Jennifer Golbeck: The curly fry conundrum: Why social media "likes" say more than you might think

Prevoditelj: Senzos Osijek Recezent: Stjepan Mateljan

If you remember that first decade of the web,

Ako se sjećate tog prvog desetljeća na Internetu,

it was really a static place.

bilo je to vrlo statično mjesto.

You could go online, you could look at pages,

Mogli ste otići online, mogli ste gledati stranice,

and they were put up either by organizations

a one su bile postavljene od strane organizacija

who had teams to do it

koje su imale timove za to

or by individuals who were really tech-savvy

ili vrlo tehnološki stručni pojedinci

for the time.

za to doba.

And with the rise of social media

S porastom društvenih medija

and social networks in the early 2000s,

i društvenih mreža ranih 2000 - tih,

the web was completely changed

Internet se u potpunosti promijenio

to a place where now the vast majority of content

u mjesto gdje velika većina sadržaja

we interact with is put up by average users,

s kojim dolazimo u kontakt je postavljena od strane prosječnih korisnika,

either in YouTube videos or blog posts

u YouTube videima ili objavama na blogu

or product reviews or social media postings.

ili recenzijama proizvoda ili objavama društvenih medija.

And it's also become a much more interactive place,

Također je postao puno više interaktivno mjesto,

where people are interacting with others,

gdje ljudi uzajamno djeluju,

they're commenting, they're sharing,

komentiraju, dijele,

they're not just reading.

nije da samo čitaju.

So Facebook is not the only place you can do this,

Facebook nije jedino mjesto gdje to možete činiti,

but it's the biggest,

ali je najveće,

and it serves to illustrate the numbers.

i služi za ilustriranje podataka.

Facebook has 1.2 billion users per month.

Facebook ima 1.2 milijardu korisnika na mjesec.

So half the Earth's Internet population

Stoga pola Zemljine internetske populacije

is using Facebook.

koristi Facebook.

They are a site, along with others,

Oni su stranica, zajedno s drugima,

that has allowed people to create an online persona

koja je dozvolila ljudima da stvore online personu

with very little technical skill,

s vrlo malo tehničkih vještina,

and people responded by putting huge amounts

a ljudi su na to odgovorili postavljanjem

of personal data online.

gomile osobnih podataka na mrežu.

So the result is that we have behavioral,

Rezultat je da imamo podatke o ponašanju,

preference, demographic data

podatke o sklonostima, demografske podatke,

for hundreds of millions of people,

za stotine milijuna ljudi,

which is unprecedented in history.

što je jedinstven slučaj u povijesti.

And as a computer scientist, what this means is that

Kao računalnome znanstveniku, to znači da

I've been able to build models

sam mogla napraviti modele

that can predict all sorts of hidden attributes

koji mogu predvidjeti razne skrivene osobine

for all of you that you don't even know

za sve vas, a da ni ne znate

you're sharing information about.

da dijelite informacije o tome.

As scientists, we use that to help

Kao znanstvenici, koristimo to kao pomoć

the way people interact online,

u načinu kako ljudi komuniciraju online,

but there's less altruistic applications,

ali ima i manje altruističnih primjena,

and there's a problem in that users don't really

a problem je u tome što korisnici

understand these techniques and how they work,

ne razumiju baš te tehnike i kako one rade,

and even if they did, they don't have a lot of control over it.

a čak i kada bi razumjeli, nemaju puno kontrole nad time.

So what I want to talk to you about today

Ono o čemu želim pričati danas

is some of these things that we're able to do,

su neke stvari koje možemo napraviti,

and then give us some ideas of how we might go forward

i onda dati neke ideje kako bismo mogli napredovati

to move some control back into the hands of users.

da vratimo dio kontrole natrag korisnicima.

So this is Target, the company.

Ovo je tvrtka Target.

I didn't just put that logo

Nisam ja stavila taj logo

on this poor, pregnant woman's belly.

na trbuh ove jadne, trudne žene.

You may have seen this anecdote that was printed

Možda ste čuli ovu anegdotu

in Forbes magazine where Target

objavljenu u časopisu Forbes;

sent a flyer to this 15-year-old girl

Target je poslao letak 15 godišnjoj djevojci

with advertisements and coupons

s reklamama i kuponima

for baby bottles and diapers and cribs

za dječje bočice, pelene i kolijevke

two weeks before she told her parents

dva tjedna prije no što je rekla

that she was pregnant.

roditeljima da je trudna.

Yeah, the dad was really upset.

Da, otac je bio zbilja uzrujan.

He said, "How did Target figure out

Rekao je, “Kako je Target saznao

that this high school girl was pregnant

da je ova srednjoškolka trudna

before she told her parents?"

prije nego što je ona rekla roditeljima?”

It turns out that they have the purchase history

Ispostavilo se da imaju prošlost kupovine

for hundreds of thousands of customers

za stotine tisuća kupaca

and they compute what they call a pregnancy score,

te izračunavaju ono što oni nazivaju domet trudnoće,

which is not just whether or not a woman's pregnant,

koji govori ne samo je li žena trudna,

but what her due date is.

već i koji joj je termin porođaja.

And they compute that

Oni to računaju

not by looking at the obvious things,

ne gledajući samo očite stvari

like, she's buying a crib or baby clothes,

poput: kupuje li kolijevku ili odjeću za bebe,

but things like, she bought more vitamins

nego i stvari kao što su: kupila je

than she normally had,

više vitamina no obično,

or she bought a handbag

ili kupila je ručnu torbu

that's big enough to hold diapers.

u koju stanu pelene.

And by themselves, those purchases don't seem

Same po sebi, te se kupovine

like they might reveal a lot,

ne čine kao da otkrivaju mnogo,

but it's a pattern of behavior that,

ali uzorak takvog ponašanja,

when you take it in the context of thousands of other people,

kada ga se stavi u kontekst tisuća drugih ljudi,

starts to actually reveal some insights.

zapravo otkriva neka shvaćanja.

So that's the kind of thing that we do

To je način na koji predviđamo

when we're predicting stuff about you on social media.

stvari o vama u društvenim medijima.

We're looking for little patterns of behavior that,

Tražimo male obrasce ponašanja koji,

when you detect them among millions of people,

kada ih otkrijete među milijunima ljudi,

lets us find out all kinds of things.

otkrivaju nam razne stvari.

So in my lab and with colleagues,

U laboratoriju s mojim kolegama,

we've developed mechanisms where we can

razvili smo mehanizme pomoću kojih možemo

quite accurately predict things

prilično precizno predvidjeti stvari

like your political preference,

poput vaše političke preferencije,

your personality score, gender, sexual orientation,

tipa osobnosti, spol, seksualnu orijentaciju,

religion, age, intelligence,

vjeru, dob, inteligenciju,

along with things like

zajedno sa stvarima poput

how much you trust the people you know

koliko vjerujete ljudima koje poznajete

and how strong those relationships are.

i koliko su jake te veze.

We can do all of this really well.

Sve to jako dobro radimo.

And again, it doesn't come from what you might

I opet, to ne dolazi od onoga

think of as obvious information.

što vam se čini kao očiti izvor informacija.

So my favorite example is from this study

Moj najdraži primjer iz ovog je istraživanja

that was published this year

koje je objavljeno ove godine

in the Proceedings of the National Academies.

u broju časopisa Postupci Nacionalne Akademije S.A.D.

If you Google this, you'll find it.

Ako ga guglate, pronaći ćete ga.

It's four pages, easy to read.

Ima četiri stranice, lak je za čitanje.

And they looked at just people's Facebook likes,

Gledali su samo “lajkove” na Facebooku,

so just the things you like on Facebook,

samo one stvari koje vam se sviđaju,

and used that to predict all these attributes,

te su ih koristili za predviđanje ovih obilježja,

along with some other ones.

zajedno s još nekim drugima.

And in their paper they listed the five likes

U svojem radu nabrojali su pet “lajkova“

that were most indicative of high intelligence.

koji su najviše ukazivali na visok stupanj inteligencije.

And among those was liking a page

A među njima bila je stranica

for curly fries. (Laughter)

za kovrčave krumpiriće. (Smijeh)

Curly fries are delicious,

Kovrčavi krumpirići zbilja su ukusni,

but liking them does not necessarily mean

no ako ste ih označili da vam se sviđaju,

that you're smarter than the average person.

ne znači da ste pametniji od prosječne osobe.

So how is it that one of the strongest indicators

Stoga, kako da je jedan od snažnijih indikatora

of your intelligence

vaše inteligencije

is liking this page

“lajkanje” stranice

when the content is totally irrelevant

čiji je sadržaj u potpunosti ireleventan

to the attribute that's being predicted?

za atribut koji predviđa?

And it turns out that we have to look at

Ispostavlja se da moramo sagledati

a whole bunch of underlying theories

golemi broj temeljnih teorija

to see why we're able to do this.

da bismo mogli vidjeti zašto možemo to predvidjeti.

One of them is a sociological theory called homophily,

Jedna od njih je sociološka teorija imena homofilija,

which basically says people are friends with people like them.

koja u biti govori da su ljudi prijatelji s onima koji su im slični.

So if you're smart, you tend to be friends with smart people,

Stoga, ako ste pametni, težite biti prijatelj s takvima

and if you're young, you tend to be friends with young people,

ako ste mladi, težite biti prijatelj s mladim ljudima,

and this is well established

i to je uzorak koji je utvrđen

for hundreds of years.

već stotinama godina.

We also know a lot

Također znamo mnogo o tome

about how information spreads through networks.

kako se informacije šire mrežama.

It turns out things like viral videos

Ispostavlja se da se stvari poput viralnih videa,

or Facebook likes or other information

“lajkova” na Facebooku ili drugih informacija

spreads in exactly the same way

šire na potpuno jednak način

that diseases spread through social networks.

kako se zarazne bolesti šire, ali kroz društvene mreže.

So this is something we've studied for a long time.

To je nešto što smo proučavali dugo vremena.

We have good models of it.

Imamo dobre modele tog obrasca.

And so you can put those things together

Možete posložiti pojave poput te

and start seeing why things like this happen.

i uvidjeti zašto se takve stvari događaju.

So if I were to give you a hypothesis,

Pa, ako bi vam dala hipotezu,

it would be that a smart guy started this page,

ona bi glasila da je pametan tip pokrenuo tu stranicu,

or maybe one of the first people who liked it

ili nekoliko prvih ljudi koji su ju “lajkali”

would have scored high on that test.

su postigli visoke rezultate na ispitu inteligencije.

And they liked it, and their friends saw it,

Svidjela im se, njihovi su je prijatelji vidjeli,

and by homophily, we know that he probably had smart friends,

i prema načelu homofilije znamo da su vjerojatno imali pametne prijatelje

and so it spread to them, and some of them liked it,

pa se to proširilo do njih, neki od njih su ju “lajkali”,

and they had smart friends,

a i oni su imali pametne prijatelje,

and so it spread to them,

and so it propagated through the network

pa se proširilo do njih,

tako se propagira kroz mrežu

to a host of smart people,

koristeći pametne ljude kao domaćine,

so that by the end, the action

stoga je na kraju čin sviđanja

of liking the curly fries page

stranice kovrčavih krumpirića

is indicative of high intelligence,

postao indikacija visoke inteligencije,

not because of the content,

ne zbog njezina sadržaja,

but because the actual action of liking

već zato što sam pritisak na gumb sviđanja

reflects back the common attributes

zrcali učestale atribute

of other people who have done it.

ljudi koji su ga pritisnuli.

So this is pretty complicated stuff, right?

Ovo su poprilično komplicirane stvari, jelda?

It's a hard thing to sit down and explain

Teško je sjesti i objasniti to

to an average user, and even if you do,

prosječnom korisniku, no čak i ako to učinite,

what can the average user do about it?

što može prosječan korisnik u vezi toga napraviti?

How do you know that you've liked something

Kako znate da nešto što ste “lajkali”

that indicates a trait for you

indicira vašu osobinu

that's totally irrelevant to the content of what you've liked?

koja je u potpunosti nepovezana sa sadržajem onoga što ste “lajkali”?

There's a lot of power that users don't have

Postoji puno snage koju korisnici nemaju

to control how this data is used.

za kontrolu kako se koriste ti podatci.

And I see that as a real problem going forward.

i vidim to kao velik problem gledajući unaprijed.

So I think there's a couple paths

Mislim da postoji nekoliko načina

that we want to look at

koje želimo sagledati

if we want to give users some control

ako bismo korisnicima donekle dali kontrolu

over how this data is used,

nad time kako se koriste ti podatci,

because it's not always going to be used

jer neće uvijek biti iskorišteni

for their benefit.

u njihovu korist.

An example I often give is that,

Primjer koji često dajem jest,

if I ever get bored being a professor,

ako mi ikad dosadi biti profesor,

I'm going to go start a company

pokrenut ću kompaniju

that predicts all of these attributes

koja predviđa sve te atribute

and things like how well you work in teams

i stvari kao što su, radite li dobro u timu,

and if you're a drug user, if you're an alcoholic.

koristite li droge, jeste li alkoholičar.

We know how to predict all that.

Sve to znamo predvidjeti.

And I'm going to sell reports

I prodavat ću izvješća

to H.R. companies and big businesses

kompanijama ljudskih resursa i velikim tvrtkama

that want to hire you.

koje vas žele zaposliti.

We totally can do that now.

Mi to možemo napraviti.

I could start that business tomorrow,

Mogla bih pokrenuti taj posao sutra,

and you would have absolutely no control

a vi ne biste imali apsolutno nikakvu kontrolu

over me using your data like that.

nada mnom koja koristim vaše podatke na taj način.

That seems to me to be a problem.

To mi se čini kao problem.

So one of the paths we can go down

Jedan put kojim možemo poći

is the policy and law path.

je put politike i prava.

And in some respects, I think that that would be most effective,

I u nekim pogledima, mislim da bi to bilo najučinkovitije,

but the problem is we'd actually have to do it.

no problem je da bismo to zbilja morali napraviti.

Observing our political process in action

Promatrajući naše političke procese u akciji

makes me think it's highly unlikely

čini mi se da je malo vjerojatno

that we're going to get a bunch of representatives

da ćemo dobiti gomilu predstavnika

to sit down, learn about this,

koji će sjesti, naučiti o tome

and then enact sweeping changes

i onda donijeti dalekosežne promjene

to intellectual property law in the U.S.

u američke zakone o intelektualnom vlasništvu

so users control their data.

da bi korisnici mogli kontrolirati svoje podatke.

We could go the policy route,

Možemo krenuti putem politike,

where social media companies say,

gdje kompanije društvenih medija kažu:

you know what? You own your data.

znate što? Vi posjedujete podatke.

You have total control over how it's used.

Imate potpunu kontrolu nad time kako se koriste.

The problem is that the revenue models

Problem je u tome što se modeli prihoda

for most social media companies

za većinu kompanija društvenih medija

rely on sharing or exploiting users' data in some way.

oslanjaju na dijeljenje ili, na neki način, iskorištavanje podataka korisnika.

It's sometimes said of Facebook that the users

Ponekad se kaže za Facebook da korisnici

aren't the customer, they're the product.

nisu mušterije, već proizvod.

And so how do you get a company

Stoga kako natjerati kompaniju

to cede control of their main asset

da korisnicima vrati kontrolu

back to the users?

nad njihovom glavnom imovinom?

It's possible, but I don't think it's something

Moguće je, ali ne mislim da je to promjena

that we're going to see change quickly.

koju ćemo ubrzo vidjeti.

So I think the other path

Stoga mislim da će drugi put

that we can go down that's going to be more effective

kojim možemo poći biti djelotvorniji

is one of more science.

i to je onaj sa više znanosti.

It's doing science that allowed us to develop

Prije svega, znanost nam dopušta razvijanje

all these mechanisms for computing

svih ovih mehanizama za računanje

this personal data in the first place.

osobnih podataka.

And it's actually very similar research

I zapravo bismo morali provesti

that we'd have to do

vrlo slično istraživanje

if we want to develop mechanisms

ako bismo htjeli razviti mehanizme

that can say to a user,

koji bi korisniku rekli:

"Here's the risk of that action you just took."

“Ovo je rizik radnje koju ste upravo napravili.”

By liking that Facebook page,

Lajkanjem te Facebook stranice

or by sharing this piece of personal information,

ili dijeljenjem ove osobne informacije,

you've now improved my ability

upravo ste poboljšali moju sposobnost

to predict whether or not you're using drugs

predviđanja koristite li droge,

or whether or not you get along well in the workplace.

ili slažete li se na radnom mjestu.

And that, I think, can affect whether or not

A to mislim da bi moglo

people want to share something,

utjecati na želju ljudi da nešto podijele,

keep it private, or just keep it offline altogether.

zadrže to privatnim, ili uopće ne stave online.

We can also look at things like

Možemo se također osvrnuti na stvari

allowing people to encrypt data that they upload,

poput dopuštanja ljudima da kodiraju podatke koje su objavili,

so it's kind of invisible and worthless

pa oni postaju nevidljivi i beskorisni

to sites like Facebook

stranicama poput Facebooka

or third party services that access it,

ili uslugama trećih stranaka koje im imaju pristup,

but that select users who the person who posted it

ali da odabrani korisnici

want to see it have access to see it.

za koje vlasnici žele da vide informacije ih mogu i vidjeti.

This is all super exciting research

Ovo je vrlo uzbudljivo istraživanje

from an intellectual perspective,

iz intelektualne perspektive,

and so scientists are going to be willing to do it.

stoga će ga znanstvenici biti voljni napraviti.

So that gives us an advantage over the law side.

To nam daje prednost pred zakonskom stranom.

One of the problems that people bring up

Jedan od problema koji ljudi spomenu

when I talk about this is, they say,

kada pričam o ovome jest, kažu,

you know, if people start keeping all this data private,

znate, ako ljudi počnu zadržavati sve svoje podatke privatnima,

all those methods that you've been developing

sve ove metode koje ste razvijali

to predict their traits are going to fail.

za predviđanje njihovih osobina će propasti.

And I say, absolutely, and for me, that's success,

A ja kažem, apsolutno, i za mene je to uspjeh,

because as a scientist,

jer kao znanstveniku,

my goal is not to infer information about users,

moj cilj nije dolaziti do informacija o korisnicima,

it's to improve the way people interact online.

već poboljšati način kako ljudi međusobno komuniciraju online.

And sometimes that involves inferring things about them,

Ponekad to uključuje saznavanje njihovih informacija,

but if users don't want me to use that data,

ali ako korisnici ne žele da koristim te podatke,

I think they should have the right to do that.

mislim da bi trebali imati to pravo i učiniti.

I want users to be informed and consenting

Želim da korisnici budu informirani

users of the tools that we develop.

i da prihvaćaju alate koje razvijamo.

And so I think encouraging this kind of science

Mislim da ohrabrivanje ove vrste znanosti

and supporting researchers

i podupiranje istraživača

who want to cede some of that control back to users

koji žele povratiti dio kontrole korisnicima

and away from the social media companies

od kompanija društvenih medija

means that going forward, as these tools evolve

znači da idući naprijed, kako se ova oruđa razvijaju

and advance,

i napreduju,

means that we're going to have an educated

imat ćemo obrazovanu

and empowered user base,

i osposobljenu bazu korisnika,

and I think all of us can agree

te mislim da se svi možemo složiti

that that's a pretty ideal way to go forward.

da je to prilično idealan način za napredovanje.

Thank you.

Hvala.

(Applause)

(Pljesak)