ladies and gentlemen please put your hands together for Ken Goldberg so we
hands together for Ken Goldberg so we
hands together for Ken Goldberg so we keep seeing all these headlines that are
keep seeing all these headlines that are
keep seeing all these headlines that are saying that robots and artificial
saying that robots and artificial
saying that robots and artificial intelligence are gonna steal all of our
intelligence are gonna steal all of our
intelligence are gonna steal all of our jobs and I want us to ask a question
jobs and I want us to ask a question
jobs and I want us to ask a question this morning I wanna I want to ask is AI
this morning I wanna I want to ask is AI
this morning I wanna I want to ask is AI really a threat or could we see it as an
really a threat or could we see it as an
really a threat or could we see it as an opportunity so I grew up in in the town
opportunity so I grew up in in the town
opportunity so I grew up in in the town of Bethlehem Pennsylvania the home of
of Bethlehem Pennsylvania the home of
of Bethlehem Pennsylvania the home of Bethlehem Steel where my father worked
Bethlehem Steel where my father worked
Bethlehem Steel where my father worked most of my friends parents worked and in
most of my friends parents worked and in
most of my friends parents worked and in the 60s I watched as technologies came
the 60s I watched as technologies came
the 60s I watched as technologies came in there were new technologies emerging
in there were new technologies emerging
in there were new technologies emerging that that changed the steel and the
that that changed the steel and the
that that changed the steel and the steel the company couldn’t adapt so over
steel the company couldn’t adapt so over
steel the company couldn’t adapt so over time they started letting people go
time they started letting people go
time they started letting people go people started struggling and eventually
people started struggling and eventually
people started struggling and eventually Bethlehem Steel went bankrupt now the
Bethlehem Steel went bankrupt now the
Bethlehem Steel went bankrupt now the this was a terrible decline I had to
this was a terrible decline I had to
this was a terrible decline I had to watch it with my at my entire city went
watch it with my at my entire city went
watch it with my at my entire city went downhill and today this is playing out
downhill and today this is playing out
downhill and today this is playing out in towns all across America in fact all
in towns all across America in fact all
in towns all across America in fact all across the world people are struggling
across the world people are struggling
across the world people are struggling they’re losing their jobs there’s a lot
they’re losing their jobs there’s a lot
they’re losing their jobs there’s a lot of inequity and the technology of our
of inequity and the technology of our
of inequity and the technology of our time the technology today is this new
time the technology today is this new
time the technology today is this new wave of autumn artificial intelligence
wave of autumn artificial intelligence
wave of autumn artificial intelligence and the artificial intelligence is this
and the artificial intelligence is this
and the artificial intelligence is this new wave is due to three factors one
new wave is due to three factors one
new wave is due to three factors one massive amounts of data are now
massive amounts of data are now
massive amounts of data are now available number two that there’s huge
available number two that there’s huge
available number two that there’s huge amounts of computing to be able to
amounts of computing to be able to
amounts of computing to be able to crunch on this data and number three
crunch on this data and number three
crunch on this data and number three there’s a number of new algorithms that
there’s a number of new algorithms that
there’s a number of new algorithms that are able to process this to learn to
are able to process this to learn to
are able to process this to learn to recognize patterns and be able to
recognize patterns and be able to
recognize patterns and be able to predict events in the future and some
predict events in the future and some
predict events in the future and some are saying that this actually could lead
are saying that this actually could lead
are saying that this actually could lead to a hypothetical transition point where
to a hypothetical transition point where
to a hypothetical transition point where robots suddenly become so intelligent
robots suddenly become so intelligent
robots suddenly become so intelligent and artificial intelligence starts to
and artificial intelligence starts to
and artificial intelligence starts to take over and actually surpasses humans
take over and actually surpasses humans
take over and actually surpasses humans leaving us essentially with nothing to
leaving us essentially with nothing to
leaving us essentially with nothing to do
do
do taking all of our jobs and this is what
taking all of our jobs and this is what
taking all of our jobs and this is what we is causing a huge amount of anxiety
we is causing a huge amount of anxiety
we is causing a huge amount of anxiety all across the the country and around
all across the the country and around
all across the the country and around the world now I’ve been studying robots
the world now I’ve been studying robots
the world now I’ve been studying robots and artificial intelligence for about 35
and artificial intelligence for about 35
and artificial intelligence for about 35 years
years
years and as today I run a laboratory at UC
and as today I run a laboratory at UC
