- Zeynep Tufekciassociate professor of information and library science at the University of North Carolina at Chapel Hill. She is also a faculty associate at the Harvard Berkman Klein Center for Internet & Society. Her book is titled Twitter and Tear Gas: The Power and Fragility of Networked Protest. Her recent piece for The New York Times is headlined “We Already Know How to Protect Ourselves from Facebook.”
Facebook CEO Mark Zuckerberg faced off with lawmakers in a marathon 5-hour hearing Tuesday about how the voter-profiling company Cambridge Analytica harvested the data of more than 87 million Facebook users, without their permission, in efforts to sway voters to support President Donald Trump. We speak with Zeynep Tufekci, associate professor of information and library science at the University of North Carolina at Chapel Hill. She is also a faculty associate at the Harvard Berkman Klein Center for Internet & Society. Her book is titled “Twitter and Tear Gas: The Power and Fragility of Networked Protest.”
AMY GOODMAN: Facebook CEO Mark Zuckerberg faced off with senators Tuesday in a marathon 5-hour hearing on the privacy scandals plaguing the social network. Zuckerberg was called to answer questions about how the voter-profiling company Cambridge Analytica harvested the data of more than 87 million Facebook users, without their permission, in efforts to sway voters to support President Donald Trump. In the first of two days of hearings, Zuckerberg repeatedly apologized.
MARK ZUCKERBERG: We didn’t take a broad enough view of our responsibility, and that was a big mistake. And it was my mistake. And I’m sorry. I started Facebook, I run it, and I’m responsible for what happens here.
AMY GOODMAN: The Facebook data was first obtained by a Cambridge University academic named Aleksandr Kogan, whose company, Global Science Research, built an app that paid Facebook users to take a personality test and agree to have their data collected. The app also collected data on these users’ friends, meaning it actually collected personal information from tens of millions of users without their knowledge. Cambridge Analytica then bought this data in order to turn a voter-profiling company into a powerful psychological tool, which began launching targeted political ads aimed at carrying out Robert Mercer’s far-right political agenda.
Democratic Senator Kamala Harris of California questioned Zuckerberg about why it took Facebook 27 months—more than two years—to alert users to the Cambridge Analytica breach.
SEN. KAMALA HARRIS: Are you aware of anyone in leadership at Facebook who was in a conversation where a decision was made not to inform your users, or do you believe no such conversation ever took place?
MARK ZUCKERBERG: I’m not sure whether there was a conversation about that, but I can tell you the thought process at the time, of the company, which was that, in 2015, when we heard about this, we banned the developer, and we demanded that they delete all the data and stop using it, and same with Cambridge Analytica. They told us they had.
SEN. KAMALA HARRIS: And I heard your testimony in that regard, but I’m talking about notification of the users. And this relates to the issue of transparency and the relationship of trust, informing the user about what you know in terms of how their personal information has been misused. And I’m also concerned that—when you personally became aware of this, did you or senior leadership do an inquiry to find out who at Facebook had this information, and did they not have a discussion about whether or not the users should be informed, back in December of 2015?
MARK ZUCKERBERG: Senator, in retrospect, I think we clearly view it as a mistake that we didn’t inform people. And we did that based on false information, that we thought that the case was closed and that the data had been deleted.
SEN. KAMALA HARRIS: So there was a decision made, on that basis, not to inform the users. Is that correct?
MARK ZUCKERBERG: That’s my understanding, yes.
SEN. KAMALA HARRIS: OK, and—
MARK ZUCKERBERG: But I—in retrospect, I think that was a mistake. And knowing what we know now, we should have handled a lot of things here differently.
AMY GOODMAN: That is Mark Zuckerberg answering the questions of California Senator Kamala Harris.
We begin today’s show with Zeynep Tufekci. She’s an associate professor of information and library science at University of North Carolina at Chapel Hill, also faculty associate at the Harvard Berkman Center for Internet & Society, author of Twitter and Tear Gas: The Power and Fragility of Networked Protest. Her recent piece for The New York Times is headlined “We Already Know How to Protect Ourselves from Facebook.” More than 2 million people have viewed her recent TED talk titled “We’re Building a Dystopia Just to Make People Click on Ads.”
Professor Zeynep Tufekci, welcome to Democracy Now! Can you talk about what happened yesterday, what Mark Zuckerberg said, what he didn’t say, and what was and wasn’t asked by the senators, in the first of two days of hearings?
ZEYNEP TUFEKCI: So, what was really interesting yesterday is that the senators started asking questions that sounded fine, and probably because the staffers sort of prepared the questions well, and then they got lost in asking the questions. They weren’t able to understand how Facebook actually worked. They kept asking sort of technically weird questions that didn’t make sense.
