so we are living in fascinating era of artificial intelligence
artificial intelligence
artificial intelligence when these technologies are becoming
when these technologies are becoming
when these technologies are becoming ubiquitous in our lives we should ask
ubiquitous in our lives we should ask
ubiquitous in our lives we should ask some serious questions beyond just
some serious questions beyond just
some serious questions beyond just pattern recognitions this technology’s
pattern recognitions this technology’s
pattern recognitions this technology’s fail to be empathetic towards human
fail to be empathetic towards human
fail to be empathetic towards human being they don’t understand what it
being they don’t understand what it
being they don’t understand what it means to be fair and they have no notion
means to be fair and they have no notion
means to be fair and they have no notion of equality so people raised so many
of equality so people raised so many
of equality so people raised so many questions about ethics of artificial
questions about ethics of artificial
questions about ethics of artificial intelligence in MIT Media Lab my work
intelligence in MIT Media Lab my work
intelligence in MIT Media Lab my work revolves around how can we envision
revolves around how can we envision
revolves around how can we envision future that machines and humans can work
future that machines and humans can work
future that machines and humans can work together they can trust each other but
together they can trust each other but
together they can trust each other but then what do we mean by trust in what we
then what do we mean by trust in what we
then what do we mean by trust in what we trust we can’t even trust our social
trust we can’t even trust our social
trust we can’t even trust our social organisations our markets I want to take
organisations our markets I want to take
organisations our markets I want to take you to the three amazing places in the
you to the three amazing places in the
you to the three amazing places in the world first we go to Western Ghats of
world first we go to Western Ghats of
world first we go to Western Ghats of India it is one of the most biodiverse
India it is one of the most biodiverse
India it is one of the most biodiverse place on the planet Earth on the west
place on the planet Earth on the west
place on the planet Earth on the west side of it you see Arabian Sea and as
side of it you see Arabian Sea and as
side of it you see Arabian Sea and as this mountain clouds gather farmers in
this mountain clouds gather farmers in
this mountain clouds gather farmers in India start preparing their land this is
India start preparing their land this is
India start preparing their land this is just a small land it’s one hectare of
just a small land it’s one hectare of
just a small land it’s one hectare of land that these farmers have but amount
land that these farmers have but amount
land that these farmers have but amount of time and efforts this paint just
of time and efforts this paint just
of time and efforts this paint just preparing this land is enormous their
preparing this land is enormous their
preparing this land is enormous their battle is not just with the nature and
battle is not just with the nature and
battle is not just with the nature and waiting for water the battle starts now
waiting for water the battle starts now
waiting for water the battle starts now when they take their 500 kilograms or
when they take their 500 kilograms or
when they take their 500 kilograms or 600 kilograms rice to the market but
600 kilograms rice to the market but
600 kilograms rice to the market but these markets are composed with lot of
these markets are composed with lot of
these markets are composed with lot of people who have self interests some of
people who have self interests some of
people who have self interests some of these people have more information than
these people have more information than
these people have more information than others it’s called
others it’s called
others it’s called information asymmetry in economics what
information asymmetry in economics what
information asymmetry in economics what that does to these markets people who
that does to these markets people who
that does to these markets people who have more information can manipulate
have more information can manipulate
have more information can manipulate people who have no information or less
people who have no information or less
people who have no information or less informations and farmers are one of
informations and farmers are one of
informations and farmers are one of those people who have less
those people who have less
those people who have less resources to kind of interpret and
resources to kind of interpret and
resources to kind of interpret and predict how markets will move by the
predict how markets will move by the
predict how markets will move by the time their goods reach to consumer their
time their goods reach to consumer their
time their goods reach to consumer their profit is one painting or less than one
profit is one painting or less than one
profit is one painting or less than one penny so the future the hard work they
penny so the future the hard work they
penny so the future the hard work they have put together to educate their kids
have put together to educate their kids
have put together to educate their kids or to have better life goes to Penny’s
or to have better life goes to Penny’s
or to have better life goes to Penny’s last 20 years there are around 300,000
last 20 years there are around 300,000
last 20 years there are around 300,000 farmer suicides in India this number is
farmer suicides in India this number is
farmer suicides in India this number is shocking and why it happens it’s just
shocking and why it happens it’s just
shocking and why it happens it’s just because unreliable reputation where
because unreliable reputation where
because unreliable reputation where farmers cannot trust any other in the
farmers cannot trust any other in the
farmers cannot trust any other in the market and inequality that this market
