Press "Enter" to skip to content

How computers predict us? | Nimrod Kozlovski | TEDxTelAviv


when I was 16 I had a date I get to a

when I was 16 I had a date I get to a house the door opens and it’s a father

house the door opens and it’s a father

house the door opens and it’s a father the father sits me down at the living

the father sits me down at the living

the father sits me down at the living room couch and starts bombarding me with

room couch and starts bombarding me with

room couch and starts bombarding me with questions what’s the last book you read

questions what’s the last book you read

questions what’s the last book you read what are your parents do what is your

what are your parents do what is your

what are your parents do what is your favorite subject in school your daughter

favorite subject in school your daughter

favorite subject in school your daughter at the moment and it just won’t stop

at the moment and it just won’t stop

at the moment and it just won’t stop what are your hobbies do you have a yard

what are your hobbies do you have a yard

what are your hobbies do you have a yard about an hour later the father stands up

about an hour later the father stands up

about an hour later the father stands up in mumbles I stayed up for no reason you

in mumbles I stayed up for no reason you

in mumbles I stayed up for no reason you don’t look like a bad kid and he called

don’t look like a bad kid and he called

don’t look like a bad kid and he called her downstairs what I went through that

her downstairs what I went through that

her downstairs what I went through that night was an elaborate algorithmic

night was an elaborate algorithmic

night was an elaborate algorithmic questioning the algorithm originator the

questioning the algorithm originator the

questioning the algorithm originator the father probably profiled in his mind

father probably profiled in his mind

father probably profiled in his mind some kind of a dangerous yard less

some kind of a dangerous yard less

some kind of a dangerous yard less maniac and he was trying to way to their

maniac and he was trying to way to their

maniac and he was trying to way to their potential risk of their shitty to his

potential risk of their shitty to his

potential risk of their shitty to his beloved daughter people always try to

beloved daughter people always try to

beloved daughter people always try to assess what is the next move of an

assess what is the next move of an

assess what is the next move of an unpredictable stranger computers on the

unpredictable stranger computers on the

unpredictable stranger computers on the other hand are totally predictable they

other hand are totally predictable they

other hand are totally predictable they live in a binary universe zeros and ones

live in a binary universe zeros and ones

live in a binary universe zeros and ones but in our world something very strange

but in our world something very strange

but in our world something very strange happened computers predict humans not

happened computers predict humans not

happened computers predict humans not the other way around as much as we think

the other way around as much as we think

the other way around as much as we think of ourself as unique one-of-a-kind

of ourself as unique one-of-a-kind

of ourself as unique one-of-a-kind exceptional we can be coded by what I

exceptional we can be coded by what I

exceptional we can be coded by what I call the algorithmic Looking Glass

computers predicting humans a man rents

computers predicting humans a man rents an apartment using Airbnb how can we

an apartment using Airbnb how can we

an apartment using Airbnb how can we tell if he’s not going to win the

tell if he’s not going to win the

tell if he’s not going to win the apartment by hosting a huge party let

apartment by hosting a huge party let

apartment by hosting a huge party let the computers

the computers

the computers how do computers do that it’s a set of a

how do computers do that it’s a set of a

how do computers do that it’s a set of a prefix computer steps digging to

prefix computer steps digging to

prefix computer steps digging to enormous amount of data concerning us

enormous amount of data concerning us

enormous amount of data concerning us and extracting insights algorithms

and extracting insights algorithms

and extracting insights algorithms operate in the background of our life

few years ago I was invited to a

few years ago I was invited to a convention in Qatar Doha I doubted that