and as today I run a laboratory at UC Berkeley where we study robots we try to
Berkeley where we study robots we try to
Berkeley where we study robots we try to develop them and we take advantage of
develop them and we take advantage of
develop them and we take advantage of these new advances in artificial
these new advances in artificial
these new advances in artificial intelligence to do what we call cloud
intelligence to do what we call cloud
intelligence to do what we call cloud robotics where we’re using massive
robotics where we’re using massive
robotics where we’re using massive amounts of data computing and machine
amounts of data computing and machine
amounts of data computing and machine learning to help robots do things better
learning to help robots do things better
learning to help robots do things better but I don’t believe that robots are
but I don’t believe that robots are
but I don’t believe that robots are going to take over or that they’re any
going to take over or that they’re any
going to take over or that they’re any kind of a threat and here’s why let’s
kind of a threat and here’s why let’s
kind of a threat and here’s why let’s look at the biggest result in artificial
look at the biggest result in artificial
look at the biggest result in artificial intelligence today this is that the
intelligence today this is that the
intelligence today this is that the result that a computer made by Google
result that a computer made by Google
result that a computer made by Google and designed by Google has beaten the
and designed by Google has beaten the
and designed by Google has beaten the world champion at the game of Go this
world champion at the game of Go this
world champion at the game of Go this created headlines all over the world and
created headlines all over the world and
created headlines all over the world and it was a significant breakthrough but
it was a significant breakthrough but
it was a significant breakthrough but what’s important to keep in mind is that
what’s important to keep in mind is that
what’s important to keep in mind is that this is a game you’re playing on a game
this is a game you’re playing on a game
this is a game you’re playing on a game board and everything is in Nice black
board and everything is in Nice black
board and everything is in Nice black and white very well defined the real
and white very well defined the real
and white very well defined the real world where we all live
world where we all live
world where we all live where we have our jobs isn’t like that
where we have our jobs isn’t like that
where we have our jobs isn’t like that it’s very messy and complex and high
it’s very messy and complex and high
it’s very messy and complex and high dimensional so if we want robots that
dimensional so if we want robots that
dimensional so if we want robots that are going to be able to drive cars or
are going to be able to drive cars or
are going to be able to drive cars or help senior citizens or work in complex
help senior citizens or work in complex
help senior citizens or work in complex warehouses and factories or work in
warehouses and factories or work in
warehouses and factories or work in operating rooms things are far more
operating rooms things are far more
operating rooms things are far more complex so I want to make this important
complex so I want to make this important
complex so I want to make this important distinction between the results that
distinction between the results that
distinction between the results that we’re seeing sometimes in in games which
we’re seeing sometimes in in games which
we’re seeing sometimes in in games which are which are significant which are
are which are significant which are
are which are significant which are interesting but that doesn’t mean that
interesting but that doesn’t mean that
interesting but that doesn’t mean that we’re suddenly going to be able to solve
we’re suddenly going to be able to solve
we’re suddenly going to be able to solve all the world’s problems with AI and
all the world’s problems with AI and
all the world’s problems with AI and robotics
robotics
robotics so rather than this idea of the
so rather than this idea of the
so rather than this idea of the singularity this point when robots are
singularity this point when robots are
singularity this point when robots are going to take over I want to offer an
going to take over I want to offer an
going to take over I want to offer an alternative and I call this multiplicity
alternative and I call this multiplicity
alternative and I call this multiplicity now multiplicity is where AI helps
now multiplicity is where AI helps
now multiplicity is where AI helps diverse groups of people work together
diverse groups of people work together
diverse groups of people work together in new ways and it’s it’s a way to think
in new ways and it’s it’s a way to think
in new ways and it’s it’s a way to think of it as that it’s taking
of it as that it’s taking
of it as that it’s taking ái and combining it with IA
ái and combining it with IA
ái and combining it with IA intelligence amplification so we want to
intelligence amplification so we want to
intelligence amplification so we want to amplify our abilities and our ability to
amplify our abilities and our ability to
amplify our abilities and our ability to work together in new ways now it’s based
work together in new ways now it’s based
work together in new ways now it’s based on an insight that comes from the
on an insight that comes from the
on an insight that comes from the foundations of artificial intelligence
foundations of artificial intelligence
