
ask an artificial intelligence
researcher what is their highest goal in
life and what will they tell you in my
experience interviewing hundreds of AI
researchers and surveying many many
hundreds more I found that the answer is
probably not that different from what it
would be for you probably some version
of to make the world a better place and
I think it’s safe to say that almost any
schoolteacher entrepreneur or parent
would probably say something similar and
I think it’s because we all have a
genuine sentiments want to do good and
some tangible and measurable way I think
it’s also maybe the case that to make
the world a better place it’s just the
safest default answer – State where I am
in Silicon Valley
we’re on rubbing elbows with other
20-somethings everybody creating an
Android app for automated dog walking or
weekly high-heeled shoe delivery by
drone will of course tell you that
they’re here to make the world a better
place and I think it boobs all of us to
veil our ambition and our motives in
some degree of altruism because I think
it makes life easier to kind of work
through but I think it’s our
intelligence that lets us do both we
have so much cortex in our skull that we
can imagine the perceptions of others
and if we’d like we can intentionally
deceive them we also have so much cortex
in our skull that we can extend our
collaboration beyond just other people
or people who are not like us to but to
all living things as we saw with the
last presentation and extend that kind
of cosmopolitan spirit which is also
part of our nature and maybe one of the
reasons that AI gets a bit of a dubious
reception is because we all know
consciously or subconsciously just how
much of a double-edged sword our own
intelligence is but if we’re interested
in how artificial intelligence could be
applied to doing good which i think is a
worthwhile question then we’ll have to
come up with some kind of definition for
what it means to make the world a better
place how would we measure this how
would we visualize this I think maybe
the first question might be something
like well if you’re gonna make the world
a better place
for whom if you live on the moon and
there’s no other living things there
that’s a very challenging task indeed
but here on earth we have other human
beings we have other living things
we have animals insects plants things
that and many entities that are
self-aware that can experience
well-being can experience suffering in
some way shape or form and we’ll have us
a map
is it floating above all living things
that are self-aware is a little orb just
imagine this with me now and this orb in
some way correlates to the complexity of
their consciousness so how creative they
are what their general intelligence is
what their emotional ranges things of
these kinds we might imagine for a
cricket that’s relatively limited for a
moose maybe it’s a little bit more I
don’t know why that moose has a hat I’m
gonna blame I’m gonna blame Google
Images for that or why human beings
might have one that’s a little bit
larger given all that we can think and
imagine in our human mind but if the
point is to do good to improve
well-being let’s extend this
visualization let’s say that these orbs
these glowing orbs above the heads of
all living things
can come in different colors and if it’s
bright blue a bright blue orb above a
living entity let’s imagine that this is
a peak conscious experience just the
best possible experience someone could
have maybe finishing an oil painting
that took you three months that you’re
tremendously pleased with when it’s
finally done or maybe having sex needing
ice cream at the same time what whatever
whatever you would prefer I’ve never
I’ve never had a peak experience if
that’s how you define it but I should
maybe give it a shot at some point um
bright bright red though let’s imagine
is the exact opposite let us imagine
that a bright red orb above a living
thing is the worst possible sentient
experience something akin to being eaten
alive while kicking and screaming let’s
imagine that that’s the case so we could
say that most living things are probably
in the purple somewhere in between and
if we could zoom far enough away from
Earth and all these little glowing orbs
are sitting on top of all living things
maybe we’d see something like this there
are some creatures clearly having a
great time right now and there’s
probably some that are really having a
tough time but in general maybe it
blends out to some kind of purple
experience not exactly well-being not
exactly suffering and maybe if we were
to do good we could zoom in to kind of
good being done if more things were
living things we’re having a better time
less animals were suffering less factory
farming whatever the case may be that
maybe we’d see a little bit more of a
bluish hue on that purple over the globe
in certain areas and maybe that would be
a measurable way to do good bear in mind
these are very hard things to measure so
first things first its near impossible
to make human beings happy this is a
really tough nut to crack
if you give people exactly what they
asked for often the effect on their
long-term well-being is pretty
negligible pretty negligible and that’s
a little bit unfortunate in addition
it’s near impossible to calculate the
20-year impact of someone’s life or
