J. Nathan Matias: The first speaker I'd like to welcome is Ethan Zuckerman, if you want to come up. He is director of the Center for Civic Media, Professor of the Practice at the MIT Media Lab, my PhD adviser, and a longtime friend and mentor. Thank you Ethan.


Ethan Zuckerman: Hello, Nathan. Hey, every­body! So, one thing I didn’t under­stand until I start­ed advis­ing PhD stu­dents is that you actu­al­ly spend a lot more time argu­ing than you do nec­es­sar­i­ly teach­ing per se. And Nathan and I had a long argument—really in some ways sort of a six‐year argu­ment. And to maybe over­sim­pli­fy the terms of the argu­ment, my argu­ment was basi­cal­ly if you want to change the world, you need a good under­ly­ing the­o­ry of who you’re gonna influ­ence and how you’re gonna influ­ence them. And unless you have that, don’t wor­ry about how effec­tive your tech­niques are.

Nathan’s basic argu­ment was, Great. Agreed. Useful to have a the­o­ry. But unless you can actu­al­ly mea­sure what you’re doing, how do you know that your pas­sion­ate, well‐meaning action makes things bet­ter instead of mak­ing things worse?”

And when you argue for six years you sort of end up at a point where you real­ize you’re both right, and it’s real­ly most­ly a mat­ter of pri­or­i­ties. It’s real­ly impor­tant both to be able to have some sort of oper­a­tional the­o­ry behind what you’re doing, but it’s also real­ly impor­tant make sure that when you’re sort of gun­ning the gas and try­ing to make the change that you want to you’re actu­al­ly dri­ving in the direc­tion you want to go in.

So that’s sort of the back­ground behind Nathan invit­ing me here to talk about this ques­tion of how peo­ple are using data for social change. And so before div­ing into some of the ways that specif­i­cal­ly in our lab we’ve been think­ing about that, I want to talk a lit­tle bit about how I… I think maybe how Nathan, how a lot of peo­ple around here at MIT think social and civic change sort of ends up hap­pen­ing.

So let me start by say­ing that when I talk about data and social change I’m not talk­ing about Cambridge Analytica. And I’m not talk­ing about Cambridge Analytica for at least two rea­sons. One is that I think gen­er­al­ly, every time you have a pres­i­den­tial elec­tion some­one stands up to say, Oh it was me! It was me. I did it. I found the data that sud­den­ly made some­one une­lec­table elec­table. It was all us and we’ll take the cred­it and con­sult­ing gigs for the next four years.”

But the rea­son I real­ly don’t want to talk about this is that I think less and less, the forms of change that are acces­si­ble to most peo­ple— Not nec­es­sar­i­ly the forms of change that are impor­tant. But the forms of change that peo­ple are often well‐positioned to engage with don’t nec­es­sar­i­ly look like this:

I think that it’s becom­ing hard­er for many peo­ple to feel like they can achieve social change either through the bal­lot box, or through protest, which is sort of our main mech­a­nism where when we can’t win argu­ments at the bal­lot box we stand up and show that we’re not hap­py about things. We failed to elect Hillary Clinton, the largest march in his­to­ry brings peo­ple out into the streets to remind peo­ple that we’ve just elect­ed a ser­i­al sex­u­al harass­er. I want to make the case that both of those meth­ods are actu­al­ly suf­fer­ing as a form of social change.

And the argu­ment real­ly sort of hinges on this graph. This graph is a com­pi­la­tion of social sur­veys going back into the late 1950s and it asks a very sim­ple ques­tion. How much do you trust the gov­ern­ment in Washington to do the right thing?

Now, if we asked that in this room, you know maybe par­tic­u­lar­ly dur­ing this regime, it’s an aca­d­e­m­ic insti­tu­tion, we tend to lean pret­ty lib­er­al… You’re going to have a lot of peo­ple say­ing that they’re not par­tic­u­lar­ly trust­ing. But I want you to move back in his­to­ry.