And even more striking, there were times that Facebook’s own CEO, Mark Zuckerberg, couldn’t answer fairly basic questions on how the platform worked. For example, he was asked, “Can Facebook track users across devices? Does Facebook track people’s browsing or their activities when they’re logged out?” The answers to both are yes, and Zuckerberg struggled and said, “I’ll have my team get back to you.” To me, this is kind of like this interesting moment where we are finally struggling to understand and grapple with the new information commons, even after all these scandals, all this brouhaha.
And I wrote a piece recently for Wired, too, where I listed the 14 years of apologies. So this isn’t even the first time Facebook CEO is apologizing. He’s been apologizing nonstop, before even Facebook was founded. He was apologizing for the previous version of—the initial prototype for Facebook, Facemash.
So, I mean, 15 years later in, we’re finally starting to deal with the kind of power, a platform with 2 billion users, a pretty much significant amount of information flow, socialization, civic functions, politics happens. I did find it quite interesting that there’s finally some questions on its power, whether it had competition, how—you know, because Facebook is essentially without competition at this point. That’s what makes it partly so powerful.
There were a lot of questions that weren’t really asked. I mean, Facebook tracks people who opt—who don’t even use the platform. They have shadow profiles. Its data—Mark Zuckerberg constantly tried to say, “We’ll keep the data within our walls.” But the problem is how much data they collect in the first place. I mean, of course, it’s better if they don’t just recklessly give away the data, as they did probably more than once, not just Cambridge Analytica. But even if they shut that down, as they did in 2015, and even if they do a good job, collecting this much data on 2 billion people, and then selling their eyeballs to whomever is paying Facebook, selling their attention—that’s the product of Facebook—is a huge problem. The fact that if you want to do politics, if you want to socialize, if you have—if you’re an immigrant and have family around the world, the fact that that’s the platform you kind of have to be on is a huge problem. It was touched upon, but not really delved into.
And maybe what I didn’t really see is—and which is what I wrote in The New York Times op-ed that you mentioned—is that: What are we going to do about it? I mean, we actually know enough. We don’t really need Mark Zuckerberg to explain the very basics of Facebook to a bunch of senators who don’t seem to even understand that. We need to sit down and say, “How do we deal with the new information commons? How do we deal with the new public sphere as it operates?”
AMY GOODMAN: I want to turn to Democratic Senator Dick Durbin of Illinois questioning Facebook CEO Mark Zuckerberg.
SEN. DICK DURBIN: Would you be comfortable sharing with us the name of the hotel you stayed in last night?
MARK ZUCKERBERG: Umm, uh, no.
SEN. DICK DURBIN: If you’ve messaged anybody this week, would you share with us the names of the people you’ve messaged?
MARK ZUCKERBERG: Senator, no, I would probably not choose to do that publicly here.
SEN. DICK DURBIN: I think that may be what this is all about: your right to privacy, the limits of your right to privacy, and how much you give away in modern America in the name of, quote, “connecting people around the world”—a question, basically, of what information Facebook’s collecting, who they’re sending it to, and whether they ever asked me, in advance, my permission to do that. Is that a fair thing for a user of Facebook to expect?
MARK ZUCKERBERG: Yes, Senator. I think everyone should have control over how their information is used. And as we’ve talked about in some of the other questions, I think that that is laid out in some of the documents. But, more importantly, you want to give people control in the product itself.
AMY GOODMAN: So, that was Mark Zuckerberg’s response. Zeynep Tufekci, just explain exactly what’s going on here. And also, this bigger point that Kamala Harris raised, when she said, “For more than two years you decided not to tell anyone about the fact—
ZEYNEP TUFEKCI: Right.
AMY GOODMAN: —”that 80 million users had their information given over.”
ZEYNEP TUFEKCI: Right. So, the first thing, what happened is, if your friend had downloaded an app, then your information got transferred to the app maker. And while Cambridge Analytica is in the news because of its political implications, there were maybe tens of thousands, maybe even more, apps that had that kind of access, until 2015 or so. So I would personally be surprised if the number of people whose information was taken that way and then transferred to other parties and is existing somewhere on the, you know, dark web is not pretty much everybody who’s been on the site. That was about a billion people at the time. So, that’s the important thing, is I think that kind of data harvesting probably is much larger than just the Cambridge Analytica app, which is just one app among many.
So, the second thing about the privacy controls is, Facebook does two things here. One, they keep trying to say, “We give you control. We give you control. We’ll keep the data ourselves.” And even if you take that at face value—but as the Cambridge Analytica app scandal sort of shows you, they weren’t—but even if they do this, from going forward, what happens is, it’s quite hard for people to understand exactly what kind of data is collected, and a lot of their controls have been obscure. Like, there’s no one little click saying, “Do not collect data about me.” It’s like you have to kind of figure out. You have to go into a million different menus. And, you know, I have a technical background. I’ve been studying this stuff for a long time. And I sometimes get lost in the weeds of their menus and can’t figure out how to do this. How is an ordinary person supposed to figure this out? So, that’s the first thing.