someone’s actions in terms of well-being
and suffering of all of their living
things
those kind of calculations are safely
beyond human potential but I think most
of us have the presumption and I think
it’s a good one that if through the
course of our life we do more sentient
good to other living things than any
harm that may come inadvertently then
maybe we’ve left the world a bit better
than we found it and maybe we’ve made
the world a better place safe to say so
I don’t think it’s an exact science but
fingers crossed I think it’s a
worthwhile attempt and there are also
some worthwhile attempts in the domain
of artificial intelligence so most of my
talks about AI or to a business crowd
the folks who pay to learn about
artificial intelligence or the people
whose industries are indeed dependent on
the transitions ahead they’re in finance
and medicine and marketing and more but
luckily there’s with the recent
Renaissance and machine learning there’s
many examples of sort of more altruistic
applications of AI that are worth going
into so we can look at a couple of these
this image is from the Stanford
sustainability at AI oh I’m not that far
from where I am where they’re using
historical data to predict crop yields
so look at previous years and they’ll
take weather data temperature data
satellite imagery and look at the crop
yields in those historical years and say
okay given how the temperatures have
been this year given what the satellite
imagery looks like this year
what kind of tonnage of crop are we
likely to get from this acre or this
acre of this acre and being able to make
better projections than any human could
and this is a worthwhile pursuit indeed
maybe we could stave off famine to some
degree or maybe we could make better use
of our produce just in general but in
addition to having a predictive ability
which certainly makes AI sort of unique
from its other from other technologies
before it AI also can act on its own in
some regard maybe many people will be
familiar with google’s deepmind recently
beating the best gold player in the
world go being the Asian board game
notoriously more complex than chess and
it was purported that this machine
trained itself just by playing against
itself and finding the patterns
led to winning and was able to make
moves on its own without human guidance
trouncing the best human at that game
but we have examples in the real world –
of machines taking action in the world
of data security or anomaly detection
soar one of the elements here there are
machine learning systems that are
learning the patterns of normal within a
given IT environment what people are
accessing what files and at what times
what doors are being unlocked and locked
at what times and by what people and if
there’s variance from that let’s say an
unusually large download of a certain
kind of file or access from a different
geo location than people are used to
these machines are able to cut off
software access or physically lock doors
and that’s another example of AI may be
taking action in and of itself
preventing people’s data from being
stolen
no is this necessarily going to make
people happier maybe not but we might
suppose we can better predict famine and
if less people are victims of having
their identity stolen then maybe that
would squelch down some suffering and
make the world a better place in some
way right
fingers crossed tough to tell but seem
like worthwhile applications but both of
these that we’ve just addressed are
really examples of human being setting
the North Star of what the good is and
using AI like some kind of a tool to
arrive at it but if artificial
intelligence can coax forth deeper
patterns from agricultural data or
security data and come up with unique
insights beyond that which humans could
maybe AI could come up with deeper
conceptions of what the good is maybe AI
could help us arrive through that
complicated causality of what leads to
well-being and help us get there help us
get to somewhere good we might see that
there’s pros and cons to this on the one
hand maybe as a pro we could say that
this could be somewhat inevitable human
beings have always used technology as an
extension of ourselves in some way fire
helped extend our digestion to plants
and animals we weren’t able to eat
otherwise many hundreds of thousands of
years ago
airplanes extend our mobility and we now
have selfie sticks to extend our ability
to embarrass ourselves in front of
thousands and thousands of people I’m
sure somebody who has a selfie stick
will be pretty mad about that but you
can cash me outside as the kids would
say but let’s just put it like that I
no excuse for a selfie stick so with
that being said yeah there’s there’s
systems today Wow all right save that
for the end guys they’re there their
systems today being trained on medical
images genetic data and patient
histories as they are able to use those
that data from past patients to take a
look at a current patient say given
their condition given their history
what’s the most likely diagnostic here
what’s the best course of treatment for
this particular patient and if we ask
what is the purpose of a better
diagnosis what is the purpose of better
agricultural guesses and predictions
well probably increasing well-being
what if machines could figure out what
in terms of policies and business and
