Popping the Myth of the Filter Bubble
Why Jacob Weisberg will argue that the Internet expands our political perspectives at the Slate/Intelligence Squared debate on April 17.
Last year, when cyber-Cassandras began to fret over Google personalization, The Slate Group Editor-in-Chief Jacob Weisberg ran an experiment. “I asked a few of my Twitter followers and Facebook friends to search four terms that seemed likely to show ideological fragmentation,” he recounts. The terms? “ ‘John Boehner,’ ‘Barney Frank,’ ‘Ryan Plan,’ and ‘Obamacare.’ ” After the guinea pigs, including a Wall Street Republican and a “quasi-socialist,” found little variation in their results, Weisberg became an evangelist for the notion that filter bubbles exist only in theory. (To the claim that we seek out like-minded voices online even without Google’s help, he replies, “Why assume that when people have more options, they will choose to live in an echo chamber?”)
Far from pinching our political field of vision, Weisberg thinks the Web bombards us with new ideas. He’ll argue that the Internet is not closing our minds when it comes to politics at the Slate/Intelligence Squared live debate on April 17. Weisberg began his career at the New Republic and has appeared on the pages of Slate, the New York Times Magazine, New York Magazine, and Vanity Fair, among other publications. Recently, I talked to him about the Internet and politics, news customization, and his “nightmare vision” of a reader who only cares about the Chicago Bulls, online gambling, and beer.
Slate: You’ll be debating whether or not the Internet closes our minds when it comes to politics. Who is “the Internet?”
Jacob Weisberg: In this case, it’s websites that either provide or filter the news in some way. It’s really a response to the theory that because so many tools and sites on the Web aim to provide some kind of personalization, they narrow our political perspective and leave us only talking to people who agree with us. So in one sense, that’s Google and Facebook, and in another sense it’s different kinds of partisan websites.
Slate: Could “the Internet” be users too? People whose impulses to self-segregate are abetted by the technology?
Weisberg: That probably is part of the argument—that on the Internet, people sequester themselves and end up engaging mostly with those they already agree with. But I don’t think that tendency is any more prevalent on the Web than it is in real life. To the extent there’s any evidence at all, it points more in the direction of people engaging with a broader range of opinions, because of the broader range that’s available. Basically I think this is an issue in theory and not in practice. In my own experience both as a consumer of information and someone who pays close attention to politics, the Internet has the effect of broadening my range of sources and voices. It doesn’t leave me narrower in any respect.
Slate: So, for you, the converse is true? Instead of closing our minds in politics, the Internet is opening our minds?
Weisberg: Exactly. The entire march of technology and information for hundreds of years has been toward a broader diversity of sources. If you wanted a period when the media or technology was really constricting viewpoints to a narrow range, that would have been before the Internet, when people could only get their news from local newspapers, a few national magazines, and three national television networks. It would be very strange if the effect of the march of progress was to close down our perspectives instead of opening them up.
Slate: Do you search Google with personalization on or off?
Weisberg: It’s funny: The personalization option was added pretty recently. It’s opt out. I actually did some tests on it right before they made it an explicit option, when it was supposed to be built in. I ran searches and I had other people run the same searches on political topics. And I couldn’t detect anything at all that amounted to political filtering. The filtering had more to do with location, with demographics. Anyway, I haven’t turned off Google personalized search yet. And I don’t really feel like you need to. But it’d be worth running the test again.
Slate: In a round table discussion published by On the Media in June 2011, you clarify that you “won’t say [you’re] not worried about the effects of personalization.”