In 1964, 77% of peo­ple answered the ques­tion I trust the gov­ern­ment in Washington to do the right thing all or most of the time.” We slipped below 20% on this met­ric dur­ing the Obama admin­is­tra­tion; we’ve nev­er recov­ered. We’ve sim­ply stayed there.

We’ve had a rad­i­cal shift over the course of about fifty years in America about how we inter­act with insti­tu­tions of all sorts. We haven’t just had a col­lapse of con­fi­dence in gov­ern­ment insti­tu­tions, we’ve actu­al­ly had a col­lapse of con­fi­dence in all sorts of large insti­tu­tions. When you look at how we used to feel about church­es, banks, labor unions, the med­ical sys­tem, in almost every case if some­thing is big enough that we are not inter­act­ing with an indi­vid­ual, we’re inter­act­ing with an enti­ty, we have lost trust in it.

And by the way, it’s not just us. This turns out to be pret­ty com­mon around the globe. It’s pret­ty com­mon in high‐development nations. It’s pret­ty com­mon in high‐democratic nations. The main nations where trust in insti­tu­tions of all sorts are increas­ing are closed soci­eties. So trust is increas­ing in China, it’s increas­ing in the United Arab Emirates, it’s increas­ing in India—which is real­ly bad news for India. It’s decreas­ing almost every­where else.

So here’s the prob­lem with this. If you don’t trust the gov­ern­ment to do the right thing; if you think either they’re going to do the wrong thing or in some ways even worse; that they’re so incom­pe­tent that they aren’t going to be able to make the changes that you care about, elect­ing peo­ple to office and hav­ing that be your main the­o­ry of change? That doesn’t work any­more.

Weirdly enough, tak­ing to the streets and protest­ing? That may also not work any­more. When we think about the high points of the Civil Rights Movement, when we think about the March on Washington, it was a March on Washington. It was a way of cre­at­ing pub­lic pres­sure on Lyndon Johnson to go car­ry out some of these civ­il rights ini­tia­tives put for­ward under Kennedy, advo­cat­ed for by King. But it was a way of pres­sur­ing a gov­ern­ment that was able to go and make those changes. I’m argu­ing that some large num­ber of peo­ple, includ­ing maybe some peo­ple in this room, don’t believe that we’re in good posi­tions to make those changes any­more.

So what do you do? I went back and read Larry Lessig. I went back to a book that Lessig wrote almost twen­ty years ago. And Lessig says some­thing inter­est­ing but real­ly basi­cal­ly pret­ty sim­ple. He says look, when we think about how soci­eties get gov­erned, we tend to assume that it’s by law. We pass laws and that pre­vents peo­ple from engag­ing in cer­tain behav­iors because if you engage in that behav­ior the police will pick you up, they’ll drag you into court, you might find your­self end­ing up in prison. It’s a good rea­son not to do cer­tain things.

But Larry’s big point is that there’s oth­er ways that we gov­ern behav­ior. Code can gov­ern behav­ior. Larry’s clas­sic exam­ple on this is you stick a disk into your com­put­er (back in the days when com­put­ers had those things). And if it’s an audio disc, your com­put­er says, Hi! Would you like me to copy this for you?” And if it’s a video disc your com­put­er says, I will nev­er let you copy this. I have locked down my entire oper­at­ing sys­tem to make sure this is nev­er copied.” That isn’t law. There’s copy­right law cov­er­ing both of those. That’s code decid­ing that one behav­ior is okay or not.

And we also have cer­tain behav­iors that’re shaped either by mar­kets; it’s cheap or it’s expen­sive to do it. Or by norms. You guys aren’t chal­leng­ing me on every point by jump­ing up and yelling at me because there’s a pret­ty strong social norm that says I have the mic and I speak, and then there’s ques­tions and answers and then you tell me what an idiot I am. But not until I’ve had the chance to get a round of applause, so on and so forth.