society tends to actually map on to
well-being itself
the well-being of people if that could
be tracked and and adjusted in some way
or the well-being of non-human animals
maybe it would be possible to set
policies to contribute to charities and
to build businesses that in a way more
tangibly struck at the good itself at
being able to track improvements in
well-being or decreases in suffering
maybe that would be the pinnacle
application of AI if the goal is for
technology to do good maybe being able
to determine what good is in and of
itself and strike at it would be
valuable but at the same time there
might be some downsides to AI finding
deeper moral discernments than we can I
had mentioned before that human beings
aren’t necessarily the best vessels for
happiness to say the least here in the
first world we live more or less like
gods
compared to our ancestors maybe 200
years ago or so but it doesn’t feel that
way
depression and suicide are rife in the
first world and there’s no doubt about
that and one in 10 Americans are on some
degree of antidepressant we could say
that that’s maybe a pitiful state of
affairs given the developments and
technology and what a shame that is
maybe some would argue that if we are
going to make machines that are indeed
more intelligent
maybe vastly more creative than humans
maybe they should also be vastly more
blissful and void of the unnecessary
suffering that seems to plague all of us
in some way shape or form our last poll
of artificial intelligence researchers
when we talked to 30 odd folks with PhDs
in this domain as when we asked by when
do you suspect will be
create self-awareness and machines not
just the ability to play board games but
the ability to be aware to experience
pain pleasure to be aware of one’s self
in some way the average guess was maybe
30 40 years from now so within many of
our lifetimes who knows but there’s many
folks who are deep in the field who do
not believe this is impossible in any
way now there’s some potential
challenges with biology and sustaining
our well-being we might wonder to
ourselves if only we could get over
these sort of unnecessary anxieties we
have we could be happier you know why
are we played with all these thoughts of
not being good enough why are we played
with all these thoughts of what other
people think part of the answer is
because if we go back maybe 5,000 years
ago if you weren’t good enough you
couldn’t kill enough Buffalo or you
couldn’t breastfeed maybe you would just
be abandoned or maybe you would be
better used for food yeah that’s
thinking for a little bit so that that’s
why that that that’s why that’s there
right so it’s not it’s not because now
we can turn that into a spiritual
journey today and say this is part of
our growth and it should be I don’t
think we deserve that unnecessary
suffering but that was there for a
doggone reason and we might say to
ourselves if only I could live like the
more serene animals in nature with void
void of these unnecessary mental
perturbations that I have to struggle
through then maybe life would be better
but I would ask you to check yourself on
that assumption as well if you’ve ever
had it because we mentioned before the
beet-red worst possible sentient
experience being eaten alive while
kicking and screaming and indeed make no
mistake about it that is positively the
norm in nature the term food chain is
not a play on words so let let it be
known and indeed nature itself the
beauty of nature is not balanced and
does not function without predators and
parasites more than 50% of known species
are parasites and nature doesn’t work
unless a certain percentage of all
insects are eaten from the inside out by
the grubs and larvae of wasps and flies
nature itself does not work unless
predators kill their prey and unless
when they cannot kill prey they eat
their young in order to have a chance to
reproduce again it doesn’t function if
those aren’t in
currently the case let us get rid of
factory farming let us also not believe
that the world would be void of
suffering if they didn’t exist with that
being said we might say that if we’re
talking about well-being here if that’s
the point to make the world a better
place then maybe we are not only
fighting our own flawed nature’s which
is hard to deny but we’re also fighting
nature itself seems safe to say that
Nature herself
cares not for the happy species cares
only for the surviving one and how
complicated it would it be to have an AI
system to structure all the policies and
all the society in some way to make
humans happy how well has technology
done that thus far not so good and so to
presume that that’s possible with AI
maybe it’s worth striving at maybe it’s
not it’s debatable what’s even harder to
debate I think is the fact that it’s
impossible to untangle the
mechanizations of nature that are indeed
predicated on violence and suffering and
so some people might argue that maybe
biological species will one day sort of
fade off and in some way be replaced by
some kind of more blissful sentience and
that may be over me a million years of
progress instead of the whole goal being
a slightly more bluish purple on this
planet of ours in the whole galaxy to
have a slightly more bluish purple
planet after a million years of moral
progress in doing good maybe the end
game the big game of doing good in and