convention in Qatar Doha I doubted that

convention in Qatar Doha I doubted that the Qataris would let me in an Israeli

the Qataris would let me in an Israeli

the Qataris would let me in an Israeli but surprisingly I get a signed visa

but surprisingly I get a signed visa

but surprisingly I get a signed visa handwritten by the Emile’s daughter I’m

handwritten by the Emile’s daughter I’m

handwritten by the Emile’s daughter I’m going to Qatar but the one entity I

going to Qatar but the one entity I

going to Qatar but the one entity I thought would really suspicious about me

thought would really suspicious about me

thought would really suspicious about me traveling to Qatar is the credit card

traveling to Qatar is the credit card

traveling to Qatar is the credit card company I thought the credit card

company I thought the credit card

company I thought the credit card company would block my card right away

company would block my card right away

company would block my card right away so I prepared for my travel

so I prepared for my travel

so I prepared for my travel I took a money belt and Heidi was in my

I took a money belt and Heidi was in my

I took a money belt and Heidi was in my socks I travelled there and everything

socks I travelled there and everything

socks I travelled there and everything works just fine my card walk just as

works just fine my card walk just as

works just fine my card walk just as usual I came back home and asked a

usual I came back home and asked a

usual I came back home and asked a friend of mine work at the credit card

friend of mine work at the credit card

friend of mine work at the credit card company how come Nimrod he said you’ve

company how come Nimrod he said you’ve

company how come Nimrod he said you’ve done nothing suspicious

done nothing suspicious

done nothing suspicious you flew oil jordanian like you always

you flew oil jordanian like you always

you flew oil jordanian like you always do at the airport you ate the shwarma in

do at the airport you ate the shwarma in

do at the airport you ate the shwarma in ten minutes later as always totally

ten minutes later as always totally

ten minutes later as always totally regretful you add another shwarma you

regretful you add another shwarma you

regretful you add another shwarma you shopped at the local alternative forget

shopped at the local alternative forget

shopped at the local alternative forget we know your style the ole time there

we know your style the ole time there

we know your style the ole time there were no transactions in Israel so your

were no transactions in Israel so your

were no transactions in Israel so your profile was perfectly normal for you

profile was perfectly normal for you

profile was perfectly normal for you that’s what algorithms try to do for us

that’s what algorithms try to do for us

that’s what algorithms try to do for us they try to understand what are the

they try to understand what are the

they try to understand what are the patterns of our normal behavior and try

patterns of our normal behavior and try

patterns of our normal behavior and try to sort out anomalies one thing that

to sort out anomalies one thing that

to sort out anomalies one thing that could really change our life is if

could really change our life is if

could really change our life is if algorithms could predict crime let me

algorithms could predict crime let me

algorithms could predict crime let me take you to fast future attribute

take you to fast future attribute

take you to fast future attribute screening technology

screening technology

screening technology it is better known by its original

it is better known by its original

it is better known by its original American name the malicious intense

American name the malicious intense

American name the malicious intense detection program it is a complex

detection program it is a complex

detection program it is a complex portable lab that’s supposed to assess

portable lab that’s supposed to assess

portable lab that’s supposed to assess how dangerous people are in public

how dangerous people are in public

how dangerous people are in public spaces streets terminal Federal Building

spaces streets terminal Federal Building

spaces streets terminal Federal Building it’s a complex set of long-distance

it’s a complex set of long-distance

it’s a complex set of long-distance cameras and sensors all of them trying

cameras and sensors all of them trying

cameras and sensors all of them trying to collect information about people