foundations of artificial intelligence there’s a technique that’s known as
there’s a technique that’s known as
there’s a technique that’s known as decision trees and decision tree is a
decision trees and decision tree is a
decision trees and decision tree is a way of taking a bunch of data and then
way of taking a bunch of data and then
way of taking a bunch of data and then tuning the tree to be able to process
tuning the tree to be able to process
tuning the tree to be able to process new data and be able to make decisions
new data and be able to make decisions
new data and be able to make decisions from it and it’s very effective it’s
from it and it’s very effective it’s
from it and it’s very effective it’s been used for many years but about 20
been used for many years but about 20
been used for many years but about 20 years ago a group at UC Berkeley figured
years ago a group at UC Berkeley figured
years ago a group at UC Berkeley figured out that if you take not just one tree
out that if you take not just one tree
out that if you take not just one tree but you take a collection of trees and
but you take a collection of trees and
but you take a collection of trees and you train them all but slightly
you train them all but slightly
you train them all but slightly differently and then combine the results
differently and then combine the results
differently and then combine the results at the end they’ll perform better than
at the end they’ll perform better than
at the end they’ll perform better than any single tree so the key idea here is
any single tree so the key idea here is
any single tree so the key idea here is that each of the trees has to be
that each of the trees has to be
that each of the trees has to be different there has to be a diversity in
different there has to be a diversity in
different there has to be a diversity in the forest of trees so if you have just
the forest of trees so if you have just
the forest of trees so if you have just the same trees it doesn’t work but when
the same trees it doesn’t work but when
the same trees it doesn’t work but when you have a diversity of trees you
you have a diversity of trees you
you have a diversity of trees you actually get provably better results
actually get provably better results
actually get provably better results than you get with a single tree this
than you get with a single tree this
than you get with a single tree this idea is really interesting because it’s
idea is really interesting because it’s
idea is really interesting because it’s saying that diversity is very powerful
saying that diversity is very powerful
saying that diversity is very powerful it actually helps us get better results
it actually helps us get better results
it actually helps us get better results and similar things have been shown
and similar things have been shown
and similar things have been shown recently for collections of people so
recently for collections of people so
recently for collections of people so they’re new experiments that have come
they’re new experiments that have come
they’re new experiments that have come out showing that if you take a group of
out showing that if you take a group of
out showing that if you take a group of people are all very similar homogeneous
people are all very similar homogeneous
people are all very similar homogeneous and they and try and solve problems or
and they and try and solve problems or
and they and try and solve problems or innovate they do they perform reasonably
innovate they do they perform reasonably
innovate they do they perform reasonably well but if you can get a diverse group
well but if you can get a diverse group
well but if you can get a diverse group where there’s a lot of differences of
where there’s a lot of differences of
where there’s a lot of differences of opinions they almost always perform
opinions they almost always perform
opinions they almost always perform significantly better and the reason is
significantly better and the reason is
significantly better and the reason is that what you want is different forms of
that what you want is different forms of
that what you want is different forms of intelligence to be in the mix so
intelligence to be in the mix so
intelligence to be in the mix so essentially you are some people have
essentially you are some people have
essentially you are some people have analytic intelligence some have logical
analytic intelligence some have logical
analytic intelligence some have logical intelligence some have musical
intelligence some have musical
intelligence some have musical intelligence and what we want to do is
intelligence and what we want to do is
intelligence and what we want to do is start thinking about adding artificial
start thinking about adding artificial
start thinking about adding artificial intelligence into this mix so using it
intelligence into this mix so using it
intelligence into this mix so using it not seeing it as something in opposition
not seeing it as something in opposition
not seeing it as something in opposition to us but seeing it as something that
to us but seeing it as something that
to us but seeing it as something that can essentially enhance our diversity
can essentially enhance our diversity
can essentially enhance our diversity and our ability to work together now
and our ability to work together now
and our ability to work together now this is not science fiction or some sort
this is not science fiction or some sort
this is not science fiction or some sort of strange prediction distant in the
of strange prediction distant in the
of strange prediction distant in the future this is something’s happening all