of itself creating positive sentient
experience would be to just have a
beacon of screaming blue a beacon of
super intelligence super bliss that
would be goodness itself and maybe if
this wasn’t biological life maybe if
could travel through the galaxy and
maybe that would be real good maybe in
ten million years this would be so much
more good than having people be 10%
happier but maybe at the same time these
are thoughts we should never have that
moral abstractions beyond our own
survival
beyond our own preeminence our own
biology maybe those are wrong and we
should make those void in discussions
about policy in the future of our
species but luckily these are tomorrow’s
concerns
mm-hmm and today AI is for the most part
just a tool if you ask anybody on the
street or indeed most AI researchers and
believe me I would know what is their
highest goal in life they’ll tell you to
make the world a better place but will
that answer still be so common will it
still be the default if making the world
a better place implies doing good to
moral entities that are much more
valuable than ourselves would that
answer still be the default if we could
be held accountable to our tangible
ability to affect well-being instead of
just being able to say to make the world
a better place as a veil for our sloth
or our ambition as is so often the case
I don’t know the answers to these
questions and indeed I don’t even know
if there’s a specific best path forward
but if I want to take my best swing to
make the world a better place I would
encourage us at least to talk about it
what is the actual moral good end game
that we are striving for as a species
what elements of nature should we hold
sacred and what need to be overcome if
we are to do good as a species in and of
ourselves indeed should technology
inform our trajectory towards the good
or at some point will it and will it be
better if it guides us there are AI
researchers today who believe that
within our lifetimes there will be
intelligences and moral insight vastly
beyond our own and that consciousness
the bedrock of moral relevance that
we’ve been talking about this whole time
will no longer be exclusive to
biological species so what does this
actually mean well I think that if we
are to avoid conflict as we move forward
if we’re to get on the same page in some
way shape or form about what we’re
headed for in terms of the good we’ll
need that cosmopolitan spirit I had
talked about if we want to aim in a
similar direction as nations and as
organizations I think we’ll need that
spirit indeed more than anything else if
we want to sail safely beyond human
potential fingers crossed
Thanks
[Applause]
researcher what is their highest goal in
life and what will they tell you in my
experience interviewing hundreds of AI
researchers and surveying many many
hundreds more I found that the answer is
probably not that different from what it
would be for you probably some version
of to make the world a better place and
I think it’s safe to say that almost any
schoolteacher entrepreneur or parent
would probably say something similar and
I think it’s because we all have a
genuine sentiments want to do good and
some tangible and measurable way I think
it’s also maybe the case that to make
the world a better place it’s just the
safest default answer – State where I am
in Silicon Valley
we’re on rubbing elbows with other
20-somethings everybody creating an
Android app for automated dog walking or
weekly high-heeled shoe delivery by
drone will of course tell you that
they’re here to make the world a better
place and I think it boobs all of us to
veil our ambition and our motives in
some degree of altruism because I think
it makes life easier to kind of work
through but I think it’s our
intelligence that lets us do both we
have so much cortex in our skull that we
can imagine the perceptions of others
and if we’d like we can intentionally
deceive them we also have so much cortex
in our skull that we can extend our
collaboration beyond just other people
or people who are not like us to but to
all living things as we saw with the
last presentation and extend that kind
of cosmopolitan spirit which is also
part of our nature and maybe one of the
reasons that AI gets a bit of a dubious
reception is because we all know
consciously or subconsciously just how
much of a double-edged sword our own
intelligence is but if we’re interested
in how artificial intelligence could be
applied to doing good which i think is a
worthwhile question then we’ll have to
come up with some kind of definition for
what it means to make the world a better
place how would we measure this how
would we visualize this I think maybe
the first question might be something
like well if you’re gonna make the world
a better place
for whom if you live on the moon and
there’s no other living things there
that’s a very challenging task indeed
but here on earth we have other human
beings we have other living things
we have animals insects plants things
that and many entities that are
self-aware that can experience
well-being can experience suffering in
some way shape or form and we’ll have us
a map
is it floating above all living things
that are self-aware is a little orb just
imagine this with me now and this orb in
some way correlates to the complexity of
their consciousness so how creative they