to collect information about people

to collect information about people passing by what kind of information

passing by what kind of information

passing by what kind of information basically everything we got pulls bread

basically everything we got pulls bread

basically everything we got pulls bread weight body heat muscle stiffness sweat

weight body heat muscle stiffness sweat

weight body heat muscle stiffness sweat circles even the pitch of our voice is

circles even the pitch of our voice is

circles even the pitch of our voice is something they collect why because we

something they collect why because we

something they collect why because we have an assumption that those body

have an assumption that those body

have an assumption that those body signals connect to malicious intent and

signals connect to malicious intent and

signals connect to malicious intent and the technology is just getting better

the technology is just getting better

the technology is just getting better and better imagine for instance an

and better imagine for instance an

and better imagine for instance an artificial nose that can replace a dog

artificial nose that can replace a dog

artificial nose that can replace a dog that was trained to sense for the smell

that was trained to sense for the smell

that was trained to sense for the smell of fear or a long-distance camera that

of fear or a long-distance camera that

of fear or a long-distance camera that can manage to track from afar a person

can manage to track from afar a person

can manage to track from afar a person that is avoiding eye contact what the

that is avoiding eye contact what the

that is avoiding eye contact what the technology is doing it is trying to

technology is doing it is trying to

technology is doing it is trying to imitate how we as human beings have our

imitate how we as human beings have our

imitate how we as human beings have our senses and intuitions in assessing for

senses and intuitions in assessing for

senses and intuitions in assessing for danger but the technology does it more

danger but the technology does it more

danger but the technology does it more accurate and 24/7 at the same time there

accurate and 24/7 at the same time there

accurate and 24/7 at the same time there is another complex technology called

is another complex technology called

is another complex technology called video analysis it looks at the scene and

video analysis it looks at the scene and

video analysis it looks at the scene and trying to understand whether there’s a

trying to understand whether there’s a

trying to understand whether there’s a risk in a certain scenario how can we

risk in a certain scenario how can we

risk in a certain scenario how can we tell a person is running towards another

tell a person is running towards another

tell a person is running towards another person back could we tell if he’s going

person back could we tell if he’s going

person back could we tell if he’s going to attack him or just hug an old friend

to attack him or just hug an old friend

to attack him or just hug an old friend the computer would know after time

the computer would know after time

the computer would know after time learning from past videos any fallouts

learning from past videos any fallouts

learning from past videos any fallouts go on the system instantly would connect

go on the system instantly would connect

go on the system instantly would connect information look for information about

information look for information about

information look for information about the suspect in governmental databases

the suspect in governmental databases

the suspect in governmental databases perhaps even private ones

perhaps even private ones

perhaps even private ones he still scores as whiskey step aside

he still scores as whiskey step aside

he still scores as whiskey step aside sir okay that’s fine

sir okay that’s fine

sir okay that’s fine it probably breaks millions of privacy

it probably breaks millions of privacy

it probably breaks millions of privacy expectations we have but let’s just

expectations we have but let’s just

expectations we have but let’s just assume that we are always one bombing

assume that we are always one bombing

assume that we are always one bombing away from giving up our freedoms it

away from giving up our freedoms it

away from giving up our freedoms it reminds me of an Iranian stand-up

reminds me of an Iranian stand-up

reminds me of an Iranian stand-up comedian I met once in New York he said