future this is something’s happening all
future this is something’s happening all the time right now so every time you
the time right now so every time you
the time right now so every time you type in a keyword on to Google and it
type in a keyword on to Google and it
type in a keyword on to Google and it seems like magic that it pops up the the
seems like magic that it pops up the the
seems like magic that it pops up the the results that you often want well what’s
results that you often want well what’s
results that you often want well what’s going on underneath the behind the
going on underneath the behind the
going on underneath the behind the scenes there is that Google is actually
scenes there is that Google is actually
scenes there is that Google is actually employing something very similar to
employing something very similar to
employing something very similar to random forests there’s many different
random forests there’s many different
random forests there’s many different algorithms many different methods that
algorithms many different methods that
algorithms many different methods that are operating to compute on the data
are operating to compute on the data
are operating to compute on the data that it has and it’s constantly
that it has and it’s constantly
that it has and it’s constantly collecting new data from humans so we
collecting new data from humans so we
collecting new data from humans so we are constantly feeding algorithms at
are constantly feeding algorithms at
are constantly feeding algorithms at Google to be able to give us these
Google to be able to give us these
Google to be able to give us these results so this is an example of
results so this is an example of
results so this is an example of multiplicity in action and the same
multiplicity in action and the same
multiplicity in action and the same things the similar ideas are happening
things the similar ideas are happening
things the similar ideas are happening to recommend books and songs and news
to recommend books and songs and news
to recommend books and songs and news feeds and products to us it’s this
feeds and products to us it’s this
feeds and products to us it’s this combination of enhancing our ability to
combination of enhancing our ability to
combination of enhancing our ability to work together in new ways so I’ve been
work together in new ways so I’ve been
work together in new ways so I’ve been writing articles trying to make the case
writing articles trying to make the case
writing articles trying to make the case that it’s not us versus the machine that
that it’s not us versus the machine that
that it’s not us versus the machine that this is some kind of dangerous threat to
this is some kind of dangerous threat to
this is some kind of dangerous threat to us but actually the machine artificial
us but actually the machine artificial
us but actually the machine artificial intelligence has a potential to help us
intelligence has a potential to help us
intelligence has a potential to help us work together in new ways to be able to
work together in new ways to be able to
work together in new ways to be able to amplify our diversity and our different
amplify our diversity and our different
amplify our diversity and our different dimensions of intelligence and we should
dimensions of intelligence and we should
dimensions of intelligence and we should consider artificial intelligence as just
consider artificial intelligence as just
consider artificial intelligence as just one more dimension now there’s a
one more dimension now there’s a
one more dimension now there’s a precedent when technology had a huge
precedent when technology had a huge
precedent when technology had a huge benefit for society and let’s go back a
benefit for society and let’s go back a
benefit for society and let’s go back a hundred years and the technology at that
hundred years and the technology at that
hundred years and the technology at that time the the issue there was that most
time the the issue there was that most
time the the issue there was that most people were working on farms and the new
people were working on farms and the new
people were working on farms and the new technology at that time was steam
technology at that time was steam
technology at that time was steam engines combines
engines combines
engines combines it started to come into farms and people
it started to come into farms and people
it started to come into farms and people started to realize that that was going
started to realize that that was going
started to realize that that was going to change the amount of jobs amount of
to change the amount of jobs amount of
to change the amount of jobs amount of work that was available in farms so in
work that was available in farms so in
work that was available in farms so in 1910 only 10% of Americans completed
1910 only 10% of Americans completed
1910 only 10% of Americans completed high school because most of them worked
high school because most of them worked
high school because most of them worked on farms but as they realized that there
on farms but as they realized that there
on farms but as they realized that there were going to be these changes as the
were going to be these changes as the
were going to be these changes as the technology was going to come in and
technology was going to come in and
technology was going to come in and change this a group of educators got
change this a group of educators got
change this a group of educators got together and they said we need to start
together and they said we need to start
together and they said we need to start thinking about having training a future
