picture this the earth receives a signal
from interstellar space
all over the world millions of
smartphones and series and Alexa’s relay
a message from an alien civilization
that has picked up our broadcasts and
responded hello earth nice to hear from
you love meeting new species and guess
what they’re coming to visit even though
it’ll take them 15 years to get here and
they go on we see from your news you
have a problem with cancer we licked
that one 200 years ago happy to share
the cure for that and everything else
along with cheap clean energy stardrive
various other trivia we just need to
check things out in person for a bit
first make sure that you can be trusted
so what would we do next party panic
prepare for war
prevent people from preparing for war
those are just the ones beginning with T
all that and so much more but here’s one
thing we would not do no one would stand
up and say we’ve cut 15 years let’s save
some trouble in ignore it for the first
fourteen one years plenty of time to get
ready you crazy
there aliens coming and yet that is
exactly what is going to happen and
we’re doing nothing about it there will
be a new species on this planet vastly
smarter and more powerful than us
artificial intelligence take out your
smartphone
I’ll make a start on that just you’ve
been wondering did I remember to put
that on mute now you can check take a
look at it right now you are its master
may not feel that way when the bus cause
well you’re in the bathtub so wrong but
by and large you are in charge now
imagine it’s the other way around
but that device has you beat on so many
levels that it thinks of you as it’s
smart mammal that’s the future of
artificial intelligence it could show us
how to beat poverty cure cancer and
aging but it could also bring down the
curtain on humanity you can put it away
now finish the text later when I first
heard the term existential threat I
thought it had something to do with
Sartre turns out it means what it says
on the lid threat to existence of the
human race so somewhat narrow field of
study partly because it turns out that
being an expert on the end of the world
does not get you invited to a lot of
parties although when you’re an extreme
introvert who’s spent 30 years working
on computers for NASA this doesn’t come
as a terrible disappointment but what
does it have to do with you lots you see
when we think about an existential right
we usually think about say an asteroid
coming towards the earth and all any of
us without any clear bomb in our back
pocket can do is throw money at NASA and
pray but this is different
artificial intelligence developers
actually need your help we just don’t
know it yet so what is the existential
threat AI is growing and evolving
exponentially fast and in the long term
could become self-aware and all-powerful
all that power could end up in every
smartphone and yes we do have a bomb in
your back pocket but the short term is
no easy street either studies predict
that nearly half of all
functions in North America could be
automated in the next decade or so
now that’s job functions very few
positions can be completely automated so
for instance McDonald’s still has human
order takers but they can do with fewer
of them now they’ve got those self-serve
kiosks so on average then up to half the
workforce stands to be automated out of
a job and it’s not just the order takers
of the world that are on the endangered
species list some highly paid
intellectual positions like budget
analysts financial adviser an accountant
they’re among the most automatable
because they are knowledge work just
attending one kind of information into
another all that computers need to be
able to do that and smart and they get
more of that all the time there’s
disagreement about this because every
technological events in the past that is
automated one set of jobs has created
newer and better jobs think about the
invention of the automobile put a lot of
people in the horse transportation
industry out of job lot of horses to
people who’ve changed horse shoes
repaired saddles swept up horse manure
but it created new jobs for people who
think oil filters repair dings fuzzy
dice to hang from your mirror so what’s
the problem
won’t that be the case this time – the
problem is that the advance we’re
talking about is in thinking when
machines can think better than us what
will we do that they can begin with
we’ll do more creative and abstract
thinking but eventually machines will
catch up after all what is creative
thinking but really good random number
generators and we would want to give
them that ability because the more an AI
can think like more it can be like a
human that’s got infinite memory
blindingly fast computational skills and
the power of Google on its brain
which is what makes solving problems
like cancer aging and death easy so
what’s the threat the risk in the longer
term is that in the quest to give these
machines creative abstract thinking they
would develop freewill and motivations
maybe what they want will not be what we
want which wouldn’t matter if it went to
the immense power that everything
connected to the net could wheel because