comedian I met once in New York he said

comedian I met once in New York he said I just finished the luckiest day of my

I just finished the luckiest day of my

I just finished the luckiest day of my life out of 11 times I traveled to the

life out of 11 times I traveled to the

life out of 11 times I traveled to the airport

airport

airport 11 times I was randomly selected for a

11 times I was randomly selected for a

11 times I was randomly selected for a security check but guess what

security check but guess what

security check but guess what even that doesn’t make us predictable

even that doesn’t make us predictable

even that doesn’t make us predictable enough for computers because most action

enough for computers because most action

enough for computers because most action in our life do not have real-time

in our life do not have real-time

in our life do not have real-time information and we don’t live in the

information and we don’t live in the

information and we don’t live in the terminal most of us so the technology

terminal most of us so the technology

terminal most of us so the technology cannot be used everywhere or it just

cannot be used everywhere or it just

cannot be used everywhere or it just tends to make too many mistakes so the

tends to make too many mistakes so the

tends to make too many mistakes so the system does something else to make up

system does something else to make up

system does something else to make up for that the systems we have try to

for that the systems we have try to

for that the systems we have try to assess how dangerous how risky a person

assess how dangerous how risky a person

assess how dangerous how risky a person is and not how dangerous the action

is and not how dangerous the action

is and not how dangerous the action itself we assess people there’s a TV

itself we assess people there’s a TV

itself we assess people there’s a TV show called blackmailer

show called blackmailer

show called blackmailer in the TV show black mirror there’s an

in the TV show black mirror there’s an

in the TV show black mirror there’s an episode called nosedive in this episode

episode called nosedive in this episode

episode called nosedive in this episode it describes an imaginary ward where

it describes an imaginary ward where

it describes an imaginary ward where everybody has a cute symbol from 1 to 5

everybody has a cute symbol from 1 to 5

everybody has a cute symbol from 1 to 5 that describes him our rating people are

that describes him our rating people are

that describes him our rating people are aware of the rating and obsessed with it

aware of the rating and obsessed with it

aware of the rating and obsessed with it they try to satisfy the algorithm every

they try to satisfy the algorithm every

they try to satisfy the algorithm every action they do in their life immediately

action they do in their life immediately

action they do in their life immediately impacts the rating and in turn it impact

impacts the rating and in turn it impact

impacts the rating and in turn it impact their life if you have seen the show you

their life if you have seen the show you

their life if you have seen the show you must say ok it’s a science fiction it’s

must say ok it’s a science fiction it’s

must say ok it’s a science fiction it’s a TV show

it is more realistic than most of us

it is more realistic than most of us think it is already happening it is

think it is already happening it is

think it is already happening it is already happening in China in China

already happening in China in China

already happening in China in China there’s a great movement of people

there’s a great movement of people

there’s a great movement of people migrating to the big cities let’s create

migrating to the big cities let’s create

migrating to the big cities let’s create a need for credit economy so the Chinese

a need for credit economy so the Chinese

a need for credit economy so the Chinese started the project similar to something

started the project similar to something

started the project similar to something we know in every western country credits

we know in every western country credits

we know in every western country credits going but the problem is that the most

going but the problem is that the most

going but the problem is that the most of the villagers in China they never had

of the villagers in China they never had

of the villagers in China they never had a bank account or they never had payment