market and inequality that this market
market and inequality that this market causes second I’m gonna take you to the
causes second I’m gonna take you to the
causes second I’m gonna take you to the one of the most beautiful cities in the
one of the most beautiful cities in the
one of the most beautiful cities in the world New York City Wall Street on Wall
world New York City Wall Street on Wall
world New York City Wall Street on Wall Street people wake up every day and
Street people wake up every day and
Street people wake up every day and they’re super smart engineers
they’re super smart engineers
they’re super smart engineers mathematicians physicists or business
mathematicians physicists or business
mathematicians physicists or business people their goal is as the Sun rises
people their goal is as the Sun rises
people their goal is as the Sun rises they hope that market will rise prices
they hope that market will rise prices
they hope that market will rise prices will go up and eventually they will have
will go up and eventually they will have
will go up and eventually they will have algorithms and systems that nobody can
algorithms and systems that nobody can
algorithms and systems that nobody can imagine will beat the market these
imagine will beat the market these
imagine will beat the market these technologies are so complex that no
technologies are so complex that no
technologies are so complex that no human being can imagine how fast the
human being can imagine how fast the
human being can imagine how fast the trades are happening and what’s their
trades are happening and what’s their
trades are happening and what’s their impact on our society I will show you
impact on our society I will show you
impact on our society I will show you what happened on May 6th 2010 after
what happened on May 6th 2010 after
what happened on May 6th 2010 after noontime 2:47 p.m. accident church stock
noontime 2:47 p.m. accident church stock
noontime 2:47 p.m. accident church stock was trading at 40 in a minute stock
was trading at 40 in a minute stock
was trading at 40 in a minute stock price went to a painting and everybody
price went to a painting and everybody
price went to a painting and everybody started wondering what’s going on there
started wondering what’s going on there
started wondering what’s going on there is huge turbulence in the market and
is huge turbulence in the market and
is huge turbulence in the market and people cannot make sense of what’s
people cannot make sense of what’s
people cannot make sense of what’s happening Dow dropped two highest 9.2
happening Dow dropped two highest 9.2
happening Dow dropped two highest 9.2 percent that intraday swing that market
percent that intraday swing that market
percent that intraday swing that market has seen this is the second highest
has seen this is the second highest
has seen this is the second highest interest wing in the history of Wall
interest wing in the history of Wall
interest wing in the history of Wall Street what’s happening in 36 minutes we
Street what’s happening in 36 minutes we
Street what’s happening in 36 minutes we lost not thousand
lost not thousand
lost not thousand million but one trillion dollars of
million but one trillion dollars of
million but one trillion dollars of wealth and we still don’t know how and
wealth and we still don’t know how and
wealth and we still don’t know how and why this happened
why this happened
why this happened so we are building these complex systems
so we are building these complex systems
so we are building these complex systems and markets and organizations and we are
and markets and organizations and we are
and markets and organizations and we are putting our dreams and 401ks and all the
putting our dreams and 401ks and all the
putting our dreams and 401ks and all the efforts that we put in our life into
efforts that we put in our life into
efforts that we put in our life into savings and the value of those savings
savings and the value of those savings
savings and the value of those savings is one penny in the end so we are fooled
is one penny in the end so we are fooled
is one penny in the end so we are fooled by this randomness that this nature and
by this randomness that this nature and
by this randomness that this nature and the world present to us and we still
the world present to us and we still
the world present to us and we still don’t know how to deal with it you can
don’t know how to deal with it you can
don’t know how to deal with it you can say well on Wall Street we have
say well on Wall Street we have
say well on Wall Street we have reputation we have credit agencies
reputation we have credit agencies
reputation we have credit agencies however credit agencies fail because we
however credit agencies fail because we
however credit agencies fail because we seen in 2000 crashes that a lot of
seen in 2000 crashes that a lot of
seen in 2000 crashes that a lot of credit ratings were manipulated so who
credit ratings were manipulated so who
credit ratings were manipulated so who do you trust the third place I want to
do you trust the third place I want to
do you trust the third place I want to take you to the amazing hyper-connected
take you to the amazing hyper-connected
take you to the amazing hyper-connected world that we live in the internet on
world that we live in the internet on
world that we live in the internet on internet as social interactions are
internet as social interactions are
internet as social interactions are growing there is unprecedented growth
growing there is unprecedented growth
growing there is unprecedented growth among the data that we have and this
among the data that we have and this
among the data that we have and this data has enabled the new kind of
data has enabled the new kind of
data has enabled the new kind of economies called
economies called
economies called sharing markets or gig economies for
sharing markets or gig economies for
sharing markets or gig economies for example Airbnb uber mechanical turn our
example Airbnb uber mechanical turn our