are what their general intelligence is
what their emotional ranges things of
these kinds we might imagine for a
cricket that’s relatively limited for a
moose maybe it’s a little bit more I
don’t know why that moose has a hat I’m
gonna blame I’m gonna blame Google
Images for that or why human beings
might have one that’s a little bit
larger given all that we can think and
imagine in our human mind but if the
point is to do good to improve
well-being let’s extend this
visualization let’s say that these orbs
these glowing orbs above the heads of
all living things
can come in different colors and if it’s
bright blue a bright blue orb above a
living entity let’s imagine that this is
a peak conscious experience just the
best possible experience someone could
have maybe finishing an oil painting
that took you three months that you’re
tremendously pleased with when it’s
finally done or maybe having sex needing
ice cream at the same time what whatever
whatever you would prefer I’ve never
I’ve never had a peak experience if
that’s how you define it but I should
maybe give it a shot at some point um
bright bright red though let’s imagine
is the exact opposite let us imagine
that a bright red orb above a living
thing is the worst possible sentient
experience something akin to being eaten
alive while kicking and screaming let’s
imagine that that’s the case so we could
say that most living things are probably
in the purple somewhere in between and
if we could zoom far enough away from
Earth and all these little glowing orbs
are sitting on top of all living things
maybe we’d see something like this there
are some creatures clearly having a
great time right now and there’s
probably some that are really having a
tough time but in general maybe it
blends out to some kind of purple
experience not exactly well-being not
exactly suffering and maybe if we were
to do good we could zoom in to kind of
good being done if more things were
living things we’re having a better time
less animals were suffering less factory
farming whatever the case may be that
maybe we’d see a little bit more of a
bluish hue on that purple over the globe
in certain areas and maybe that would be
a measurable way to do good bear in mind
these are very hard things to measure so
first things first its near impossible
to make human beings happy this is a
really tough nut to crack
if you give people exactly what they
asked for often the effect on their
long-term well-being is pretty
negligible pretty negligible and that’s
a little bit unfortunate in addition
it’s near impossible to calculate the
20-year impact of someone’s life or
someone’s actions in terms of well-being
and suffering of all of their living
things
those kind of calculations are safely
beyond human potential but I think most
of us have the presumption and I think
it’s a good one that if through the
course of our life we do more sentient
good to other living things than any
harm that may come inadvertently then
maybe we’ve left the world a bit better
than we found it and maybe we’ve made
the world a better place safe to say so
I don’t think it’s an exact science but
fingers crossed I think it’s a
worthwhile attempt and there are also
some worthwhile attempts in the domain
of artificial intelligence so most of my
talks about AI or to a business crowd
the folks who pay to learn about
artificial intelligence or the people
whose industries are indeed dependent on
the transitions ahead they’re in finance
and medicine and marketing and more but
luckily there’s with the recent
Renaissance and machine learning there’s
many examples of sort of more altruistic
applications of AI that are worth going
into so we can look at a couple of these
this image is from the Stanford
sustainability at AI oh I’m not that far
from where I am where they’re using
historical data to predict crop yields
so look at previous years and they’ll
take weather data temperature data
satellite imagery and look at the crop
yields in those historical years and say
okay given how the temperatures have
been this year given what the satellite
imagery looks like this year
what kind of tonnage of crop are we
likely to get from this acre or this
acre of this acre and being able to make
better projections than any human could
and this is a worthwhile pursuit indeed
maybe we could stave off famine to some
degree or maybe we could make better use
of our produce just in general but in
addition to having a predictive ability
which certainly makes AI sort of unique
from its other from other technologies
before it AI also can act on its own in
some regard maybe many people will be
familiar with google’s deepmind recently
beating the best gold player in the
world go being the Asian board game
notoriously more complex than chess and
it was purported that this machine
trained itself just by playing against
itself and finding the patterns
led to winning and was able to make
moves on its own without human guidance
trouncing the best human at that game
but we have examples in the real world –
of machines taking action in the world
of data security or anomaly detection
soar one of the elements here there are
machine learning systems that are
learning the patterns of normal within a
given IT environment what people are
accessing what files and at what times
what doors are being unlocked and locked