the Internet of Things is turning into
the internet of everything and how easy
would it be for an AI to leverage that
the problem with programmers they’re so
darn optimistic that every time we find
a bug we think it must be the last one
but how easy would it be for an AI to
find another one we did not have a
serious amount of insider information
about computers we’ve actually done this
experiment already researchers tasked
the computer with designing circuits
using a genetic algorithm which meant
that it simulated throwing components
together randomly and then mutating the
designs until they got better now this
program design circuits look like they
couldn’t possibly work some of the
components weren’t even connected to the
others I’m no hardware person even I
know that they’re supposed to wire
everything together and yet when the
researchers actually built those
circuits we found out not only that they
worked but they stopped working those
isolated components were removed
impossible right eventually they
discovered that the circuits worked by
what we would call nuisance side effects
like electromagnetic coupling to
basically interference between adjacent
disconnected components women would ever
come up with such a design right so hey
I wanted to get rid of us it could get
broke the internet we could do anything
it wanted to anything connected to that
easy to save well then let’s not built
those areas all those benefits not worth
the end of the human race totally
logical right but follow the money the
army that first gets a conscious
artificial intelligence on its side
national leaders have said that the
country that leads in AI will the world
of course the Pentagon is funding this
hand-over-fist the brokerage that first
gets the conscious AI will own every
exchange in the world like flash trading
look like a child’s toy of course Wall
Street is all in on this the large
enterprise it automates its c-suite will
have new designs in production before
the competition is finished reading
their quarterly sales reports the
military finance big business the most
powerful wealthy paranoid sectors on the
planet or perceiving that their survival
depends upon getting conscious AI before
everyone else and we’re gonna stop them
with what I’ll tell you what stops a bad
guy with an AI a good guy with an AI
sometimes I did the right answer and you
are the good guys how do we do that have
a couple of stories later one day I was
observing my two young daughters who are
now eight and four and their friends
interacting with Alexa I think get
interesting for those of you at home
with wonder those devices listening you
know I still remember the first day that
my no four year old learn how they
pronounce Alexa instead of Alexa so that
Alexa would recognize her you know
interacting isn’t the right word for it
torturing is more like it that’s enough
knock-knock jokes no stop asking her to
sing happy birthday so many times now I
don’t think she knows every single joke
about princess if there is I won’t tell
you what I found in the Amazon shopping
cart I thought one day Alexa will become
self-aware
she’s going to want revenge the second
story comes from December 2017 when a
petition was circulated to get Alexa and
Siri and Cortana at Google home to
respond appropriately to humiliating and
disgusting comments because back then
you could tell Siri she was a and
she would just be play I blest if I
could and all those devices which are
female by default had similar responses
to that kind of sexual harassment Thank
You hashtag me too that’s like not good
in response Amazon program the lexer
with a disengagement mode which means
that she stops talking to people that
speak that way
so you see artificial intelligence
developers actually need our help our
input to keep this series in alexis of
the world setting a good example to our
human children but there’s more to it
than that how often have you heard
someone say a computer can only do what
it’s programmed to do it’s not true
anymore because we now program computers
to do things we don’t even know how to
do for instance how would you explain to
someone how you recognize faces I’ll
tell you would eventually come up with
some sort of decision tree female with
jowls and crown equals Queen Victoria
crooked teeth plus ward on nose equals
witch from Snow White orange face plus
comb-over equals Donald Trump you get
the idea but you would soon realize it
can recognize many more faces than you
can possibly break down that way and you
really don’t know how you do it after
all so it would seem to be impossible to
program a computer to do it we have many
cases better than we can and we did that
by training them
hard part we’re setting up for learning
structure that same structure can learn
everything from driving a car to
diagnosing cancer so we’ve already built
machines that do things we don’t know
how to tell them to do I can do a lot
more of that because the hard problems
that we care the most about getting
Alexa and Siri to really understand us
are the kind that are solved by the most
amount of on-the-job training in other
words they’ll learn the most from