example Airbnb uber mechanical turn our TaskRabbit where you can hire people in
TaskRabbit where you can hire people in
TaskRabbit where you can hire people in an instant rate or you can work for
an instant rate or you can work for
an instant rate or you can work for somebody if you have access to Internet
somebody if you have access to Internet
somebody if you have access to Internet so the goal is somebody who has some
so the goal is somebody who has some
so the goal is somebody who has some need in uber you want to go from one
need in uber you want to go from one
need in uber you want to go from one place to other place that you need so
place to other place that you need so
place to other place that you need so you’re putting a request for uber and
you’re putting a request for uber and
you’re putting a request for uber and somebody who has resources it’s either
somebody who has resources it’s either
somebody who has resources it’s either it’s time or some sort of bungalow or a
it’s time or some sort of bungalow or a
it’s time or some sort of bungalow or a flat or a car that he can rent here or
flat or a car that he can rent here or
flat or a car that he can rent here or on Mechanical Turk for example if I’m
on Mechanical Turk for example if I’m
on Mechanical Turk for example if I’m training artificial intelligence
training artificial intelligence
training artificial intelligence algorithm I can put my data in the real
algorithm I can put my data in the real
algorithm I can put my data in the real world and ask people to label it so that
world and ask people to label it so that
world and ask people to label it so that I can train my machine learning
I can train my machine learning
I can train my machine learning algorithm and bake make better in
algorithm and bake make better in
algorithm and bake make better in inferences on the data now this is super
inferences on the data now this is super
inferences on the data now this is super appealing because it’s so easy for
appealing because it’s so easy for
appealing because it’s so easy for people who don’t have jobs to get on
people who don’t have jobs to get on
people who don’t have jobs to get on these markets this is one of the example
these markets this is one of the example
these markets this is one of the example from being a Turk or paper written by
from being a Turk or paper written by
from being a Turk or paper written by David Martin is
David Martin is
David Martin is most amazing paper I ever read on this
most amazing paper I ever read on this
most amazing paper I ever read on this gig economies
gig economies
gig economies it says I’m having a hard time and
it says I’m having a hard time and
it says I’m having a hard time and especially financially roommates running
especially financially roommates running
especially financially roommates running out of money but only thing I can do to
out of money but only thing I can do to
out of money but only thing I can do to pay my bills is get on internet and
pay my bills is get on internet and
pay my bills is get on internet and start talking talking means go on
start talking talking means go on
start talking talking means go on Mechanical Turk and start answering
Mechanical Turk and start answering
Mechanical Turk and start answering these questions that requesters are
these questions that requesters are
these questions that requesters are posting to workers now there is a huge
posting to workers now there is a huge
posting to workers now there is a huge inequality in these markets because
inequality in these markets because
inequality in these markets because think about it
think about it
think about it you’re working with strangers you are
you’re working with strangers you are
you’re working with strangers you are hiding a stranger or you’re working for
hiding a stranger or you’re working for
hiding a stranger or you’re working for a stranger and sometimes people get
a stranger and sometimes people get
a stranger and sometimes people get experiences with it that their work get
experiences with it that their work get
experiences with it that their work get rejected for example on Mechanical Turk
rejected for example on Mechanical Turk
rejected for example on Mechanical Turk if you are a requester you hold the
if you are a requester you hold the
if you are a requester you hold the power to accept or reject the data that
power to accept or reject the data that
power to accept or reject the data that somebody has produced so worker can
somebody has produced so worker can
somebody has produced so worker can spend hours working on and creating the
spend hours working on and creating the
spend hours working on and creating the data sets but requester in a minute can
data sets but requester in a minute can
data sets but requester in a minute can reject that data without paying a single
reject that data without paying a single
reject that data without paying a single penny to the worker so this is one of
penny to the worker so this is one of
penny to the worker so this is one of the court from requester to worker he
the court from requester to worker he
the court from requester to worker he says you received a bonus from some
says you received a bonus from some
says you received a bonus from some company for work related to something
company for work related to something
company for work related to something and the value of your bonus is one cent
and the value of your bonus is one cent
and the value of your bonus is one cent now this is shame because bonus aren’t
now this is shame because bonus aren’t
now this is shame because bonus aren’t is one one one penny again same problems
is one one one penny again same problems
is one one one penny again same problems inequality and broken reputation leads
inequality and broken reputation leads
inequality and broken reputation leads to the breakdown in trust and we see
to the breakdown in trust and we see
to the breakdown in trust and we see internet markets are failing so if we go
internet markets are failing so if we go