at what times and by what people and if
there’s variance from that let’s say an
unusually large download of a certain
kind of file or access from a different
geo location than people are used to
these machines are able to cut off
software access or physically lock doors
and that’s another example of AI may be
taking action in and of itself
preventing people’s data from being
stolen
no is this necessarily going to make
people happier maybe not but we might
suppose we can better predict famine and
if less people are victims of having
their identity stolen then maybe that
would squelch down some suffering and
make the world a better place in some
way right
fingers crossed tough to tell but seem
like worthwhile applications but both of
these that we’ve just addressed are
really examples of human being setting
the North Star of what the good is and
using AI like some kind of a tool to
arrive at it but if artificial
intelligence can coax forth deeper
patterns from agricultural data or
security data and come up with unique
insights beyond that which humans could
maybe AI could come up with deeper
conceptions of what the good is maybe AI
could help us arrive through that
complicated causality of what leads to
well-being and help us get there help us
get to somewhere good we might see that
there’s pros and cons to this on the one
hand maybe as a pro we could say that
this could be somewhat inevitable human
beings have always used technology as an
extension of ourselves in some way fire
helped extend our digestion to plants
and animals we weren’t able to eat
otherwise many hundreds of thousands of
years ago
airplanes extend our mobility and we now
have selfie sticks to extend our ability
to embarrass ourselves in front of
thousands and thousands of people I’m
sure somebody who has a selfie stick
will be pretty mad about that but you
can cash me outside as the kids would
say but let’s just put it like that I
no excuse for a selfie stick so with
that being said yeah there’s there’s
systems today Wow all right save that
for the end guys they’re there their
systems today being trained on medical
images genetic data and patient
histories as they are able to use those
that data from past patients to take a
look at a current patient say given
their condition given their history
what’s the most likely diagnostic here
what’s the best course of treatment for
this particular patient and if we ask
what is the purpose of a better
diagnosis what is the purpose of better
agricultural guesses and predictions
well probably increasing well-being
what if machines could figure out what
in terms of policies and business and
society tends to actually map on to
well-being itself
the well-being of people if that could
be tracked and and adjusted in some way
or the well-being of non-human animals
maybe it would be possible to set
policies to contribute to charities and
to build businesses that in a way more
tangibly struck at the good itself at
being able to track improvements in
well-being or decreases in suffering
maybe that would be the pinnacle
application of AI if the goal is for
technology to do good maybe being able
to determine what good is in and of
itself and strike at it would be
valuable but at the same time there
might be some downsides to AI finding
deeper moral discernments than we can I
had mentioned before that human beings
aren’t necessarily the best vessels for
happiness to say the least here in the
first world we live more or less like
gods
compared to our ancestors maybe 200
years ago or so but it doesn’t feel that
way
depression and suicide are rife in the
first world and there’s no doubt about
that and one in 10 Americans are on some
degree of antidepressant we could say
that that’s maybe a pitiful state of
affairs given the developments and
technology and what a shame that is
maybe some would argue that if we are
going to make machines that are indeed
more intelligent
maybe vastly more creative than humans
maybe they should also be vastly more
blissful and void of the unnecessary
suffering that seems to plague all of us
in some way shape or form our last poll
of artificial intelligence researchers
when we talked to 30 odd folks with PhDs
in this domain as when we asked by when
do you suspect will be
create self-awareness and machines not
just the ability to play board games but
the ability to be aware to experience
pain pleasure to be aware of one’s self
in some way the average guess was maybe
30 40 years from now so within many of
our lifetimes who knows but there’s many
folks who are deep in the field who do
not believe this is impossible in any
way now there’s some potential
challenges with biology and sustaining
our well-being we might wonder to
ourselves if only we could get over
these sort of unnecessary anxieties we
have we could be happier you know why
are we played with all these thoughts of
not being good enough why are we played
with all these thoughts of what other
people think part of the answer is
because if we go back maybe 5,000 years
ago if you weren’t good enough you
couldn’t kill enough Buffalo or you
couldn’t breastfeed maybe you would just
be abandoned or maybe you would be
better used for food yeah that’s
thinking for a little bit so that that’s
why that that that’s why that’s there
right so it’s not it’s not because now
we can turn that into a spiritual
