SPEAKER: Continuing the theme
of a power shift in examining who is in charge, our next
speaker is Clay Shirky, who again I'm really excited
to be able to introduce. At previous Zeitgeist, we had
James Surowiecki of The New Yorker and of The Wisdom of
Crowds talk about how collective intelligence made a
difference in how the world works. And the internet had
made this possible. Certainly, it has made
Google and many other companies possible. And I think if that was the
wisdom of crowds, Clay Shirky is going to talk about the
actions of crowds, and how we can understand them. He's on the faculty of
the Interactive Telecom Program at NYU. I think of him as one of the
sanest voices in technology policy and its interaction
with social and political life that there is.
It's hard to strike the balance
between unreasonable pessimism and unreasonable optimism,
but Clay Shirky consistently does it. Please join me in
welcoming Clay Shirky. [APPLAUSE] CLAY SHIRKY: Thanks very much. So I wrote a book that came
out this spring called Here Comes Everybody. And the thesis of the book
in five words is group action just got easier. What I looked at was the ways
in which all of our new network technologies — the internet,
mobile phones, applications built on top of them — are
changing the way groups come together and get things done. Now you've got that book in
your schwag bag, so I'm not going to talk about
that this morning. But it is the theme I'm working
from, very much exemplified by Jennifer's talk
before the break. What I'm going to talk about
today is something I've been thinking about since then,
which is the role of inconvenience in the
design of social systems.
So let's start here. So here we have a photo —
two athletic young men pursuing a basketball. Now this is curious because
these guys are quite well paid. Right? And you would think that one or
the other of them would have had the common sense to go by a
sporting goods store on the way to the game, and buy a second
basketball so they didn't have to spend the whole hour
fighting over this one. [LAUGHTER] In fact, these guys have got
enough money — maybe we could just get a basketball for
every member of both teams. It's only about $50 a piece. Of course it's a
ridiculous idea. The design of basketball
requires inconvenience to function. That's how it works. It needs scarcity, it
fails with abundance. So hold that thought. I'm going to tell you about
something that happened to a friend of mine.
Last November, a former student
of mine who's gone on to become a colleague and a friend, broke
off her engagement with her longtime fiance. And in addition to all of the
sort of emotional horror and administrative trivia that you
have to go through when something like that happens,
she also had to engage in the 21st century ritual of
the changing of the relationship status. She had to go on to Facebook
and she had to change her relationship status from
engaged to single. Because what she didn't want to
have happen, is she didn't want to have the news that she had
broken up with her fiance show up on the Facebook news feed.
Her problem was that she had a
lot of friends on Facebook, but she also had a lot of
"friends." Which is to say people she'd gone to high
school with, people she knew from two jobs ago —
all of the rest. And having it show up in this
environment, having that be the way she tells people —
something she really wanted to avoid. Even worse, many of her
recently former fiance's friends had also friended her,
and she especially didn't want them to find out this way. Because that would be a
terrible thing to do. And so in order to perform this
changing of the relationship status, she goes on to Facebook
and she finds this, which is Facebook's privacy policy. Very well thought out,
very clearly stated. Then she goes here, which is
the interface for managing privacy preferences. She reads the privacy policy,
she read the instructions, she goes to the interface, she
checks the appropriate check boxes, and finally
she can do it. And she goes through and she
changes her status — I'm not seeing the next slide.
There it is. From engaged back to single,
and presses the button. Two seconds later, absolutely
everybody in her contact list got in their news feed — your
friend just changed her relationship status from
engaged to single. Total privacy wipe out. Complete disaster. Exactly the opposite effect of
the one she wanted to create. Her email started going
off, IM, Twitter. The same thing was happening
to her former fiance. Suddenly everybody
knew all at once. So the question we want to
ask in situations like that is, who's at fault? What went wrong? How did this happen? So one answer is my
friend's at fault. She didn't really
understand Facebook.