a bank account or they never had payment

a bank account or they never had payment history so how can you decide who to

history so how can you decide who to

history so how can you decide who to lend money to or to end out a flat so

lend money to or to end out a flat so

lend money to or to end out a flat so the Chinese government decided to do

the Chinese government decided to do

the Chinese government decided to do something different they approached the

something different they approached the

something different they approached the technology companies in China and they

technology companies in China and they

technology companies in China and they asked them to look at the information

asked them to look at the information

asked them to look at the information they have mobile internet social

they have mobile internet social

they have mobile internet social purchase activities of their users and

purchase activities of their users and

purchase activities of their users and come up with social scoring you you buy

come up with social scoring you you buy

come up with social scoring you you buy diapers online you’re probably a good

diapers online you’re probably a good

diapers online you’re probably a good family man you have a lot of friends

family man you have a lot of friends

family man you have a lot of friends online that you play online games with

online that you play online games with

online that you play online games with for 10 hours straight not so good you

for 10 hours straight not so good you

for 10 hours straight not so good you shared an aggressive post you can see

shared an aggressive post you can see

shared an aggressive post you can see awaiting takes a nosedive at the moment

awaiting takes a nosedive at the moment

awaiting takes a nosedive at the moment the leading project called sesame is not

the leading project called sesame is not

the leading project called sesame is not a must but the Chinese government plans

a must but the Chinese government plans

a must but the Chinese government plans in three years by 2020 that every

in three years by 2020 that every

in three years by 2020 that every citizen in China will have his score a

citizen in China will have his score a

citizen in China will have his score a number between 350 and 750 a number for

number between 350 and 750 a number for

number between 350 and 750 a number for each citizen is call people that will

each citizen is call people that will

each citizen is call people that will have high score

have high score

have high score they’ll have great life they will be

they’ll have great life they will be

they’ll have great life they will be able to walk in governmental jobs in

able to walk in governmental jobs in

able to walk in governmental jobs in China and they will be able to get an

China and they will be able to get an

China and they will be able to get an extended line of credit people with low

extended line of credit people with low

extended line of credit people with low scores not so good

scores not so good

scores not so good people with low scores would not be able

people with low scores would not be able

people with low scores would not be able to get passport and travel abroad and

to get passport and travel abroad and

to get passport and travel abroad and they would not be able to sleep in

they would not be able to sleep in

they would not be able to sleep in luxurious hotels and it’s not only that

luxurious hotels and it’s not only that

luxurious hotels and it’s not only that your waiting is expected to drop or rise

your waiting is expected to drop or rise

your waiting is expected to drop or rise depending on who your friends or family

depending on who your friends or family

depending on who your friends or family are hey you look around do you feel

are hey you look around do you feel

are hey you look around do you feel comfortable sitting with someone that

comfortable sitting with someone that

comfortable sitting with someone that has a score of 350 – have you thought

has a score of 350 – have you thought

has a score of 350 – have you thought how your score will be reflected by his

how your score will be reflected by his

how your score will be reflected by his code let’s be honest we do not predict

code let’s be honest we do not predict

code let’s be honest we do not predict we profile people we do not predict we

we profile people we do not predict we

we profile people we do not predict we profile people

for filing and waiting systems are going

for filing and waiting systems are going to be a growing part of our lives and

to be a growing part of our lives and

to be a growing part of our lives and it’s not only China they’ll be

it’s not only China they’ll be

it’s not only China they’ll be everywhere they are efficient and they

everywhere they are efficient and they

everywhere they are efficient and they are needed and the technology they are

are needed and the technology they are

are needed and the technology they are based on predictive analytics is already

based on predictive analytics is already

based on predictive analytics is already mature enough it’s already marketed as

mature enough it’s already marketed as

mature enough it’s already marketed as off-the-shelf software you can buy two

off-the-shelf software you can buy two

off-the-shelf software you can buy two the same way you buy Microsoft Office so

the same way you buy Microsoft Office so

the same way you buy Microsoft Office so soon every entity private or public will

soon every entity private or public will

soon every entity private or public will start waiting us they will start

start waiting us they will start

start waiting us they will start creating risk models and try to see how

creating risk models and try to see how

creating risk models and try to see how we score awaited with these risk models

we score awaited with these risk models

we score awaited with these risk models every entity your government tinder eBay

every entity your government tinder eBay

every entity your government tinder eBay uber everyone every entity will score

uber everyone every entity will score

uber everyone every entity will score and wave you there will be algorithmic

and wave you there will be algorithmic

and wave you there will be algorithmic lookingglass wherever you go in the last

lookingglass wherever you go in the last

lookingglass wherever you go in the last 15 years

15 years

15 years I personally devoted my time to study

I personally devoted my time to study

I personally devoted my time to study the relationship we have with algorithms

the relationship we have with algorithms

the relationship we have with algorithms well not exactly 15 years I will seek

well not exactly 15 years I will seek

well not exactly 15 years I will seek for a day the most important observation

for a day the most important observation

for a day the most important observation to me came from a group of fellows at

to me came from a group of fellows at

to me came from a group of fellows at Yale who had a sticker on their computer

Yale who had a sticker on their computer

Yale who had a sticker on their computer saying do you want your life be dictated