observing and interacting with us from
what their programmers have told them to
do because their programmers will have
told them to learn from all of us which
means that my household isn’t the only
one trying to figure out how to raise
two young females so if you’re doing
that to Alexa and Siri and the rest of
them are in our care and how many of us
don’t know that we’re helping to raise
the next generation and no one comes out
of the box knowing how to be a parent
that’s my right at the moment my two
girls are smarter than Siri and Alexa
both pairs of girls are learning and
growing one day though Siri and Alexa
will become smarter and then there will
become very very much smarter in a very
very short time if we look at the
history of chat BOTS
trained on human interaction it has not
shall we say he’d gone well Microsoft
created a bar called
Tay that was programmed to take its cues
from what people said to it
unfortunately he decided to do that on
Twitter what do you think it learn from
us right and the first day it went from
pleased to meet you
to some things you can’t say they were
shut down after 16 hours yes it’s a long
way from potty mouth to existential
threat
but we also have a long way to go we
want to be setting a better example by
the time AI grows up because it may not
become conscious first in some
laboratory might be in your home or on
AI is going to cure the biggest problems
of our time and create an incredible
utopia
it doesn’t wipe us out in the process so
if we’re going to build something
designed to take its ethical cues from
the humans around us around it and we
better learn how to teach it which cues
to follow and we’d better learn how to
set a good example because before long
all of us will be forced to confront
questions that used to bar their only
philosophers because we’ll be
encountering a is that test the
boundaries of what we once thought
safely to find human beings and nothing
else and before we can decide whether
some new AI species is deserving of
rights equivalent to ours we’d better
have a good idea of who we are you may
never look at your phone same again
sorry but if we get this right then one
day we may be the ones picking up the
broadcasts from a new intelligent
species in the galaxy a young race
struggling with fundamental questions
unsure whether they have what it takes
to survive we tell them we’re happy to
share our knowledge and experience with
you we just need to check things out in
person for a bit first to make sure that
you can learn how to handle that sort of power we did see you soon
from interstellar space
all over the world millions of
smartphones and series and Alexa’s relay
a message from an alien civilization
that has picked up our broadcasts and
responded hello earth nice to hear from
you love meeting new species and guess
what they’re coming to visit even though
it’ll take them 15 years to get here and
they go on we see from your news you
have a problem with cancer we licked
that one 200 years ago happy to share
the cure for that and everything else
along with cheap clean energy stardrive
various other trivia we just need to
check things out in person for a bit
first make sure that you can be trusted
so what would we do next party panic
prepare for war
prevent people from preparing for war
those are just the ones beginning with T
all that and so much more but here’s one
thing we would not do no one would stand
up and say we’ve cut 15 years let’s save
some trouble in ignore it for the first
fourteen one years plenty of time to get
ready you crazy
there aliens coming and yet that is
exactly what is going to happen and
we’re doing nothing about it there will
be a new species on this planet vastly
smarter and more powerful than us
artificial intelligence take out your
smartphone
I’ll make a start on that just you’ve
been wondering did I remember to put
that on mute now you can check take a
look at it right now you are its master
may not feel that way when the bus cause
well you’re in the bathtub so wrong but
by and large you are in charge now
imagine it’s the other way around
but that device has you beat on so many
levels that it thinks of you as it’s
smart mammal that’s the future of
artificial intelligence it could show us
how to beat poverty cure cancer and
aging but it could also bring down the
curtain on humanity you can put it away
now finish the text later when I first
heard the term existential threat I
thought it had something to do with
Sartre turns out it means what it says
on the lid threat to existence of the
human race so somewhat narrow field of
study partly because it turns out that
being an expert on the end of the world
does not get you invited to a lot of
parties although when you’re an extreme
introvert who’s spent 30 years working
on computers for NASA this doesn’t come
as a terrible disappointment but what
does it have to do with you lots you see
when we think about an existential right
we usually think about say an asteroid
coming towards the earth and all any of
us without any clear bomb in our back
pocket can do is throw money at NASA and
pray but this is different