So what I’m inter­est­ed in is what hap­pens when we turn Lessig inside out. And I don’t mean that lit­er­al­ly. I think that would be messy. But the invert­ed Lessig is a way of think­ing about hav­ing mul­ti­ple paths to social change. Which is to say it’s pos­si­ble to make change in the world by pass­ing laws. But it’s also pos­si­ble to cre­ate new code; to cre­ate new inter­ven­tions in the mar­ket; to shape norms. And that these may turn out actu­al­ly to be even more pow­er­ful ways to make social change.

Anyone know who this is? Anyone rec­og­nize this per­son? Yell it out. Kim Davis. Why do we know who Kim Davis is? Kim Davis was a coun­ty clerk in Kentucky who refused to issue mar­riage licens­es to same‐sex cou­ples after the Supreme Court deci­sion mak­ing it pos­si­ble for same‐sex cou­ples to mar­ry. This is why law is so pow­er­ful. The Supreme Court made a deci­sion. Everybody in the United States oth­er than Kim Davis sud­den­ly said, Okay. Now we’re going to mar­ry same sex‐couples.” One woman stood up and said we were going to do it any­more, and law ulti­mate­ly made it so that she would have to do it or that some­one would come into her place to do it.

Law is a love­ly way to make change. It’s real­ly clean. It’s real­ly exacting—you make change across the board. The trick is pass­ing laws can get hard when you have a par­a­lyzed Congress. Passing laws has a lot to do with professionalism—having peo­ple who are used to argu­ing court cas­es, who are used to lob­by­ing. It can feel very far from what ordi­nary peo­ple are able to do.

So there’s oth­er forms of change that we’re start­ing to see. Don’t like cli­mate change? Hey, buy an elec­tric car. This starts becom­ing a real­ly inter­est­ing way of tak­ing on cer­tain issues that we haven’t been able to tack­le through law, but we might be able to tack­le through mar­kets. And if Elon Musk and Tesla’s a lousy exam­ple, think about some­thing like rooftop solar pan­els. Turns out to be an inter­est­ing way of think­ing about market‐based the­o­ries of change.

In a room like this one, always fun to talk about things like pri­va­cy and secu­ri­ty. If you don’t like the NSA read­ing your mail, and I cer­tain­ly don’t, we couldn’t get laws passed under Obama to get rid of this. We’re prob­a­bly not going to have it hap­pen under Trump.

But some peo­ple have done some real­ly inter­est­ing work around nor­mal­iz­ing cryp­tog­ra­phy and get­ting it baked into soft­ware that some of us use every day. And that becomes a very pow­er­ful way to make social change.

And then final­ly, I think frankly the most under­used and poorly‐understood path to social change is through chal­leng­ing norms. And I think Black Lives Matter has been an amaz­ing exam­ple of this.

People rec­og­nize that pho­to all the way to left there? So that’s Mike Brown. And that’s Michael Brown’s pho­to tak­en from Facebook, broad­cast on CNN and on oth­er net­works the day that he was killed by police in Ferguson, Missouri.

So, help me out here. What does Mike Brown look like in that pho­to where he’s wear­ing the red shirt? This is real­ly hard in audi­ences full of white peo­ple. I get much bet­ter answers in more diverse audi­ences. Please. Yeah, thank you. He’s an 18 year old kid. He’s try­ing to look tough. He’s being shot from below. He’s flash­ing a peace sign; it looks like a gang sign. You know, I know that I want­ed to look tough when I was 18. Same thing for him.

What’s Mike Brown look in this oth­er pho­to that sort of shows up simul­ta­ne­ous­ly? Right. He’s the same age in the two pho­tos. They’re with­in six months of one anoth­er. But what end­ed up hap­pen­ing was that that pho­to on the left (also from Facebook) was the pho­to that any­body and every­one end­ed up using. The pho­to on the right, very rarely used.