saying do you want your life be dictated

saying do you want your life be dictated by an algorithm don’t you think you are

by an algorithm don’t you think you are

by an algorithm don’t you think you are entitled to know that the computer made

entitled to know that the computer made

entitled to know that the computer made a decision about you and not a person

a decision about you and not a person

a decision about you and not a person and what the decision is based on and

and what the decision is based on and

and what the decision is based on and where did you get the information about

where did you get the information about

where did you get the information about you form a new let it gather all this

you form a new let it gather all this

you form a new let it gather all this information others might be concerned

information others might be concerned

information others might be concerned with what I call the distorted looking

with what I call the distorted looking

with what I call the distorted looking glass they question the scoring itself

glass they question the scoring itself

glass they question the scoring itself who is behind the algorithm what our is

who is behind the algorithm what our is

who is behind the algorithm what our is interest is agenda

interest is agenda

interest is agenda maybe the designer of the algorithm is a

maybe the designer of the algorithm is a

maybe the designer of the algorithm is a racist and it doesn’t even know can

racist and it doesn’t even know can

racist and it doesn’t even know can someone help you’re waiting on purpose

someone help you’re waiting on purpose

someone help you’re waiting on purpose and if you’re waiting is wrong can you

and if you’re waiting is wrong can you

and if you’re waiting is wrong can you question the decision is there anyone

question the decision is there anyone

question the decision is there anyone you can talk to about you’re waiting is

you can talk to about you’re waiting is

you can talk to about you’re waiting is there anyone there at all to my research

there anyone there at all to my research

there anyone there at all to my research the nature of the opposition became

the nature of the opposition became

the nature of the opposition became clear if algorithms are the new policing

clear if algorithms are the new policing

clear if algorithms are the new policing power we believe they need to answer to

power we believe they need to answer to

power we believe they need to answer to the rules the same way we limit the real

the rules the same way we limit the real

the rules the same way we limit the real policemen and its power the same way we

policemen and its power the same way we

policemen and its power the same way we forbid private businesses from

forbid private businesses from

forbid private businesses from arbitrarily discriminating their

arbitrarily discriminating their

arbitrarily discriminating their customers we believe it’s our right to

customers we believe it’s our right to

customers we believe it’s our right to police the policing algorithms if there

police the policing algorithms if there

police the policing algorithms if there is an algorithmic Looking Glass watching

is an algorithmic Looking Glass watching

is an algorithmic Looking Glass watching observing reflecting you all day today

observing reflecting you all day today

observing reflecting you all day today and you don’t even know it you are in a

and you don’t even know it you are in a

and you don’t even know it you are in a much more situation than those people in

much more situation than those people in

much more situation than those people in China they at least know the ranking and

China they at least know the ranking and

China they at least know the ranking and can understand its logic but we we are

can understand its logic but we we are

can understand its logic but we we are surrounded by so many weighting

surrounded by so many weighting

surrounded by so many weighting profiling scoring algorithms by so many

profiling scoring algorithms by so many

profiling scoring algorithms by so many entities and we don’t even know we may

entities and we don’t even know we may

entities and we don’t even know we may end up trying to play the looking glass

end up trying to play the looking glass

end up trying to play the looking glass try to manipulate the algorithm we may

try to manipulate the algorithm we may

try to manipulate the algorithm we may end up just as any website trying to

end up just as any website trying to

end up just as any website trying to tweak the Google algorithm to be ranked

tweak the Google algorithm to be ranked

tweak the Google algorithm to be ranked higher in a search result even if you

higher in a search result even if you

higher in a search result even if you accept the claim that you can’t figure

accept the claim that you can’t figure

accept the claim that you can’t figure out how it works

out how it works

out how it works because you’re getting its way aren’t

because you’re getting its way aren’t

because you’re getting its way aren’t you still entitled to legal protection

you still entitled to legal protection

you still entitled to legal protection from unfair invasive discriminatory

from unfair invasive discriminatory

from unfair invasive discriminatory arbitrary algorithm that is going to

arbitrary algorithm that is going to

arbitrary algorithm that is going to affect your life

we are at the age of the algorithmic

we are at the age of the algorithmic looking-glass we must take action and

looking-glass we must take action and

looking-glass we must take action and define now what our rights versus the

define now what our rights versus the

define now what our rights versus the algorithmic class we must regulate it

algorithmic class we must regulate it

algorithmic class we must regulate it now otherwise it will be too late thank

now otherwise it will be too late thank

now otherwise it will be too late thank you very much

you very much

you very much [Applause]

”

Be First to Comment

Leave a Reply

Your email address will not be published. Required fields are marked *