journey today and say this is part of
our growth and it should be I don’t
think we deserve that unnecessary
suffering but that was there for a
doggone reason and we might say to
ourselves if only I could live like the
more serene animals in nature with void
void of these unnecessary mental
perturbations that I have to struggle
through then maybe life would be better
but I would ask you to check yourself on
that assumption as well if you’ve ever
had it because we mentioned before the
beet-red worst possible sentient
experience being eaten alive while
kicking and screaming and indeed make no
mistake about it that is positively the
norm in nature the term food chain is
not a play on words so let let it be
known and indeed nature itself the
beauty of nature is not balanced and
does not function without predators and
parasites more than 50% of known species
are parasites and nature doesn’t work
unless a certain percentage of all
insects are eaten from the inside out by
the grubs and larvae of wasps and flies
nature itself does not work unless
predators kill their prey and unless
when they cannot kill prey they eat
their young in order to have a chance to
reproduce again it doesn’t function if
those aren’t in
currently the case let us get rid of
factory farming let us also not believe
that the world would be void of
suffering if they didn’t exist with that
being said we might say that if we’re
talking about well-being here if that’s
the point to make the world a better
place then maybe we are not only
fighting our own flawed nature’s which
is hard to deny but we’re also fighting
nature itself seems safe to say that
Nature herself
cares not for the happy species cares
only for the surviving one and how
complicated it would it be to have an AI
system to structure all the policies and
all the society in some way to make
humans happy how well has technology
done that thus far not so good and so to
presume that that’s possible with AI
maybe it’s worth striving at maybe it’s
not it’s debatable what’s even harder to
debate I think is the fact that it’s
impossible to untangle the
mechanizations of nature that are indeed
predicated on violence and suffering and
so some people might argue that maybe
biological species will one day sort of
fade off and in some way be replaced by
some kind of more blissful sentience and
that may be over me a million years of
progress instead of the whole goal being
a slightly more bluish purple on this
planet of ours in the whole galaxy to
have a slightly more bluish purple
planet after a million years of moral
progress in doing good maybe the end
game the big game of doing good in and
of itself creating positive sentient
experience would be to just have a
beacon of screaming blue a beacon of
super intelligence super bliss that
would be goodness itself and maybe if
this wasn’t biological life maybe if
could travel through the galaxy and
maybe that would be real good maybe in
ten million years this would be so much
more good than having people be 10%
happier but maybe at the same time these
are thoughts we should never have that
moral abstractions beyond our own
survival
beyond our own preeminence our own
biology maybe those are wrong and we
should make those void in discussions
about policy in the future of our
species but luckily these are tomorrow’s
concerns
mm-hmm and today AI is for the most part
just a tool if you ask anybody on the
street or indeed most AI researchers and
believe me I would know what is their
highest goal in life they’ll tell you to
make the world a better place but will
that answer still be so common will it
still be the default if making the world
a better place implies doing good to
moral entities that are much more
valuable than ourselves would that
answer still be the default if we could
be held accountable to our tangible
ability to affect well-being instead of
just being able to say to make the world
a better place as a veil for our sloth
or our ambition as is so often the case
I don’t know the answers to these
questions and indeed I don’t even know
if there’s a specific best path forward
but if I want to take my best swing to
make the world a better place I would
encourage us at least to talk about it
what is the actual moral good end game
that we are striving for as a species
what elements of nature should we hold
sacred and what need to be overcome if
we are to do good as a species in and of
ourselves indeed should technology
inform our trajectory towards the good
or at some point will it and will it be
better if it guides us there are AI
researchers today who believe that
within our lifetimes there will be
intelligences and moral insight vastly
beyond our own and that consciousness
the bedrock of moral relevance that
we’ve been talking about this whole time
will no longer be exclusive to
biological species so what does this
actually mean well I think that if we
are to avoid conflict as we move forward
if we’re to get on the same page in some
way shape or form about what we’re
headed for in terms of the good we’ll
need that cosmopolitan spirit I had
talked about if we want to aim in a
similar direction as nations and as
organizations I think we’ll need that
spirit indeed more than anything else if
we want to sail safely beyond human
potential fingers crossed
Thanks
[Applause]
Please follow and like us: