AlterConf San Francisco 2016 – Surveying Identity or Surveillance… by Vikram Babu


(upbeat music) – Hi everyone. – [Voiceover] Hi. – So, wait, this is not
the beginning. Sorry. (laughter) – As you can see, there
are 70 slides here. (laughter) – Cool. Everyone looks really good today. Crowd looks really diverse. And, I wonder how the agri-web sort of looks at everyone, how it identifies you through your like various forms that you filled out. So, how does it see you online, and, I suppose if you want
to go by one data point (laughter) it’s probably like the Twitter
social justice warrior list, which means Marc Andreessen
should have dropped in, and if he doesn’t, he doesn’t
really know you guys, so. So, I want to also give an update that originally the talk
was about surveillance, as well, so it was about surveying versus surveillance,
and I want to go back to something Morgan said which is bringing back the complexity of humanity, back on to the web, and I sort of dropped the
whole surveillance part of it. Only because I wanted this to
be about Utopian speculations rather than distopia, and, also for having like reasons of being undocumented in the US during
a time of like xenophobia. I’ve been surveilled, I
know what that’s like, and I think there are bigger
crises than surveillance right now, which is partly like, acknowledging identities, so, there won’t be any talk of surveillance and I don’t think I used the word white in any of the slides, so check me on that. Okay, before we start this presentation, though, if you could just fill out this onboarding form, just pick one or the
other, everyone’s cool, if you can’t pick one, like,
sorry, it doesn’t matter. Now, think of every
other form you filled out and if there’s any, like
what if any useful data, like have you put into those forms? We fill out tons of
forms and it’s just like, name, age, you know,
address, male, female, it’s pretty binary, and so, let’s take my identity, for example, You put it into some
amazingly designed form, stored in binary so we can retrieve, analyze and exchange it, and what do we get, you put all, you know, this quintillian bytes of data, and you get, whoa this is me. 30s, Asian male, maybe some income data, profession, marital status, and that’s with the
volunteered information, there’s also all this user generated data, where I check in, like what I like, voluntary, involuntary data, that’s collected, and there’s some person that all this data correlates to, and I sort of think when
you think of epidemiology, it’s like some epidemic occurs, and there’s a spike in an event, but we know very little about the people that it’s happening to, so it’s like, well, oh cool, we have all this data, but
what does that correlate to? Is there like a genetic condition here, or is there like some
aspect of their personality that’s caused this rise
in diabetes, who knows, but, back to the point, is like, really, is that me? I’m like three points of data, I don’t think so. I’m, this is who I am,
I’m yungrama on Twitter. I’m a writer, designer and entrepreneur. I’m South Indian. I’m Tamil, to be precise, which is one of the
only classical languages that’s still in use today. There’s another Tamilian here. Any other designers in the room? Couple of designers, cool. And, yeah, I’m also a
late night tweeter, so (laughs) that’s me. But, I think, you know,
we collect all this data and we say, it’s, we need to keep it simple, fast or secure, you know, and that’s why we only collect a little bit of data about people, or maybe we mostly just
don’t care about people, we don’t care about our users, so it’s like this, or maybe it’s that most of our forms were designed for like
transactions or advertising, or log ins, so it’s like a very flat, like not very informative
sort of data collection about users. And, you sort of, in this
rather crude binary sampling, you sort of lose a lot of rich metadata about people. You lose, you know, you lose
stuff about the interests, the depths of them and
their specifics about them. Again, you kind of just
have, are you male or female? Is that, you know, is that it? And then there’s, you
know, I think Okcupid kind of does something a bit better. You know, there’s actually
a checkbox for Indian there, it’s a step up from
Asian, and there’s also sapiosexual, that was, like, a shout-out to Sean from Tinder, you know. And, like, they have other sorts of gender options and relationship types, and stuff like that, but, by and large, I think when
you sample people’s identity in a really, like, crude way, you end up with something like this. (laughter) We lose a lot of information about people and we have to guess or resort to bias, so like, a 30 A M is like,
I’m 30s Asian Male, that’s me. Could be anyone, sort of. And, because of that you
strip a person of context, and it’s just a guessing game and you sort of end up with
some really strong biases, propagating from offline to online, so. Also, lately, at Okcupid,
this is how people tend to quick match, based on
like whether they’re male, female, or like, Asian, Latino, black, so, there’s a lot of like, you know, instant judgments and biases, and all that, but, there’s another fact, I’m just throwing this in there, that people tend to find each
other more attractive after they’ve spent time together, so, it’s sort of this weird thing, it’s like, well, I don’t know, swipe left, swipe right, whatever, sapiosexual, a la. (laughter) And, you know, there’s understandably, like there’s good reason to want to not share personal information, like we talk about trust
and safety as a concern and we’re just becoming aware of that, how dangerous it is online. But, I sort of think
the only way out of Hell is through it, so, we can’t just guess who our users are, or
imagine their behaviors. If we do that, we’ll
just start to imagine, they act as if, they act as though, sorry, if we imagine who are users are, we just tend to project
ourselves onto them, right? As designers, especially. Well, like, I don’t really
know anything about them, but I’m sure they like whatever I like, so, let’s just put that in there. And, it’s a problem because, well besides not having data about them, we’re all so, sort of evolutionary, hard-wired, even as infants, to recognize age and gender, race is a bit different, but, I think, by and large, those naturally evolved judgment systems, are sort of out-dated for our
modern social economy system. Some of our like human firmware about how we recognize
and sort of, I guess, identify people that our in our in group and all, is sort of outdated. And for what happens when this sort of pattern recognition meets low sampling, I think we just make
really bad assumptions and we overlook the natural
variants in the world. And we discount people’s inherent value in favor of, like, whatever
the dominant group is doing. Cool, so how’s everyone doin’ so far? I’m just sort of, this is
all stream of consciousness, I don’t know, (laughter) I don’t know how I got to 70 slides. Anyway, so because people
and their identities are not represented in any data format we just sort of disregard them. We make uninformed judgments, we gaslight them when issues arise, and we do that offline as well, and, so, we’re in a sense excluding people from meaningful
conversations and research. Instead, you know, we build out these really
impressive organizations, we’ve got all these
employee and user metrics that drive performance
towards impressive results. And, because, you know, we live in this post-racial secular
society, and all that fluff, and then what happens, Erica Bayko passes around a spreadsheet and lights it on fire, because she collects
meaningful sales salary data, and, you know, and then
when the data comes out it’s like oh snap. I’m fourth on that list I guess, somewhere between White
Female and Black Female, as of 2012. So it’s sort of like this, like, what’s up with that, why did, like why did we
not collect data sooner, and thanks for that, Erica. And I think part of this
is sort of rooted in philosophy, it’s sort of
rooted in Western secularism. And Western secularism
is sort of different in that it sort of, it seeks to bar expressions of religious diversity. Think of, like, the Hijab in Paris, and whatever the fucks going
on with Republican politics right now, (laughter) versus, you know,
according to Amartya Sen, like Eastern secularism, where it says like, if
you offer an identity for one group, you have
to offer it for everyone. So, what happens, you have 200
national languages in India, and, like, there’s a holiday every week, that’s just, that’s just it. ‘Cause if you have a
holiday for the Christians, and you have, you gotta allow a holiday for someone else, and so. Imagine personal data in the future is far more inclusive, so it’s not Western secularism, like, none of that, we
just don’t want any data because we’re post-racial and all that. It’s just like, no, no, no, we just have everyone’s data in here. Or, we could just settle
for that toggle switch. Well, I think that’s kind of crappy and we’re all really
complex individuals here, with like malleable identities and amalgams of qualities,
beliefs, expressions. And, so, I think of
census, like census data. I’m Canadian. More shout outs to the Canadians. Lot a Canadians in the Bay area. And, this is about Jean Talon who sort of, I think they
say he ran the first census. And, it’s, yeah, he
just sort of went around to villages and asked people
like what they wanted, who they were, and Canada and Estonia rank as two countries
that have the highest like family census. It’s kind of like, I think
it’s like law in Canada. You have to fill out the census thing, it’s like, what are you
going to do if I don’t. – [Voiceover] Mail the long forms back. – I and, yeah, and the
long forms back, right. Ted Harp, Stephen Harper, sorry, I’ve been livin’ in Canada awhile, (laughter) When he came into power he stripped, he said, we don’t want to
do long form census anymore. We’ve done it for, what 40, I don’t know, hundreds of years, whatever, and he’s like, no, no, no, we
don’t think we need that data, I’m just going to run this
government how I want to, and, then, you know,
Trudeau brought it back. So, how can we like think about long form identity data so we can
sort of gather more insight and foster Utopian change. Because, like, there are
these situations that happen where we don’t have any data and we can’t, like, solve issues
before they happen, right? So, we need to think about
how data might be used beyond just transactions, advertising, and log ins, towards like trust and safety and like visibility for the least visible and, so, that recent incident with Airbnb, you know, uncovering some
rather systemic issues, with hosts and guests, could that have been
avoided if there was better, like data, built into the onboarding? I’m not sure, I’m asking,
that’s a question. Anyone here work at Airbnb? I think there’s a couple, right? Hey. (laughter) Okay, so, yeah, so, with better identification, can we sort of understand these epidemics
before they sort of happen? Before they spiral into like PR nightmares and a bunch of emotional
damage for people. It’s, you know, pre-triggering. And, so, I think, in all
these things, I like, I tend to side with, like Douglas Rushkoff and Astra Taylor, and believing that the
internet was sort of built on top of a really
flawed operating system, which was, you know, capitalism. And, extractive capitalism at that. And, that unchecked online spaces tentatively make offline