This points to a real­ly inter­est­ing ques­tion of how media por­trays peo­ple. You start­ed to see activists essen­tial­ly say­ing look, if I get gunned down in the street what pho­to would peo­ple use to por­tray me?

And with­in three days of peo­ple start­ing to tweet this you had a full‐scale Tumblr cam­paign of peo­ple putting up pho­tos of them­selves at their best and at their worst. Three days into this, you had The New York Times run­ning a front page sto­ry on this cam­paign.

The way we know this cam­paign worked is right now if you pull up your lap­top, go to Google Images and pull up Mike Brown, you will have a very hard time find­ing that first image. In fact, usu­al­ly the only time you’ll find it is in a copy of my slides.

So mov­ing on from this, my lab, my group of stu­dents, who Nathan’s been one of the lead­ers in, has been work­ing on this ques­tion of how do we use these four levers of change? How do we use not just law, but also norms and mar­kets and code to make change out in the world? And so I want to talk about a cou­ple of projects that we’ve worked on in the hopes that maybe it helps us col­lec­tive­ly here think about the rig­or­ous data that we’re get­ting and the changes we want to make in our com­mu­ni­ties, whether they’re phys­i­cal com­mu­ni­ties or online com­mu­ni­ties, where we go with it.

So, I spent a good chunk of my twen­ties liv­ing and work­ing in Ghana, West Africa. And Ghana turns out to be a real­ly nice place. I high­ly rec­om­mend spend­ing your twen­ties there. If you miss the oppor­tu­ni­ty, sucks for you. But I came back to the United States and was mar­veling at how lit­tle news we got from Sub‐Saharan Africa. And so despite the fact that I can’t real­ly pro­gram my way out of a paper bag, I start­ed writ­ing code and I start­ed essen­tial­ly grab­bing every sto­ry that showed up on The New York Times and throw­ing it on a map every day to sort of show a famil­iar pat­tern.

The New York Times writes about the United States. It writes a lot about Western Europe. It writes a lit­tle bit about big world pow­ers like China and India. It writes basi­cal­ly not all about Sub‐Saharan Africa, Eastern Europe, Central Asia.

Wrote a bunch of papers about this. Predictably noth­ing hap­pened. With my dear friend Rebecca MacKinnon decid­ed that maybe what was hap­pen­ing here was a sup­ply prob­lem. Maybe if we just had more peo­ple in Africa and Central Asia writ­ing great sto­ries in English we’d be able to bal­ance the whole thing out.

We start­ed a com­mu­ni­ty called Global Voices. This com­mu­ni­ty now twelve years lat­er is 1,100 peo­ple strong; puts out an amaz­ing set of web sites in thir­ty dif­fer­ent lan­guages with news from 150 dif­fer­ent coun­tries. Now, amaz­ing thing about this. It did noth­ing in terms of chang­ing what The New York Times writes about. The one time we very clear­ly had impact was dur­ing the Arab Spring. The Times knew noth­ing about Tunisia, we had authors in Tunisia, we were able to be a great short­cut for them.

What it did do instead is help build up this amaz­ing, robust com­mu­ni­ty of peo­ple all over the world who’re real­ly inter­est­ed in this ques­tion of How is my coun­try por­trayed to the rest of the world?” When peo­ple think about Pakistan, what did they think about, and is that a fair way to think about the coun­try?

So the oth­er thing that came out of this research is a plat­form that we’ve been build­ing here for years, prob­a­bly nine years now, called Media Cloud, which is basi­cal­ly a way of say­ing… This is real­ly dis­con­cert­ing. This was like there and next, and I’m con­fused now. [This pre­vi­ous com­ment seems to be about a tech­ni­cal issue but is includ­ed in case not.] But this plat­form lets you say, how often are we talk­ing about Pakistan in The New York Times? And when we’re talk­ing about it, what are we say­ing? We’ve actu­al­ly got to the point where the Global Voices com­mu­ni­ty is using this plat­form in part to try to build its cov­er­age and to try to say here are real holes in how peo­ple are talk­ing about our coun­try.” So we’ve end­ed up with a set of tools, a tech­ni­cal solu­tion; a set of media, real­ly a norms‐based solu­tion, look­ing at this and say­ing this is how our coun­try tends to get mis­rep­re­sent­ed in the West, can we use the data to actu­al­ly put a dif­fer­ent pic­ture out there?”