internet markets are failing so if we go back to the farmer in India or people
back to the farmer in India or people
back to the farmer in India or people who are investing money on Wall Street
who are investing money on Wall Street
who are investing money on Wall Street or people who are finding new work on
or people who are finding new work on
or people who are finding new work on internet there is one thing common these
internet there is one thing common these
internet there is one thing common these all markets are leading to markets for
all markets are leading to markets for
all markets are leading to markets for laymen’s where value of goods it’s very
laymen’s where value of goods it’s very
laymen’s where value of goods it’s very hard to judge so I believe that every
hard to judge so I believe that every
hard to judge so I believe that every person on earth should get equal playing
person on earth should get equal playing
person on earth should get equal playing field in the market it’s people’s
field in the market it’s people’s
field in the market it’s people’s fundamental right whether you are in
fundamental right whether you are in
fundamental right whether you are in India as a farmer or in the America
India as a farmer or in the America
India as a farmer or in the America investing your money on Wall Street or a
investing your money on Wall Street or a
investing your money on Wall Street or a person
person
person on internet and the only way that can
on internet and the only way that can
on internet and the only way that can happen is we rethink about trust we can
happen is we rethink about trust we can
happen is we rethink about trust we can think about trust and reputation as a
think about trust and reputation as a
think about trust and reputation as a backbone of these economies but what it
backbone of these economies but what it
backbone of these economies but what it means
means
means can we trust reputation itself
can we trust reputation itself
can we trust reputation itself can we trust credit agencies can we
can we trust credit agencies can we
can we trust credit agencies can we trust 5-star ratings that we see all
trust 5-star ratings that we see all
trust 5-star ratings that we see all around us this was the uber I took from
around us this was the uber I took from
around us this was the uber I took from the school across MIT and it asked me
the school across MIT and it asked me
the school across MIT and it asked me please rate there is another economy
please rate there is another economy
please rate there is another economy based on that that creates this sample
based on that that creates this sample
based on that that creates this sample boards that passenger can see and it
boards that passenger can see and it
boards that passenger can see and it shows you five star now the problem with
shows you five star now the problem with
shows you five star now the problem with this is that most of the time people
this is that most of the time people
this is that most of the time people rate in four point seven five to five
rate in four point seven five to five
rate in four point seven five to five not below how often you have given
not below how often you have given
not below how often you have given negative rating to a Lobel driver or how
negative rating to a Lobel driver or how
negative rating to a Lobel driver or how often you have rated a driver very few
often you have rated a driver very few
often you have rated a driver very few times and research shows that around
times and research shows that around
times and research shows that around eighty percent times people vote at the
eighty percent times people vote at the
eighty percent times people vote at the higher end and why that happens it’s
higher end and why that happens it’s
higher end and why that happens it’s more because people don’t care about
more because people don’t care about
more because people don’t care about watching and they don’t have incentives
watching and they don’t have incentives
watching and they don’t have incentives to what until they had a bad experience
to what until they had a bad experience
to what until they had a bad experience or in the marketplaces like Mechanical
or in the marketplaces like Mechanical
or in the marketplaces like Mechanical Turk where you have workers and
Turk where you have workers and
Turk where you have workers and requesters giving public feedback kind
requesters giving public feedback kind
requesters giving public feedback kind of seen as a problematic thing a worker
of seen as a problematic thing a worker
of seen as a problematic thing a worker if he says this requester is bad and if
if he says this requester is bad and if
if he says this requester is bad and if it goes on internet then he may not get
it goes on internet then he may not get
it goes on internet then he may not get work again and similarly if requester he
work again and similarly if requester he
work again and similarly if requester he gives bad feedback to a worker that
gives bad feedback to a worker that
gives bad feedback to a worker that worker can send him thousands of emails
worker can send him thousands of emails
worker can send him thousands of emails complaining about that and the requester
complaining about that and the requester
complaining about that and the requester will end up spending huge amount of time
will end up spending huge amount of time
will end up spending huge amount of time rather than doing his business so this
rather than doing his business so this
rather than doing his business so this is a problem when incentives are
is a problem when incentives are
is a problem when incentives are misaligned so most of the scores lead to
misaligned so most of the scores lead to
misaligned so most of the scores lead to four point seven five out of five and
four point seven five out of five and
four point seven five out of five and it’s known as reputation inflation and
it’s known as reputation inflation and
it’s known as reputation inflation and how do we solve that so I talked about
how do we solve that so I talked about