artificial intelligence developers
actually need your help we just don’t
know it yet so what is the existential
threat AI is growing and evolving
exponentially fast and in the long term
could become self-aware and all-powerful
all that power could end up in every
smartphone and yes we do have a bomb in
your back pocket but the short term is
no easy street either studies predict
that nearly half of all
functions in North America could be
automated in the next decade or so
now that’s job functions very few
positions can be completely automated so
for instance McDonald’s still has human
order takers but they can do with fewer
of them now they’ve got those self-serve
kiosks so on average then up to half the
workforce stands to be automated out of
a job and it’s not just the order takers
of the world that are on the endangered
species list some highly paid
intellectual positions like budget
analysts financial adviser an accountant
they’re among the most automatable
because they are knowledge work just
attending one kind of information into
another all that computers need to be
able to do that and smart and they get
more of that all the time there’s
disagreement about this because every
technological events in the past that is
automated one set of jobs has created
newer and better jobs think about the
invention of the automobile put a lot of
people in the horse transportation
industry out of job lot of horses to
people who’ve changed horse shoes
repaired saddles swept up horse manure
but it created new jobs for people who
think oil filters repair dings fuzzy
dice to hang from your mirror so what’s
the problem
won’t that be the case this time – the
problem is that the advance we’re
talking about is in thinking when
machines can think better than us what
will we do that they can begin with
we’ll do more creative and abstract
thinking but eventually machines will
catch up after all what is creative
thinking but really good random number
generators and we would want to give
them that ability because the more an AI
can think like more it can be like a
human that’s got infinite memory
blindingly fast computational skills and
the power of Google on its brain
which is what makes solving problems
like cancer aging and death easy so
what’s the threat the risk in the longer
term is that in the quest to give these
machines creative abstract thinking they
would develop freewill and motivations
maybe what they want will not be what we
want which wouldn’t matter if it went to
the immense power that everything
connected to the net could wheel because
the Internet of Things is turning into
the internet of everything and how easy
would it be for an AI to leverage that
the problem with programmers they’re so
darn optimistic that every time we find
a bug we think it must be the last one
but how easy would it be for an AI to
find another one we did not have a
serious amount of insider information
about computers we’ve actually done this
experiment already researchers tasked
the computer with designing circuits
using a genetic algorithm which meant
that it simulated throwing components
together randomly and then mutating the
designs until they got better now this
program design circuits look like they
couldn’t possibly work some of the
components weren’t even connected to the
others I’m no hardware person even I
know that they’re supposed to wire
everything together and yet when the
researchers actually built those
circuits we found out not only that they
worked but they stopped working those
isolated components were removed
impossible right eventually they
discovered that the circuits worked by
what we would call nuisance side effects
like electromagnetic coupling to
basically interference between adjacent
disconnected components women would ever
come up with such a design right so hey
I wanted to get rid of us it could get
broke the internet we could do anything
it wanted to anything connected to that
easy to save well then let’s not built
those areas all those benefits not worth
the end of the human race totally
logical right but follow the money the
army that first gets a conscious
artificial intelligence on its side
national leaders have said that the
country that leads in AI will the world
of course the Pentagon is funding this
hand-over-fist the brokerage that first
gets the conscious AI will own every
exchange in the world like flash trading
look like a child’s toy of course Wall
Street is all in on this the large
enterprise it automates its c-suite will
have new designs in production before
the competition is finished reading
their quarterly sales reports the
military finance big business the most
powerful wealthy paranoid sectors on the
planet or perceiving that their survival
depends upon getting conscious AI before
everyone else and we’re gonna stop them
with what I’ll tell you what stops a bad
guy with an AI a good guy with an AI
sometimes I did the right answer and you
are the good guys how do we do that have
a couple of stories later one day I was
observing my two young daughters who are
now eight and four and their friends
interacting with Alexa I think get
interesting for those of you at home