She did her graduate thesis
work on a comparison of Friendster, Facebook,
and Meetup. If she doesn't get social
networks, it's well beyond the average user. So then you say, well,
Facebook is at fault. This privacy policy isn't
clear, the interface isn't clear, something
went wrong there. James Grimmelmann, who writes
about privacy and social networks at New York Law
School, has said that Facebook has the best privacy policy of
any social network service he's ever looked at. What went wrong here is that
changing your relationship status in a simple and binary
way is an unnatural act. We're simply no good at having
anything called privacy preferences at all. In fact, prior to the current
era the only person any of us even knew of who had anything
that could be called privacy preferences was Greta Garbo. And now suddenly all of
us have to manage the set of check boxes. Prior to the current era, the
principal guarantor of privacy wasn't law and regulation, and
it wasn't engineering and software.
It was inconvenience. That's what kept us safe. We lived most of our lives not
in private or in public, at sort of two extreme ends. We lived most of our lives
in a sphere that we called personal life. A word we hardly use anymore
except — sorry about this. A word we hardly use anymore
except with reference to technology. In personal life you could walk
around talking to your friend, in public yes, but it wasn't
like anybody was spying on you. And it wasn't like every
word you uttered was being recorded for all time. And now it is like that. It's a lot like that. And for people like my friend,
a significant piece of their social life is like that. With privacy we're not moving
from an old engineered system to a new engineered system,
we're moving from an evolved system primarily based on the
inconvenience of the real world into a system
that's engineered.
And we have to figure out how
to put the efficiency back. That's the question. There is no amount of formal
management of single check box yes or no privacy preferences
that's going to accommodate our need for privacy
as we've lived. And as I've looked around after
finishing the book, and started looking at social systems as
they're adapting, I see this pattern over and over again. Places where for the system to
work we have to have a certain kind of inefficiency.
There was a moment shortly
after Friendster launched — the original breakout social
networking service — where young teenagers who weren't
allowed to join the service nevertheless joined in droves. But there was no place for them
to list their age, because they weren't supposed to be there. So 16-year-olds listed their
age as 66, 14-year-olds listed their age as 64. Everybody was in on the joke,
and you could start to sort of look through the system. And because Friendster was
designed to increase the searchability of the system —
how many friends do you have, how easy is it to find them? The combination of those two
factors meant we came this close to having an effective
search engine for 14-year-old girls. We don't want a search engine
for 14-year-old girls. We don't want that for a lot
of really good reasons. And that's now obviously widely
understood in the social networking service.
But the general case is that
there are some kinds of inefficiency we have to have. And that the design of
inefficiency in the digital environment becomes
one of the big social challenges of the age. Here's another example
taken from another domain. Oops. Back please. Thank you. So this is Chris Avenir. He's an 18 year old. And because he's 18, he's
grown up in this environment. When he was 5, there was
already public access to the internet. When he was 10 the commercial
web was already well on its way to being built out. By the time he was 15,
Friendster, MySpace, Facebook had all launched. Before he was 20, he
goes to college. When he gets to college he sets
up a study group for his chemistry class, as
students have done since time immemorial. But because he's the age he
is, he sets up the study group on Facebook. He calls it The Dungeon slash
Ryerson College Chemistry. And it goes pretty well. He gets 146 of his fellow
students to join his chemistry study group.
They start talking
about chemistry. And there they are on Facebook. And all of a sudden the college
calls him up on charges and threatens to expel him. How many charges? 147 charges. One for setting up the Facebook
group, and one for each student who joined. They said — you can't do this,
this is a complete violation of the academic code of conduct. You must come down
and defend yourself. So here's Ryerson's
point of view. What you are doing
is publishing. What you were doing is breaking
out of the environment of the college. You were operating as if you
owned a television station, a printing press, or whatever. Here's Avenir's point of view. He's hiding nothing. He's doing this out in
the open on Facebook. And not just that, he called
his study group The Dungeon because that's the name of the
room on the Ryerson campus where study groups meet.
From his point of view, this
was simply taking real-world kinds of engagement and
moving it on to the web. And so you have this
either-or argument. Is Facebook like a
newspaper, or is Facebook like a study room? But metaphor is our great enemy
here, because if you have to answer either of those
questions, you'll get it wrong. Because it turns out
that Facebook is a lot like Facebook. It does different things
than any other system. In fact, if it didn't do
that they wouldn't have a business model at all. And it is exactly in the ways
that it differs that means there is no obvious choice from
previous experience about what to do in this example.