how do we solve that so I talked about problems but at MIT you don’t talk about
problems but at MIT you don’t talk about
problems but at MIT you don’t talk about problems
problems
problems you solve the problems and to solve this
you solve the problems and to solve this
you solve the problems and to solve this problem I led the invention of incentive
problem I led the invention of incentive
problem I led the invention of incentive compatible reputation system and the
compatible reputation system and the
compatible reputation system and the idea is let’s say you are a worker on
idea is let’s say you are a worker on
idea is let’s say you are a worker on Mechanical Turk or any of the sharing
Mechanical Turk or any of the sharing
Mechanical Turk or any of the sharing economy
economy
economy this mechanism rethinks about how
this mechanism rethinks about how
this mechanism rethinks about how repetition should work so workers
repetition should work so workers
repetition should work so workers incentives are get good work get paid
incentives are get good work get paid
incentives are get good work get paid well and get respectable communication
well and get respectable communication
well and get respectable communication from the requesters Bob a worker
from the requesters Bob a worker
from the requesters Bob a worker interacts with 15 people 15 requesters
interacts with 15 people 15 requesters
interacts with 15 people 15 requesters now if he lie in his ratings the person
now if he lie in his ratings the person
now if he lie in his ratings the person score that he’s rating high goes up on
score that he’s rating high goes up on
score that he’s rating high goes up on his task field so if he rates Alice as a
his task field so if he rates Alice as a
his task field so if he rates Alice as a top requester then next Alice’s tasks
top requester then next Alice’s tasks
top requester then next Alice’s tasks are available to the Bob and other
are available to the Bob and other
are available to the Bob and other requesters tasks go at the bottom so in
requesters tasks go at the bottom so in
requesters tasks go at the bottom so in this way Bob lies he’s inviting work
this way Bob lies he’s inviting work
this way Bob lies he’s inviting work from somebody that he doesn’t like or
from somebody that he doesn’t like or
from somebody that he doesn’t like or somebody that he will not pay him
somebody that he will not pay him
somebody that he will not pay him another hand if you’re a requester
another hand if you’re a requester
another hand if you’re a requester posting work to to the workers you’re in
posting work to to the workers you’re in
posting work to to the workers you’re in incentives get high quality work done in
incentives get high quality work done in
incentives get high quality work done in a quick amount of time but if you lie in
a quick amount of time but if you lie in
a quick amount of time but if you lie in your rating you may in white workers who
your rating you may in white workers who
your rating you may in white workers who you don’t like or they are not high
you don’t like or they are not high
you don’t like or they are not high quality so the way tasks are released
quality so the way tasks are released
quality so the way tasks are released the access to the task from Ellis is
the access to the task from Ellis is
the access to the task from Ellis is given to people who Alice’s likes first
given to people who Alice’s likes first
given to people who Alice’s likes first and then other people get the access
and then other people get the access
and then other people get the access later on
later on
later on so in this bi-directional system what’s
so in this bi-directional system what’s
so in this bi-directional system what’s happening is both workers and requesters
happening is both workers and requesters
happening is both workers and requesters have no incentive to lie because if they
have no incentive to lie because if they
have no incentive to lie because if they lie their future is super uncertain but
lie their future is super uncertain but
lie their future is super uncertain but if they tell the truth then they can
if they tell the truth then they can
if they tell the truth then they can make money or they can get their
make money or they can get their
make money or they can get their businesses done so this is called as
businesses done so this is called as
businesses done so this is called as strategy proof mechanisms but in a new
strategy proof mechanisms but in a new
strategy proof mechanisms but in a new angle that is encouraging pro-social
angle that is encouraging pro-social
angle that is encouraging pro-social behavior in these marketplaces now when
behavior in these marketplaces now when
behavior in these marketplaces now when I think about artificial intelligence
I think about artificial intelligence
I think about artificial intelligence and future of words or future of market
and future of words or future of market
and future of words or future of market there are these three things are super
there are these three things are super
there are these three things are super important empathy fairness and equality
important empathy fairness and equality
important empathy fairness and equality and when we are thinking about
and when we are thinking about
and when we are thinking about rust across this dimension we need to
rust across this dimension we need to
rust across this dimension we need to ask a question that can we design human
ask a question that can we design human
ask a question that can we design human center markets that can revolutionize
center markets that can revolutionize
center markets that can revolutionize our future that future deserves more
our future that future deserves more
our future that future deserves more than pennies and creates equal field in
than pennies and creates equal field in
than pennies and creates equal field in the marketplaces for every human being
the marketplaces for every human being
the marketplaces for every human being thank you
thank you
thank you [Applause]
”
Be First to Comment