
will be able to discover by
itself only when there is fire in the
piece it is better to carapater quickly
done and this is this new technique
artificial intelligence that me
less passionate about research
it’s called learning
developmental it is an approach that
inspired psychology in
particular of what we know about the
how children build
their knowledge about the world in
interacting with the world
when we make robots
like those on which
I work what we seek to do
it’s robots that are motivated by
one thing their curiosity and that’s
thanks to their curiosity and their
motivation that will explore their
environment interact with us
in particular and build their
knowledge to evolve
so good it’s not a new idea
learning development talent
did turing in 1950 had already evoked
this idea that
he would probably be a lot more
smart to make an intelligence of
level of a child and educate him that
to tackle directly to reproduce
the intelligence of an adult and I think
that he was right we just put 70 years
before trying
so now imagine a second
that I have reason to believe that one
artificial intelligence
developmental will be next
revolution in artificial intelligence
what does that mean
that tomorrow programmed desire to this will
more the prerogative of scientists in
their research labs that
are working on theories that
will never walk but it will become
the responsibility of all to educate
robots that will occupy our daily lives
will become your responsibility as the
mine and if we want to do it
correctly
educate these robots we will have to
learn how it works
how do they use the values that
have their transmits how they
build their knowledge because
that it’s their curiosity that will be the
motor of their learning to you
think maybe that’s all that you
not that it is the future that
Anyway you will not be there for the
see and me either but in fact I
a scoop for you wrong
the iia she’s already everywhere in our
daily from the bottom of our pocket
in our cellphone until
in the management of our markets
financial and are influenced by all
days probably unwittingly by the
data typically produced by
the pictures of cats we do and we
put on the internet by the choices we
done but also by the choices we make
not so this ia this intelligence
artificial she is in the process of
revolutionize our daily lives
it is a phenomenon of magnitude
significant
and since it is a phenomenon of magnitude
significant it is essential that the
society be well armed to be able to the
train to be interested
unfortunately for now it’s not
necessarily the case it must be said that the
headlines of the press that mix
subtly scientific words
complicated that scare everyone
alarmist terms from the images
of science fiction is in general who
involves robots that kill men
it does not help to inform the debate
hold for example a few months ago we
could read in the mainstream press
united to d
the ope a form of aggression when she
is upset what you are told
in this title it is that the purpose of the game
for it was to win a game of
domination of territory and so for
win she developed a strategy
aggressive I can tell you that if me
I had to fight against this ia
I would have also developed a strategy
aggressive to win anyway not
to let beat by intelligences
artificial apple picking
so how do we do it for
have some form of discernment among
all the media noise that we hear
around artificial intelligence
what can we do as a
that individual to be creator of
the artificial intelligence of tomorrow and
not be a consumer of technology
that we will impose on
have the right reading keys of
artificial intelligence
there are not 36 solutions there
that and that is education
it is imperative that the company seizes
of the subject for that it is imperative that
the actors of intelligence
artificial is committed to more
transparency and to demystify
of artificial intelligence and believe
I already have a lot of initiatives
there are workshops on this subject
summer schools seminars of
conferences of exhibition workshops
there are even people who go on stage
to explain to you that it’s important
but the greatest responsibility
it is you who have it you must you
seized of the subject you must learn to
master the proponents and
outsiders of intelligence
artificial as you master the
cell phone that you have in
your pocket
because if I’m right and you
become the educators of the robots of
tomorrow you have better be well
prepared then I return the question
today
do you think that will be good
parents for tomorrow’s robots my
opinion is that it starts with being
good educators for society
of today because I do not know if
the new yorker hypothesis is true or if
she is wrong
but what I do know is that together we
has all the keys in hand to avoid
that it happen I thank you for
your attention
[Applause]
[Music]
”
Be First to Comment