So clearly Avenir
embarrassed Ryerson. Clearly they overreacted,
clearly they're bad actors here. The story had a
moderately happy ending. He was graded down for the
chemistry class but not expelled, which in my view is
even still a little too far. Ryerson going a little too far. But there is a deeper
principle at work here. What did Avenir do that
upset Ryerson so much? And what he did is that by
being given control of convening large groups he
caused the two messages that exist inside educational
institutions to collide with one another. Colleges have an inside message
and an outside message. The inside message is, welcome
to the community of scholars.
We are a discursive
environment. The ideal kind of class to be
in is a small seminar where you discuss things with your peers. It's about conversation,
it's about community. The outside message from a
college is quality control of individual minds. We take students, we fill
their brains with the information they need, the
experiences they need. When they're ready we
slap a diploma on them, and ship them to you. And the thing that kept those
two messages from colliding is inconvenience in
the real world. And so as long as that happens
the inside message and the outside message don't clash. And what Avenir did was
just run Facebook right through that distinction. But just because Ryerson
College is the bad actor here doesn't mean that this is a
simple case of do as thou wilt is the whole of the law.
Because here's the thing. There may be tables down in the
dungeon at Ryerson, but there's no tables that seat 146. If you've got a group of half a
dozen people who come together for a study group, and somebody
shows up and says — look I'm just here to free
ride on you guys. I just want you to tell me how
to pass the test, but I'm not going to participate. That person gets kicked out. Small groups are really good
at defending themselves against free riders. Large groups are not. And in particular the logic of
digital networks is to reduce the harm of free riders,
rather than to reduce the threat of free riders. In fact, many of our most
important systems, starting with Napster, are
free-rider tolerant. But colleges can't afford to
build an educational experience around tolerance of free riders
any more than you could have a good game of basketball with
an abundance of basketballs. We need some way to ensure that
participants in a group are actually participating. And there's no way in a group
of 146 people that some of those people weren't just in
there essentially shoulder surfing for the answers
to the chemistry test.
And that doesn't even bring in
to account search engines. Because next year all someone
would have to do is just type in Ryerson chemistry study
group, and get all of the conversation without
even joining the group. And so instead of saying we're
going to take this principle from the 20th century, or that
principle from the 20th century — or in the case of education,
from the 12th century — what has to happen is we have to
figure out what kind of inefficiency is going to
produce the right balance between individual and group
effort in educational institutions. There's no simple porting
of old principles to take this example into account. We don't have many good
examples of places where people have figured this
out, but we do have one.
Actually, let's go back
from this for a minute. Sorry. We do have one. And this is a story that
starts in 1980, when Xerox delivers the most important
printer in the world. The most important printer in
the world was the 9700, it was part of the first model
of laser printers. But it wasn't important for
what it did, it was important for where it went and what
it had when it got there. Because where it went was
the MIT lab where Richard Stallman was a student.
And what it had when it got
there was no source code. And Richard Stallman, a
programmer of the first rank, saw that printer
and he saw the future. He said, if this continues
this way it is the end of the community of sharing
that I know and love. It is the end of programmers
sharing their knowledge. And so, another
momentous anniversary. 25 years ago this year
he founded the Free Software Foundation. The Free Software Foundation
added the GNU public license, otherwise called the GPL.
And the GPL is specifically
designed to inject a certain kind of inefficiency into the
act of sharing software. It is inefficiency
as flow control. It says programmers who abide
by these rules can share, and programmers who don't abide by
these rules are prohibited from sharing in certain ways. It is the kind of inefficiency
that increases the social density of the community
of participants. And it's not just the GPL and
it's not just software now, because the intellectual
heritage of this is quite extraordinary. All of the Creative Commons
licenses for content have brought this model to all kinds
of intellectual production. And so we talk a lot about the
really massive scale versions of this — Linux, the Apache
web server, Wikipedia. And these are all huge and
deserving of attention. But that's not the stuff
that winds me up. I'm sorry, go back. That's not the stuff
that winds me up. Here's what gets me.