We’ve also tried using this data for some very dif­fer­ent things. We’ve been using it to ask a ques­tion about how move­ments like Black Lives Matter might change what we get in the news. So this is a study that Nathan has been a key per­son on—we’ve worked togeth­er with about half a dozen oth­er peo­ple on our team—looking at how media atten­tion has been paid to var­i­ous dif­fer­ent unarmed peo­ple of col­or killed by police.

And what we were able to find going through this is that before Michael Brown, the most like­ly thing that hap­pened if you were an unarmed per­son of col­or killed by police is that no one would write about you. Zeor or one sto­ries is the most com­mon out­come. Post Mike Brown, you have a much much high­er chance of get­ting a sto­ry writ­ten about you. You also have a much high­er chance of get­ting a bunch of sto­ries writ­ten about you. And in fact there’s a lot of peo­ple who end up at a much high­er lev­el of media atten­tion post Mike Brown then they would pre Mike Brown.

In fact what hap­pens for about two years after Mike Brown’s death, you have a real shift in how the media is pay­ing atten­tion these sto­ries. Not just peo­ple who end up becom­ing famous, but peo­ple who sim­ply end up being report­ed on in their com­mu­ni­ty. What’s even more inter­est­ing in some ways is that peo­ple change how they write these sto­ries.

When some­one was killed by the police pre Mike Brown, there was about a 2% chance that you would have the name of anoth­er vic­tim in that sto­ry. Post Mike Brown it ris­es to 22%. That’s a pret­ty clear sig­nal that what’s going on is peo­ple are say­ing, This per­son was gunned down…like Mike Brown. Like Freddie Gray.” It’s a way that reporters turn this indi­vid­ual sto­ry into part of a news wave, a whole phe­nom­e­non of what we’re talk­ing about at the same time. And this looks like a strat­e­gy that activists can use to try to fig­ure out how they frame these sto­ries by say­ing this is yet anoth­er instance of this ongo­ing trend that we have to take a clos­er look at.

And now we’re see­ing peo­ple who we weren’t nec­es­sar­i­ly work­ing with using our tools, using our data, to come in and apply this to dif­fer­ent things includ­ing Sophie Chou, for­mer grad­u­ate of the Media Lab, going out and using this tool to inquire about MeToo and to look at the ways in which the MeToo move­ment has focused very heav­i­ly on per­pe­tra­tors and on celebri­ties, but not so much on the peo­ple who actu­al­ly built [inaudi­ble] the move­ment.

But look, data for change doesn’t just have to hap­pen on the screen, it can hap­pen in some very dif­fer­ent com­mu­ni­ties. We’ve been doing a lot of work in communities—favelas—in Brazil. This is a com­mu­ni­ty in Belo Horizonte where we’ve gone out and worked with com­mu­ni­ty groups around this ques­tion of what hap­pens when you get to go col­lect your own data, grab the data out of the com­mu­ni­ty that you care the most about and find ways to visu­al­ize it and think about it.

And so in this com­mu­ni­ty of Santa Marta we went out and asked peo­ple What are the prob­lems that you most care about?” And the prob­lems that we end­ed up find­ing were prob­lems that we didn’t expect peo­ple would tell us about.

