The Filter Bubble

ultimately put a lot of front-page editors out of work. Why visit your local paper’s Web site, after all, if Google’s personalized site has already pulled the best pieces?

The Internet’s impact on news was explosive in more ways than one. It expanded the news space by force, sweeping older enterprises out of its path. It dismantled the trust that news organizations had built. In its wake lies a more fragmented and shattered public space than the one that came before.

It’s no secret that trust in journalists and news providers has plummeted in recent years. But the shape of the curve is mysterious : According to a Pew poll, Americans lost more faith in news agencies between 2007 and 2010 than they did in the prior twelve years. Even the debacle over Iraq’s WMDs didn’t make much of a dent in the numbers—but whatever happened in 2007 did.

While we still don’t have conclusive proof, it appears that this, too, is an effect of the Internet. When you’re getting news from one source, the source doesn’t draw your attention much to its own errors and omissions. Corrections, after all, are buried in tiny type on an inside page. But as masses of news readers went online and began to hear from multiple sources, the differences in coverage were drawn out and amplified. You don’t hear about the New York Times’s problems much from the New York Times—but you do hear about them from political blogs, like the Daily Kos or Little Green Footballs, and from groups on both sides of the spectrum, like MoveOn or RightMarch. More voices, in other words, means less trust in any given voice.

As Internet thinker Clay Shirky has pointed out, the new, low trust levels may not be inappropriate. It may be that the broadcast era kept trust artificially high. But as a consequence, for most of us now, the difference in authority between a blog post and an article in the New Yorker is much smaller than one would think.

Editors at Yahoo News, the biggest news site on the Internet, can see this trend in action. With over 85 million daily visitors, when Yahoo links to articles on other servers—even those of nationally known papers—it has to give technicians advance warning so that they can handle the load. A single link can generate up to 12 million views. But according to an executive in the news department, it doesn’t matter much to Yahoo’s users where the news is coming from. A spicy headline will win over a more trusted news source any day. “People don’t make much of a distinction between the New York Times and some random blogger,” the executive told me.

This is Internet news: Each article ascends the most forwarded lists or dies an ignominious death on its own. In the old days, Rolling Stone readers would get the magazine in the mail and leaf through it; now, the popular stories circulate online independent of the magazine. I read the expose on General Stanley McChrystal but had no idea that the cover story was about Lady Gaga. The attention economy is ripping the binding, and the pages that get read are the pages that are frequently the most topical, scandalous, and viral.

Nor is debundling just about print media. While the journalistic hand-wringing has focused mostly on the fate of the newspaper, TV channels face the same dilemma. From Google to Microsoft to Comcast, executives are quite clear that what they call convergence is coming soon. Close to a million Americans are unplugging from cable TV offerings and getting their video online every year—and those numbers will accelerate as more services like Netflix’s movie-on-demand and Hulu go online. When TV goes fully digital, channels become little more than brands—and the order of programs, like the order of articles, is determined by the user’s interest and attention, not the station manager.

And of course, that opens the door for personalization. “Internet connected TV is going to be a reality. It will dramatically change the ad industry forever. Ads will become interactive and delivered to individual TV sets according to the user,” Google VP for global media Henrique de Castro has said. We may say good-bye, in other words, to the yearly ritual of the Super Bowl commercial, which won’t create the same buzz when everyone is watching different ads.

If trust in news agencies is falling, it is rising in the new realm of amateur and algorithmic curation. If the newspaper and magazine are being torn apart on one end, the pages are being recompiled on the other—a different way every time. Facebook is an increasingly vital source of news for this reason: Our friends and family are more likely to know what’s important and relevant to us than some newspaper editor in Manhattan.

Personalization proponents often point to social media like Facebook to dispute the notion that we’ll end up in a narrow, overfiltered world. Friend your softball buddy on Facebook, the argument goes, and you’ll have to listen to his political rants even if you disagree.

Since they have trust, it’s true that the people we know can bring some focus to topics outside our immediate purview. But there are two problems with relying on a network of amateur curators. First, by definition, the average person’s Facebook friends will be much more like that person than a general interest news source. This is especially true because our physical communities are becoming more homogeneous as well—and we generally know people who live near us. Because your softball buddy lives near you, he’s likely to share many of your views. It’s ever less likely that we’ll come to be close with people very different from us, online or off—and thus it’s less likely we’ll come into contact with different points of view.

Second, personalization filters will get better and better at overlaying themselves on individuals’ recommendations. Like your friend Sam’s posts on football but not his erratic musings on CSI? A filter watching and learning which pieces of content you interact with can start to sift one from another—and undermine even the limited leadership that a group of friends and pundits can offer. Google Reader, another product from Google that helps people manage streams of posts from blogs, now has a feature called Sort by Magic, which does precisely this.

This leads to the final way in which the future of media is likely to be different than we expected. Since its early days, Internet evangelists have argued that it was an inherently active medium. “We think basically you watch television to turn your brain off, and you work on your computer when you want to turn your brain on,” Apple founder Steve Jobs told Macworld in 2004.

Among techies, these two paradigms came to be known as push technology and pull technology. A Web browser is an example of pull technology: You put in an address, and your computer pulls information from that server. Television and the mail, on the other hand, are push technologies: The information shows up on the tube or at your doorstop without any action on your end. Internet enthusiasts were excited about the shift from push to pull for reasons that are now pretty obvious: Rather than wash the masses in waves of watered-down, lowest- common-denominator content, pull media put users in control.

The problem is that pull is actually a lot of work. It requires you to be constantly on your feet, curating your own media experience. That’s way more energy than TV requires during the whopping thirty-six hours a week that Americans watch today.

In TV network circles, there’s a name for the passive way with which Americans make most of those viewing decisions: the theory of least objectionable programming. Researching TV viewers’ behavior in the 1970s, pay-per-view innovator Paul Klein noticed that people quit channel surfing far more quickly than one might suspect. During most of those thirty-six hours a week, the theory suggests, we’re not looking for a program in particular. We’re just looking to be unobjectionably entertained.

This is part of the reason TV advertising has been such a bonanza for the channel’s owners. Because people watch TV passively, they’re more likely to keep watching when ads come on. When it comes to persuasion, passive is powerful.

While the broadcast TV era may be coming to a close, the era of least objectionable programming probably isn’t—and personalization stands to make the experience even more, well, unobjectionable. One of YouTube’s top corporate priorities is the development of a product called LeanBack, which strings together videos in a row to provide the benefits of push and pull. It’s less like surfing the Web and more like watching TV—a personalized experience that lets the user do less and less. Like the music service Pandora, LeanBack viewers can easily skip videos and give the viewer feedback for picking the next videos—thumbs up for this one, thumbs down for these three. LeanBack would learn. Over time, the vision is for LeanBack to be like your own personal TV channel, stringing together content you’re interested in while requiring less and less engagement from you.

Steve Jobs’s proclamation that computers are for turning your brain on may have been a bit too optimistic. In reality, as personalized filtering gets better and better, the amount of energy we’ll have to devote to choosing what we’d like to see will continue to decrease.

And while personalization is changing our experience of news, it’s also changing the economics that