You can go out on the web and
find pages devoted to the LEGO figurine modding community. It's not just for people who
collect LEGO figurines, it's for people who collect them
and then modify them to use in particular kinds
of LEGO dioramas. And those people will
share with you as long as you credit them. You can go out to the home
schooling community — people we think of as removing
themselves from society — and you'll see incredible
curriculum sharing. You'll see used
textbook networks. There's an incredible
amount of sharing going on in these models. And all of that derives
from the intellectual heritage of the GPL. But now that the network is
becoming so important we're seeing a rise not just in
intellectual production, but in collective action. Here we have a Facebook
group protesting banks' treatment of students. Here we have a group of
Belarussian kids protesting against the Lukashenko
government using LiveJournal, using digital photography for
sharing after the event. Here we have 40,000 kids
walking out of the L.A. Unified School District to
protest anti-immigrant policies.
A protest organized in 48 hours
entirely by MySpace and SMS. But here's a curious thing. We've got all of this fecundity
in intellectual production, and we've got this rise of
collective action. But all of the collective
action examples we have rely on stop energy. There's a huge number of
protest movements — I should say not all, but the
preponderance of them — there's a huge premium on
short-term, high-visibility protests in collective action. So what's the gap? And here's what I
think the gap is. There is no GPL
for group action.
There is no license yet for
collective action that allows groups to come together and
work together over long periods of time around goals related to
collective action, rather just intellectual production. Do this thought experiment. Assume you and a couple dozen
friends, neighbors, decide you're going to do something to
lower the carbon footprint of your neighborhood or town. And you go down to a bank
and you say — hey, we all like each other, we all
share the same goals. Why don't you give
us a bank account? The bank will laugh at you and
send you out of the room. But go away and incorporate in
Delaware, and come back and say — we're a corporation, now
give us a bank account. The banks says —
OK, sign here.
The act of incorporating,
literally of embodying the group, injects four kinds of
inconvenience into that group. The group is hard to set up,
it's hard to tear down, it's hard to join, and
it's hard to leave. And once you do that the state
is willing to defer to the group as a real entity. And that's what I
think is coming now.
There are lots of examples
of people experimenting with how you do this. There is — let me
bang through this. These are in fact the– in
Britain, there's something called the Community
Interest Company. And the Community Interest
Company is a way of starting a for-profit entity that has
social goals built into the charter. Vermont has become the first
state in the union to add something called the
virtual corporation. You can form a company in
Vermont without having to have paper filings, a physical
address, or a hierarchical management structure. And doesn't that sound friendly
to net native groups? Vermont is trying to be the
Delaware of cyberspace. The Meetup. Meetup is creating something
called the Meetup Alliance, which allows bunches of local
real world groups to coordinate their actions on regional
and national scale.
Now I can't tell you which of
these examples might succeed, but I can tell you this. The premium for and the effect
of getting this right are both going to be enormous. Because once we have groups
where the action doesn't have to be short-term, where there
is enough structure built in part around inefficiency, we
can start thinking about the design of social systems that
take digital networks, that take sharing, that take these
kinds of things for granted. Sometimes that means
reconsidering ancient habits as with figuring out what privacy
means in social networks. Sometimes it means redesigning
existing institutions. Like coming up with a new
bargain between individual and group effort in
educational institutions.
Sometimes it will mean the
formation of absolutely new institutions doing things
like taking on the environmental movement. There's a very interesting
effort going on right now called Transition Towns, which
is essentially a recipe for doing my thought experiment. Get a group of people together
and figure out how a bit at a time to convert your whole town
or neighborhood to the idea of lowering energy costs. And this is, I think, the
design challenge that we're heading into. Society is not just the sum of
all its individual members, it's also the sum of
all of its groups. We have done an amazing job of
handing over a high degree of control to individuals. And the results of that,
as we've seen, have been astonishing. The period I think we're
entering into is to figure out how to accomplish that same
task for group action. Thank you very much.