The second thing is, by promising to keep the data completely secure from now on, they’re still not dealing with the fact that they are collecting an enormous amount of data. And that is not just what you voluntarily share. So, a lot of people say, “You know what? I don’t care. I told Facebook where my college is, what my political views are, and I told Facebook the pages I like, and that’s OK.” But that’s not all it collects. It purchases data from data brokers. You know, there’s your shopping habits. It promised to merges offline data, life if you go into a store and buy something, it wants to match you to your Facebook ID. It collects data while you’re browsing across the web, by tracking pixels.
It tracks you across devices. On Android phones, there was this missed sort of—there was this obscure control. If you kind of missed it, it flashed before you. Before you knew it, you had just given it permission to read and sort of have information on all the text messages, the SMS messages you send, outside of the app. You know, you’re just texting normal people. And when people were downloading their Facebook database this week, as a result of the scandal, and thinking, “What’s in there?” a lot of people found that every single text message they had sent on their Android phone—you know, they’re just messaging their parent or their girlfriend, boyfriend, whatever—it’s all in Facebook’s databases.
So, this kind of surveillance machine, which is then used where the data is used to target you to whomever is paying Facebook, is dangerous on its own. Now, a lot of people mistakenly think that Facebook sells your data. Facebook doesn’t sell your data. Facebook sells you. Right? Facebook is selling your attention. And it’s selling people’s attention screen by screen, so you don’t really get to see on a global level what’s going on. For example, you know, a lot of media critics would remember when, leading up to the Iraq War, New York Times got a lot of things wrong. And it was horrible in its consequences, and it was a grave problem. But you could at least see it and say, “Look, this is wrong.” And you could try to get in the public sphere and try to correct it. I mean, it’s not an ideal situation, obviously. But at least you know what’s in front of you, and you can organize, and you can try to do something, whereas when misinformation is targeted on Facebook, or hate speech or things that either go viral organically or where advertisers can target you, you don’t even see it. So, it’s a—the combination of this kind of power is a huge problem that we need to deal with.
AMY GOODMAN: You know, you gave this very interesting TED talk, where you talk about artificial intelligence, this kind of ad targeting. And people know about ad targeting, maybe. But how deep it is.
ZEYNEP TUFEKCI: Yeah.
AMY GOODMAN: And you talked, for example, about identifying people who are bipolar.
ZEYNEP TUFEKCI: Yes.
AMY GOODMAN: And explain this. It is just astounding.
ZEYNEP TUFEKCI: OK, so, it is astounding. And this is the part—once again, it doesn’t get understood well, because people think like the past. Like I subscribe to Outside magazine. It’s like this travel adventure. So, if somebody wants to advertise on that to me, I’m on the subscriber list. That’s the old method, right? It’s a pretty clear, direct link.
What happens now with artificial intelligence is that computers, computational inference, can take seemingly unrelated data, things like your Facebook likes—right?—not on the topic, but just your Facebook likes, your posting frequency, sort of the semantic tone of your words, and it can predict things like whether you’re likely to enter into clinical depression or whether you’re likely, say, to enter a manic phase, if you have bipolar mental health issues, before the onset of clinical symptoms. So the computer can, you know, with enough data about you—and it’s not a lot of data. It’s definitely the kind of data Facebook has on you. You can predict people’s likelihood of entering a depressive state or a manic state in the next few months, before we even have a clinical test for it, because there are no clinical symptoms.
Now, you can imagine the kind of manipulation that it’s open to. You know, you want to target people that—you want to sell them discounted Las Vegas tickets to? You want to invite them to a casino? Well, you know what? People about to enter a manic phase would be prime, vulnerable targets, because people become compulsive spenders and gamblers, or at least have a tendency to. If you want—if you’ve got people’s personality profiles, which we know from research you can do with Facebook, people with, you know, certain kinds of personality traits are more open to voting for authoritarians when they’re afraid, according to research. You can try to target those people. And sometimes you can do all of this—because of the way current computational models work, you can do all of this without even knowing you’re doing this. You’re going to go tell the computer, “Go find me people who are going to—you know, more likely to buy tickets to Vegas, or more likely to vote for this guy when they’re afraid.” You can just sort of tell the computer what to optimize for, and it can do all of this. And you don’t even know. Like you could even not have the intention to do this, but the way machine learning works today, you could do this.
So, this is what I said. We could enter into a phase of what I term “surveillance authoritarianism,” where we don’t face the kind of 1984 model, where there’s open totalitarianism, where we’re kind of dragged off in the middle of the night, kind of situation. But we’re silently and quietly, and person by person, screen by screen, nudged and manipulated according to our individual vulnerabilities. That kind of authoritarianism would even be hard to realize. You’d just be sort of like being nudged here and there, and there are these slow changes over time. And that’s what’s scary. That’s what we’ve got to get ahead of and not end up there.