I end­ed up spend­ing a very hot day climb­ing stair­cas­es with these two gen­tle­man. And the rea­son for this is that when you’re in a favela, there aren’t a lot of streets, there are a lot of stair­cas­es. And what peo­ple real­ly cared about in this com­mu­ni­ty was cor­rimão, stair rail­ings. Because if you were an elder­ly guy climb­ing up the stairs and you fall, you break your hip. And so what this com­mu­ni­ty real­ly want­ed from the city of Belo Horizonte was rail­ings. And so by going out and map­ping where these stair­cas­es were, map­ping which ones were dan­ger­ous, map­ping which ones didn’t have rail­ings, they sud­den­ly had a pool of data that they could use to go and approach their local gov­ern­ment.

So this project has now become a very large thing, it’s called Promise Tracker. Where it has actu­al­ly caught on the most is high school stu­dents mon­i­tor­ing the qual­i­ty of their lunch. And before you laugh too much about this let me say, high school lunch is pro­tect­ed explic­it­ly in the Brazilian con­sti­tu­tion. How many calo­ries you get. What the qual­i­ty of it is. This is actu­al­ly in the con­sti­tu­tion­al doc­u­ment, because it’s real­ly impor­tant in a mid­dle income nation that peo­ple go out and get fed. And what hap­pened was high school stu­dents took this and start­ed essen­tial­ly say­ing, Look, we are being fad crap­py food.” Who cared the most about this? The local pub­lic pros­e­cu­tor, who went and used the data to go after the peo­ple sup­ply­ing the schools to say Was this cor­rup­tion?” Or was this solv­able prob­lems like the fact that many of these schools don’t have refrig­er­a­tors? So you bring the meet­ing on Monday and by Wednesday there’s noth­ing else that you can serve peo­ple.

This idea has now gone nation­al. There’s peo­ple in com­mu­ni­ties all through­out Brazil who are run­ning this. And what seems to work the best about this is that when you have this arti­sanal data; when you have col­lect­ed it your­self; when you’ve gone out and framed the ques­tion and asked it, this is the stuff that you most want to make change about.

Last thing I’m going to talk about is the data project that we’ve been work­ing on the last year or so. This is a project called Gobo. And Gobo is basi­cal­ly a way of ask­ing what infor­ma­tion you’re encoun­ter­ing out in the world. Gobo is basi­cal­ly an aggre­ga­tor that you got to con­trol. And so you link your Twitter to it, you link your Facebook to it, and you start get­ting these posts in. And unlike on Facebook, you can find why you’re see­ing what you’re see­ing.

Gobo for each post will say Here’s why we fil­tered this in or out of your feed.” And it gives you a set of con­trols that you can look at and basi­cal­ly say, You know, hon­est­ly I’d like peo­ple to be a lit­tle nicer. So I’m gonna slide my rude­ness’ fil­ter over. And frankly I lis­ten to too many men. So I’m gonna slide the gen­der’ slid­er over.” Or actu­al­ly use the Mute all men” but­ton which may be my sin­gle favorite but­ton on the Internet at the moment.

But what this is real­ly about is giv­ing you a chance to exper­i­ment with and inter­ro­gate how algo­rithms are shap­ing the world around you. And so we’re now col­lect­ing data on what do peo­ple want to do with this? Is this a way that peo­ple would want to encounter social media? Is this some­thing that some­one like Facebook should be offer­ing sort of built into their tools?

My point in all of this (now that I have talked to long as Nathan is remind­ing me) is that we ben­e­fit enor­mous­ly from hav­ing the data to link to our intu­itions of what’s impor­tant and our the­o­ries about how we want this to be used for change. Without the data, it’s very hard to move the lever. Without the lever and an intel­li­gent choice of who you’re try­ing to influ­ence and how, it’s hard to use the data in a way that’s effec­tive. So, I know that it’s been use­ful for Nathan and for me in the dif­fer­ent projects that we’ve tak­en on. I hope some of these ideas are use­ful in sort of think­ing about the var­i­ous dif­fer­ent ways you’re try­ing to use data for change in your own com­mu­ni­ties. So, thank you.


Help Support Open Transcripts

If you found this useful or interesting, please consider supporting the project monthly at Patreon or once via Square Cash, or even just sharing the link. Thanks.