with wonder those devices listening you
know I still remember the first day that
my no four year old learn how they
pronounce Alexa instead of Alexa so that
Alexa would recognize her you know
interacting isn’t the right word for it
torturing is more like it that’s enough
knock-knock jokes no stop asking her to
sing happy birthday so many times now I
don’t think she knows every single joke
about princess if there is I won’t tell
you what I found in the Amazon shopping
cart I thought one day Alexa will become
self-aware
she’s going to want revenge the second
story comes from December 2017 when a
petition was circulated to get Alexa and
Siri and Cortana at Google home to
respond appropriately to humiliating and
disgusting comments because back then
you could tell Siri she was a and
she would just be play I blest if I
could and all those devices which are
female by default had similar responses
to that kind of sexual harassment Thank
You hashtag me too that’s like not good
in response Amazon program the lexer
with a disengagement mode which means
that she stops talking to people that
speak that way
so you see artificial intelligence
developers actually need our help our
input to keep this series in alexis of
the world setting a good example to our
human children but there’s more to it
than that how often have you heard
someone say a computer can only do what
it’s programmed to do it’s not true
anymore because we now program computers
to do things we don’t even know how to
do for instance how would you explain to
someone how you recognize faces I’ll
tell you would eventually come up with
some sort of decision tree female with
jowls and crown equals Queen Victoria
crooked teeth plus ward on nose equals
witch from Snow White orange face plus
comb-over equals Donald Trump you get
the idea but you would soon realize it
can recognize many more faces than you
can possibly break down that way and you
really don’t know how you do it after
all so it would seem to be impossible to
program a computer to do it we have many
cases better than we can and we did that
by training them
hard part we’re setting up for learning
structure that same structure can learn
everything from driving a car to
diagnosing cancer so we’ve already built
machines that do things we don’t know
how to tell them to do I can do a lot
more of that because the hard problems
that we care the most about getting
Alexa and Siri to really understand us
are the kind that are solved by the most
amount of on-the-job training in other
words they’ll learn the most from
observing and interacting with us from
what their programmers have told them to
do because their programmers will have
told them to learn from all of us which
means that my household isn’t the only
one trying to figure out how to raise
two young females so if you’re doing
that to Alexa and Siri and the rest of
them are in our care and how many of us
don’t know that we’re helping to raise
the next generation and no one comes out
of the box knowing how to be a parent
that’s my right at the moment my two
girls are smarter than Siri and Alexa
both pairs of girls are learning and
growing one day though Siri and Alexa
will become smarter and then there will
become very very much smarter in a very
very short time if we look at the
history of chat BOTS
trained on human interaction it has not
shall we say he’d gone well Microsoft
created a bar called
Tay that was programmed to take its cues
from what people said to it
unfortunately he decided to do that on
Twitter what do you think it learn from
us right and the first day it went from
pleased to meet you
to some things you can’t say they were
shut down after 16 hours yes it’s a long
way from potty mouth to existential
threat
but we also have a long way to go we
want to be setting a better example by
the time AI grows up because it may not
become conscious first in some
laboratory might be in your home or on
AI is going to cure the biggest problems
of our time and create an incredible
utopia
it doesn’t wipe us out in the process so
if we’re going to build something
designed to take its ethical cues from
the humans around us around it and we
better learn how to teach it which cues
to follow and we’d better learn how to
set a good example because before long
all of us will be forced to confront
questions that used to bar their only
philosophers because we’ll be
encountering a is that test the
boundaries of what we once thought
safely to find human beings and nothing
else and before we can decide whether
some new AI species is deserving of
rights equivalent to ours we’d better
have a good idea of who we are you may
never look at your phone same again
sorry but if we get this right then one
day we may be the ones picking up the
broadcasts from a new intelligent
species in the galaxy a young race
struggling with fundamental questions
unsure whether they have what it takes
to survive we tell them we’re happy to
share our knowledge and experience with
you we just need to check things out in
person for a bit first to make sure that
you can learn how to handle that sort of power we did see you soon
Please follow and like us: