How AI and humans are keeping each other alive
[Music]
[Applause]
what do you feel
when i say artificial intelligence
not what you think it is what do you
feel
when i say artificial intelligence
that’s the question
i ask my mba students when they start
their journeys
ai will do amazing things
to humanity if
it is used right in the last 20 years
i was a part of dozens of ai projects
we use ai to save lives
we use ai to make them better but the
responses that i get about
feelings are not optimistic
they are not even positive
many people say they fear that if they
bring ai
to their companies they’ll spend
millions but eventually they’ll fail
no wonder almost 90 percent
of ai projects today fail to bring
results
many other people say they fear that in
the near future
ai will take their jobs
clearly humanity cannot enjoy
the benefits of ai when so many
wonderful people
fear ai three years ago
i decided to look for a solution for
this
big problem if there is a solution
and i started researching along my
research
i learned two things the first
that we have to divide responsibilities
what’s for humans and what’s for ai
the second is that the most successful
companies
they don’t focus on how to use ai to
replace humans
they focus on creating
coexistence between humans and ai
and then eventually after a long
research i found a solution
i call it the two piece of human
and a.i coexistence it’s about
letting ai do the predictions and then
letting humans take these predictions
and make wonderful things with them
in short predictions for ai planning for
humans
the two p’s for their research i looked
at
all of the ventures that i was a part of
and many more
it took me back to ask why i coexist
with ai so well
what makes me love ai so much
as a child i was very close to my
grandfather
shraga along my childhood he used to
tell me many stories
he was a holocaust survivor so
some of the stories were about our
family
that died in the holocaust and about his
childhood
as a partisan in the cold forests of
belarus
even took me there when i was 12.
without his intentions he planted two
things in me
one the need to be a protector of life
and the second the belief that i can
if he could then i can but
the thing is that i’m not exactly gi
jane
in fact the reason that my leg is broken
right now
is not because i did anything cool it’s
just because
i was excited about something at work
went outside of my office
walked a little faster than usual
stumbled and fell
so clearly i cannot use my physical
abilities to
be a protector gladly i was very good at
mathematics and computer programming
and then at the age of 18 i joined the
army and at the age of 19 i already
became
the head of a big technological system
a system took a lot of data and using ai
it generated predictions for terror
attacks
i still remember what my commanders told
me
your system helps saving so many lives
because it gives us time time for what
time to plan how to save as many
people as possible you know these things
are very complex
again i learned two things the first
that i can use ai to protect life
and then i fell in love with ai
and data and the second
is that the collaboration between
human and ai is so powerful it can save
lives
after the army for 15 years i worked on
dozens of ai projects and ventures
not all of them in saving lives many
just making them more fun
and easier you probably know this one
when you add pasta to your online
shopping cart
and then you immediately get a
recommendation for a sauce
here’s the sauce on my daughter’s face
that’s the thoughts that was
recommended me by the ai system and the
thing is that i did not
created the division between prediction
and planning
it’s something our brain does naturally
ever since we are born thousands of
times a day
suppose a bee would enter this room
right now
our brain will immediately start to
scan all the data that is relevant to
the event
it will classify the bee as an animal
and then
an animal that is may be dangerous for
us
and then it will
look at the speed and create a
prediction
that’s something that ai is very good at
but then it will move to the second part
the planning what’s the best plan for us
in this case some of you will decide
to run from the door and shut the door
i cannot do that my leg is broken so
i’ll probably freeze and pretend to be a
statue
you know the planning is very specific
to the person
and to the situation this is for humans
at least in the next few years
so how does it work in ai
projects this is tom
my friend he’s a senior medical doctor
in a large research hospital he called
me a year ago
and he said hug it i have plenty of data
about patients
and i think that if we use ai
we can fight colon cancer much better
it got me very curious and then he
continued
it’s a problem of no show half of the
people that we
send for a colonoscopy don’t show up
or just push back their appointments
over and over and over again
the problem is that some of them already
develop cancer
and this is a huge problem because colon
cancer
develops fast i need you
to find who are the people that will
continue pushing back
and are developing cancer
i wanted to immediately start working on
this
it’s about saving lives
but i knew that if we will not follow
the two p’s
of human and ai coexistence
we will not be among the 13 that succeed
so i asked him what will you do
if i’ll give you the system that gives
you the predictions
instead frankly
i can’t tell you yet but i promise you
that if you will identify the people
who are in danger we’ll find ways to
treat them
so i started working and a few weeks
afterwards
i started avoiding tom it’s just that
i didn’t like the accuracy of the system
it got only
to 65 accuracy
i thought it’s not good enough but
eventually i had to speak with him
and i said tom i’m sorry i don’t think
that it is possible to predict who will
be
in this dangerous situation
tom insisted that i’ll show him the
statistics
the results of the system
and then there was silence
and he said hagit
if the system
identified only 10 of the people that
are in this dangerous
situation it would have been great
your system identifies 65
of the people who are in dangerous
condition it’s amazing
you see the only ones who get
to decide what is fruitful for them
which prediction works are the decision
makers
are the terms are the clients not the
data scientists
unfortunately it usually doesn’t work
because most
projects and companies and ventures
don’t follow the two p’s
of human and ai coexistence and then
humans don’t
take the responsibility after they get
the predictions
a few weeks ago i talked with a friend
of mine who is a
data scientist i wanted to consult and
talk with him about data science
but i also told him about two companies
that i work with
on systems to predict who among their
customers
is probably going to churn
probably going to leave the company
and then he told he said hagit
i recommend you to stop working on these
projects they will not be good for you
i asked why and he told me that recently
he
worked on such a project and
he gave the company that he worked with
accurate predictions about
these people that might churn they
took this list moved it to sales the
sales people called
these people and guess what happened
they got reminded the customer got
reminded that they want to leave the
company
and use that opportunity to live
so you understand that the problem
wasn’t that they had
a system that predicted who is not happy
and
want to live but this the problem was
that they didn’t use the two pieces of
human and ai coexistence
once the managers had the predictions
they automatically moved it to the sale
they didn’t use the predictions to plan
how to save their customers
as humans to humans
their customers are humans you know so
that’s the thing
while many people fear
fear of ai i fear that if we
don’t use ai rights we will lose
our humanity because
in this era we’ll continue collecting
more
and more and more data and decision
making will become
more and more and more complex
and we have a choice we either fear ai
and then we waste all of our time in
trying to make sense
of lots of data discuss it make
predictions
or we use ai to make
quick and specific predictions
in whatever field we are healthcare
agriculture business energy education
saving our planet we use these
predictions
and we plan
how to make better things for humans
from humans to humans
from humans to humanity
it’s our choice it’s in our ends
thank you
you