thinking about having training a future
thinking about having training a future generation in a new way so they started
generation in a new way so they started
generation in a new way so they started something called the high school
something called the high school
something called the high school movement and they built high schools
movement and they built high schools
movement and they built high schools all across America and they developed
all across America and they developed
all across America and they developed new curriculum and they made a new push
new curriculum and they made a new push
new curriculum and they made a new push to have people have students stay in
to have people have students stay in
to have people have students stay in school and graduate and it’s incredibly
school and graduate and it’s incredibly
school and graduate and it’s incredibly successful because by 1950 80 % of
successful because by 1950 80 % of
successful because by 1950 80 % of Americans graduated high school and the
Americans graduated high school and the
Americans graduated high school and the benefits of this are tremendous we’ve
benefits of this are tremendous we’ve
benefits of this are tremendous we’ve seen this it then led to the advances in
seen this it then led to the advances in
seen this it then led to the advances in the GI Bill and college education and
the GI Bill and college education and
the GI Bill and college education and many of the benefits that we experience
many of the benefits that we experience
many of the benefits that we experience and enjoy today now I have to tell you
and enjoy today now I have to tell you
and enjoy today now I have to tell you Bethlehem also figured this out
Bethlehem also figured this out
Bethlehem also figured this out Bethlehem Pennsylvania invested in its
Bethlehem Pennsylvania invested in its
Bethlehem Pennsylvania invested in its schools in its technology in its
schools in its technology in its
schools in its technology in its infrastructure and and lynx highways
infrastructure and and lynx highways
infrastructure and and lynx highways linking it to other cities and today
linking it to other cities and today
linking it to other cities and today Bethlehem is a thriving metropolis
Bethlehem is a thriving metropolis
Bethlehem is a thriving metropolis so rather than the fear of this
so rather than the fear of this
so rather than the fear of this singularity that’s going to some way
singularity that’s going to some way
singularity that’s going to some way take over and steal our jobs and and and
take over and steal our jobs and and and
take over and steal our jobs and and and essentially a process I wanted us to
essentially a process I wanted us to
essentially a process I wanted us to consider the possibility of starting a
consider the possibility of starting a
consider the possibility of starting a multiplicity movement so let’s think
multiplicity movement so let’s think
multiplicity movement so let’s think about how multiplicity can change the
about how multiplicity can change the
about how multiplicity can change the way we think and learn so instead of the
way we think and learn so instead of the
way we think and learn so instead of the old school ideas of conformity where
old school ideas of conformity where
old school ideas of conformity where school is where you learn to follow
school is where you learn to follow
school is where you learn to follow rules and everything is about learning
rules and everything is about learning
rules and everything is about learning specific directions and being obedient
specific directions and being obedient
specific directions and being obedient we should start to transform our schools
we should start to transform our schools
we should start to transform our schools in our education to the new generation
in our education to the new generation
in our education to the new generation where we have computers but they use
where we have computers but they use
where we have computers but they use them to enhance our diversity our
them to enhance our diversity our
them to enhance our diversity our abilities to work together and our
abilities to work together and our
abilities to work together and our creativity the things that make us
creativity the things that make us
creativity the things that make us uniquely human so we can start
uniquely human so we can start
uniquely human so we can start emphasizing it like examples like this
emphasizing it like examples like this
emphasizing it like examples like this where these girls from Pakistan formed a
where these girls from Pakistan formed a
where these girls from Pakistan formed a robotics team and four and came to the
robotics team and four and came to the
robotics team and four and came to the United States to compete these are ways
United States to compete these are ways
United States to compete these are ways of bringing people together and diverse
of bringing people together and diverse
of bringing people together and diverse groups to think in new ways to
groups to think in new ways to
groups to think in new ways to experiment to explore to be creative to
experiment to explore to be creative to
experiment to explore to be creative to innovate this is a huge enormous
innovate this is a huge enormous
innovate this is a huge enormous potential
potential
potential so let’s together start a multiplicity
so let’s together start a multiplicity
so let’s together start a multiplicity movement to facilitate the future for
movement to facilitate the future for
movement to facilitate the future for all of us thank you
all of us thank you
all of us thank you you
”
Be First to Comment