spaces, ’cause we’re human. And, I’ll just read a bit of that quote, “If the Web is hailed for its openness and “that’s where the confusion begins, “because open is in no means equal. “And, while the internet may create space “for many voices, it also reflects “and often amplifies
real-world inequalities “in striking ways.” And, so, and so, there’s also so, we’ve got this thing of what do we do with the people online now, and there’s also, what about the next one
billion people coming online, you know, especially in India and Asia. You know, are we just
going to subject them to our sort of data standards
of personal identity? Like, male, female, white, non-white, cool, all of you guys gets
in that group, it’s chill. So, how can we support them? Or, are we just going to continue with these ethnocentricisms? Because I think, I think
in the preeminent world, you know, many folks
were subjugated to like, out to the fringes because their identity
didn’t sort of fall into the, into the sort of main group, you know? And then you had suffrage, emancipation, post-colonialism, and modernally we sort of got to spread our arms out on the internet while
it’s still new and fresh and, today, I think we are still in this sort of virtual equality, for a lot of minorities and
underrepresented groups. Oh, this one badly designed slide. Here it is. (laughter) So, when I think about virtual equality, I think about Urvashi Vaid’s
book, Virtual Equality. You know, and I think we still, and I sort of map a lot
of my ideas to her, like, her like tiers of humanity, and I think the three tiers, first you have to acknowledge people, or a group, then you have to accept them, and then you, then you
actually, at some point will get to affirming these groups, right? Like, Black Twitter, let’s chat for that, like, and we’ll get there, and we sort of like going
back and forth between all these things, but I think data especially
comes into, like, let’s just acknowledge that
these groups exist online, and that they have these like
really diverse identities. And, yeah, let’s just, get with, acknowledge, I think that was
supposed to mean something. But, we can’t, I think we
can’t get to acceptance and affirmation, like
without acknowledging. Acknowledging the presence. Because, otherwise, again, we’re back to this dominant binary, male, female, white,
non-white, gay, straight, rich, poor, able, disabled. So, let’s bring data to
sort of give us tools to properly survey people along wider axis of like gender, race, sexuality,
ability, socioeconomics. Besides just like, we
know where you’ve been and what you shopped for last. So, because I think
eventually we’ll get to like affirming people, and
the tree of diversity and variation found around. And, so, this is really about challenging, that people should conform to a form. It’s a challenge pessimism of will and the subject of identification data. It challenges that our data works today because it’s collected and it works and it’s good enough, why should I care? It challenges the notion that things have always been this way and people have always been marginalized or put to the fringes. We can’t really account for everyone. It challenges the idea
that it’s too complicated to design and build systems
that better represent the variance of people. And, so, slide 44, we’re gettin’ there. (laughter) I put in these six conjectures, kind of simple rules. This is the overview. I think people have multiple
and overlapping identities. I think omission of data is harmful. I think sampling rates
will improve over time. I think people should
still supervise their data. Data is good as it is wide. And, I think useful
identification is possible. So, one, multiple and
overlapping identities, I take three-time immigrants like myself, we’re often in a state of becoming, we’re kind of never at home, you know, our roots are never at home in new neighborhoods. So, we kind of have this third culture, and I think most people actually have this sort of neuroplastic
malleable identity. I think we’re always changing. Like, thank god, I’m not
the guy I was in college. That’s, that’d be terrible. (laughter) So, how do we make identity robust? How do we allow it to change with us? To incorporate wider aspects of ourself? Because we have neuroplastic brains, and we have a lot of
expressions and outlets, and, we’re not all,
you know, we’re not all this monocrop that’s gonna
listen to spotify’s latest feed. Because hetergenate thrives, and, you know, oh, I did use this word, so why do we insist that the internet is the domain of white
men from the global north? It seems terribly archaic and a hold over from, like, you know, an old world. So, I think we can, we’re
starting to acknowledge in design and engineering,
that the internet can represent plurality. Facebook and Pinterest gender IDs sort of move in that direction. Whoa, skipped ahead. Okay, cool. I was going to get a drink of water. (drinking) So, this one is that the omission of data is actually harmful. And, Melinda Gates, I
think, had just announced that there’s 80 million dollars going towards getting more data around girls around the world. And, she says that data just doesn’t exist for a lot of problems
we’re trying to solve. And it misses women and girls entirely, or undercounts and
undervalues their social and economic contributions
to their famillies and communities. So, it, the omission of data
can actually hamper our ability to advance a cause, and, worse, it can simply distort what
that community experiences. And, sort of just reinforce
our own stereotypes about them in its place. So, such omissions reinforce that a lot of groups of people
simply don’t matter. But, I think of Shivani
Sawyer’s work on inventure, where she’s starting to
use wider pools of data to provide loans to
people that Wall Street wouldn’t otherwise look at. Based on, like, which