Weisberg: Like a lot of changes, it will undoubtedly involve a combination of good and bad effects. I suspect that the good effects will very much outweigh the bad ones, but that doesn’t mean that everything about it will be good or has to be good. You have this first hurdle: Nobody’s figured out how to personalize the news effectively. If you go to the major news sites, there’s very little personalization that actually takes place. I think we’re dealing mainly with a theoretical problem. If the problem is solved through technology—if the news sites become much better at personalizing news—there certainly is the risk that people will not be as broadly informed about a range of things. Bad personalization would be just telling people about what they expressed an interest in. But I also think that would be a news product not many people would want. It would exclude the news, almost by definition. When there’s an earthquake in Haiti, you want to read about Haiti. But not many people would necessarily have said they were interested in Haiti or earthquakes in Haiti before one happened. So part of the problem of news personalization is that it’s hard to know what people will be interested in. But I do think that it’s theoretically possible that you could overdo it.
Slate: Your opponents fear that search personalization might harm serendipity, or chance encounters with new information, on the Web.
Weisberg: I don’t think that’s happened. The Web is all about serendipity—people passing along things that are interesting or notable. If you look at Twitter, it’s both the best tool for news personalization so far and the best tool for the discovery of serendipity. Serendipity is part of this problem of news personalization; A good news source gives you things you aren’t expecting as well as things you are. And the fear of a “Daily Me” that only tells you what you already know—that nightmare vision of someone who only cares about the Chicago Bulls, online gambling, and beer—you have to ask yourself, who would really want that? If you were any respectable consumer of news, you’d remain interested in things beyond your particular hobbies, tastes, and passions. That’s part of the reason the news hasn’t been personalized. It’s so hard to figure out how to give people both what they’ve indicated a preference for and what they as consumers of news need and want. I don’t think there’s any risk of the curator function, the editorial function, going away.
Slate: No, but what if it’s outsourced to a Google algorithm?
Weisberg: You have to differentiate between personalization that’s useful and valuable, and the kind of personalization that theoretically would be negative. So if you’re searching pizza, it’s clearly valuable for the search engine to know where you are and give you pizzerias that are in proximity to where you live and not in another state. But personalization that figures out you’re a libertarian Republican—and heavily weights your news in favor of Reason magazine and the Wall Street Journal editorial page—I think most people would view that as a negative.
Of course, there’s a lot of territory between the two. Personally, I would like to have much more personalized news. I don’t think getting news that I find more relevant would in any respect narrow my political perspective. I think it would do a better job of informing me about the issues that matter to a democratic society.
Slate: What’s the best personalization tool we have right now?
Weisberg: From my perspective, it’s Twitter. By following people on Twitter, you implicitly personalize the news: You figure out what your friends and colleagues, people you respect in different fields, are interested in, and you get a stream that’s more relevant to you. But while you can customize what areas you’re interested in, you’re not going to tune things out.
Slate: Is there something troubling about getting your information from corporate or profit-driven entities—about doing your learning, as Siva Vaidhyanathan says, in a shopping mall, not a public square?
Weisberg: You mean a corporate entity as opposed to ABC Disney or NBC Comcast or Viacom PBS? Because Google and Twitter are technology companies, are they somehow more malevolent than traditional media conglomerates? I don’t think so. In fact, in terms of the value they place on freedom of expression and freedom of information, I think Google and Twitter in particular do very well. Now, their business is very disruptive to the economic model of journalism, but that’s a separate question.
Slate: The other debate camp thinks the tech companies have a special incentive to present you with content you want to click on, so that you see their ads.
Weisberg: Doesn’t the Washington Post have an incentive to give you articles that you want to read? They [Google and Twitter] are getting more direct feedback from their users, so they’ll have better answers to certain questions. How interested were people in this? What was the response to this? Do people want more of this?
Perhaps there was a sort of “ignorance is bliss” aspect to traditional journalism. I often wonder how many people read the coverage of the Albany state government in the New York Times. Not knowing the answer to that question may be what allows the Times to invest as much as it does in such coverage! In a Google world where nobody clicked on those stories, they would go down in priority. But I don’t at all buy this idea that the world of old media was this democratic paragon now being disrupted by technology. We had a narrow, centrist view of the world that came through the major news organizations. It’s the Internet that has allowed hundreds of flowers to bloom.
Katy Waldman is a Slate assistant editor.