data correlations between whose in their network, who
do they make phone calls to, where do they travel, like
who are they as a person, sort of this wide data collection. And, there’s some correlation between the ability to pay back a loan and how many times they make a phone call to a couple of people. So, that was that point. This is sampling rates improve over time. I think the web is still young, and just as sort of, we went from like eight byte audio and eight byte gaming
to higher fidelities, I, I sort of use that as an analogy that we can forecast that identity and behavior sampling
will improve over time to approximate the true diversity and variation found in life. So, rather than this, you know, it’s sort of the expanse of what a thinking, feeling person is. And, I think this is sort of one of the things that
I’d love to get into more, if you want to talk about it, is that there could be
this open web identity, that’s maybe block chain, that gives you control
over your personal data, but you really, sort
of, it’s like open ID, or something like that. You can just define what
people have access to, about, like who you are as a person. And, maybe that identity consortium could go from a small, useful subset into like the several expanding, you know, collection off
the dimensions of humanity. Like, first gender and
then evolve into sex and then properly catalog what region you’re from, and then
specific geographies, and then, I don’t know,
just kind of create this open identity for a lot of people. And, I think we should
still supervise our data, so, even as we sort of venture forward into offering up our data voluntarily, we should still be vigilant about involuntary use of our data. So, like, turning on tracker blockings and so their not just kind of pulling data about you all the time. So, there’s no reason, I feel like you shouldn’t use DuckDuckGo or EFFs Privacy Badger, at a minimum. So, data is as good as it is wide. I think this is sort of we don’t know what we don’t know, often, and so, I think of this like, thing, the story I heard
in an Economics class, where the difference between
surveyed unemployment and real unemployment has to do with, often, whether people had phones, could afford phones in their home, and so, it’s like, well, we
called a bunch of people, and this is what we found out
the unemployment rate was, it’s like, well, what about the people that didn’t have phones? What? What happened to them? And so I think the lesson in trying to become
aware of our blind spots I think of like Clue App, and how, you know, and what they
did for period tracking, which, as a guy that grew up with all boys and in a Catholic Indian family, like, I didn’t know what a period was until my first girlfriend, and, but that’s no excuse, like that we shouldn’t
be trying to collect better personal data about people. And I think, so, the last one is that I think useful identification is possible. I’m doing a product design
fellowship right now and it’s in the financial tech sector, and I’ll probably be looking
for a job afterwards, so, shout out there. (laughter) But, you know, if you’ve
ever worked in the sector, there’s such arcane uses of personal data, like FICO scores are the worst and most opaque sort of form of data. And they just don’t move. It’s just like, okay,
you’ve been put in this, and it’s some invisible calculation with little relevance about your ability to actually pay back a loan. And, so what happens in this
world of opaque financial data, you get loan sharks,
like, predatory lenders, and payday loans, just blowing up, and that’s where we are today. But, some companies are starting to take larger sort of
collections of your data, as actuarial proof that
you can pay back a loan, you know, based on your
networks and who you know. So, their creating sort of
these credit forecasting systems outside of FICO scores. And this is just a screenshot
of something I was working on, but it’s the idea that if you
like sampled your bank history and, when you made payments
it actually will give you a good forecast of your
ability to pay back x amount of dollars within a year, rather than this FICO score that sort of came out of nowhere. And, so, I think, again, personal data and forms, I think within organizations we need to think about
how we design better forms and like start using the custom fields and slack to sort of get, if you’re gonna do it internally, like, are there more
fields that we can put in besides these really bare bones ones, so that we get a sense of what
our organizations look like. So, you know, we can
spot potential red flags before they happen. And, yeah, so survey not surveil. I think of Estonia and
Canada with their E-Census. And, it’s a pretty widespread,
sort of census in Estonia. It’s all digital and the government is continually working on nominalization and visualization of this E-Census with the hopes that in the future the census itself is unnecessary, they’re just passively
collecting all this data from various like government agencies and, you know, putting, giving, you know, giving the government of Estonia a really good insight into
what their society is like. But, as always, data is not
tantamount to social change. You know, we still have
to do the hard work a bit. But, I still think we need
to acknowledge it first and uncover that, and so, countries and companies paving the way for improved census and
better decision making based on data, suggest a future where diversity and well-being is less a slogan and much more like a science. And, so, as Marshall McLuhan says, there’s a poster outside, I just took a picture of it, “We shape the tools and
thereafter they shape us.” Yeah, and that’s it. Thanks guys. (clapping) (upbeat music)

Leave a Reply

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