an unanticipated consequence of
Jack M. Balkin
Jack Balkin: jackbalkin at yahoo.com
Bruce Ackerman bruce.ackerman at yale.edu
Ian Ayres ian.ayres at yale.edu
Mary Dudziak mary.l.dudziak at emory.edu
Joey Fishkin joey.fishkin at gmail.com
Heather Gerken heather.gerken at yale.edu
Abbe Gluck abbe.gluck at yale.edu
Mark Graber mgraber at law.umaryland.edu
Stephen Griffin sgriffin at tulane.edu
Bernard Harcourt harcourt at uchicago.edu
Scott Horton shorto at law.columbia.edu
Andrew Koppelman akoppelman at law.northwestern.edu
Marty Lederman msl46 at law.georgetown.edu
Sanford Levinson slevinson at law.utexas.edu
David Luban david.luban at gmail.com
Gerard Magliocca gmaglioc at iupui.edu
Jason Mazzone mazzonej at illinois.edu
Linda McClain lmcclain at bu.edu
John Mikhail mikhail at law.georgetown.edu
Frank Pasquale pasquale.frank at gmail.com
Nate Persily npersily at gmail.com
Michael Stokes Paulsen michaelstokespaulsen at gmail.com
Deborah Pearlstein dpearlst at princeton.edu
Rick Pildes rick.pildes at nyu.edu
Alice Ristroph alice.ristroph at shu.edu
Neil Siegel siegel at law.duke.edu
Brian Tamanaha btamanaha at wulaw.wustl.edu
Mark Tushnet mtushnet at law.harvard.edu
Adam Winkler winkler at ucla.edu
The Future of Frictionless Sharing: Facebook's New Audio-Identification Feature
"Frictionless sharing"—a buzzword that encompasses a range of software applications that post automatic updates about what we're doing for our friends to see—has failed in ways that reinforce the importance of boundaries for meaningful social behavior. When everything is shared by default, the amount of information is overwhelming, people disclose things they wish they hadn't, and people are chilled by the concern that their online activities may be broadcast to all their friends.
Facebook recently announced a new feature for its mobile app. This feature—which activates the microphone on a user's phone to identify songs, TV programs, and films playing nearby—has caught flak for its potential encroachment on our private offline conversations. Provided Facebook can mitigate the privacy risks, however, it appears that this feature marks an improved approach to frictionless sharing. The feature does not post any updates or even record any audio without the user's knowledge and consent. Instead, it makes it easier for users to voluntarily indicate exactly what they're watching or listening to. As Ryan Tate at WIRED describes it, the "aim [is] to remove every last bit of friction from the way we reference bits of pop culture."
This blog post explores the past and present of Facebook's approach to frictionless sharing, and then charts a course for Facebook's future. I argue that Facebook has implemented frictionless sharing poorly in the past, but we may be entering a period where efforts to reduce friction can work for users rather than against them. At the same time, this new approach does nothing to mitigate deeper frictions that restrict the sorts of discussions people have on the platform. It promises little more than another nudge towards monetizable conversations about what's hot in popular media.
The Past: Sharing Things We Never Wanted To
Mark Zuckerberg popularized frictionless sharing as a concept in 2011, when he introduced Facebook's plans for "real-time serendipity in a frictionless experience." Cutting past the marketing jargon, Facebook wanted to facilitate the chance discovery of things we might enjoy by making our friends' interests and activities more transparent.
Well-designed tools for frictionless sharing might provide a welcome counterbalance against technological and social trends that reduce our organic sharing. Consider Spotify, a service that discloses your music selections to your friends. In a world where most of us no longer keep vinyl record collections for friends to browse—and where economist Scott Wallsten argues that increased time online comes at the expense of socializing offline—Spotify intends its broadcast of music preferences to simulate "the thrill of looking through a friend's music collection and finding a track that blows your mind."
In practice, however, this paradigm failed to deliver because it kept insisting that people share information against their wishes. Facebook's flirtation with frictionless sharing began with the Beacon program in 2007, well before Zuckerberg hyped the concept in 2011. Beacon posted automatic Facebook updates regarding users' activities at third-party websites, including retailers like eBay and Overstock.com. Its sins were manifold—it is perhaps best known for ruining surprise Christmas gifts—but as James Grimmelmann explains Beacon's key missteps included its automatic enrollment of all Facebook users, its violation of those users' contextual privacy, and its terribly ineffective opt-out mechanisms. Aggrieved users sued and ultimately won a $9.5 million settlement. Even Zuckerberg admits Beacon was a mistake.
Notwithstanding Beacon's failure, Facebook unveiled a new wave of frictionless sharing in 2011. The new way involved "social apps" like social readers—programs that would automatically tell all your friends what you were reading on a third-party site like the Washington Post. Other genres of social apps were no less persistent. Recall how many uninvited status updates you received from social games like FarmVille.
Formally speaking, these programs improved on Beacon because users had to affirmatively opt into them. In practice, however, the improvement often went unnoticed. As Bill McGeveran describes in his comprehensive analysis of frictionless sharing, these apps were notoriously easy to activate by mistake (or without understanding their implications). Yet, like Beacon, they were also notoriously difficult to deactivate.
The automation of sharing raised problems that undermined the putative goal of greater engagement with friends' activities and reading habits. The sheer overload of information caused users to tune out; even if there was a gem to be found, it wasn't worth sifting through the dross. Even worse, as Neil Richards and Margot Kaminski argue with respect to social readers, these disclosures harmed users' privacy and deterred people from engaging with new ideas. Those who didn't understand the system inadvertently posted embarrassing disclosures—McGeveran dubs these "misclosures." Those who did understand were discouraged from reading anything so controversial they wouldn't want it to appear on Facebook.
The Near Present: Automation of User-Initiated Sharing
Facebook is backing away from the old paradigm. It recognizes "that stories people choose to explicitly share from third party apps are typically more interesting" than those the app posts automatically, and "that people often feel surprised or confused by stories that are shared without taking an explicit action." Facebook isn't ready to disable the automatic updates, but it is changing its algorithms to reduce how frequently users see them.
At the same time, Facebook appears to be moving towards new tools that remove friction from users' voluntary sharing. The recently announced audio-identification tool for its smartphone app is at the vanguard. By default, the tool is off. If a user activates the feature, the microphone records fifteen seconds of audio whenever the user begins writing a status update. Like the popular Shazam app, Facebook's app analyzes the audio to identify songs, television shows, or movies playing in the background. The app then gives the user the option of tagging the identified media in the status update. This tag also provides a link to more information—presumably the linked page will also encourage downstream users to purchase access to the song, show, or film. Importantly, no status updates are posted except by the user's voluntary action.
At first blush, this seems like an awfully complicated (not to mention invasive) way to generate a status update. What's the payoff for users? For starters, it takes the friction out of casual updates. If you're sitting on the couch and feel compelled to tell the world that you're watching The Simpsons, you can do so in just a few taps of the touchscreen. Facebook will even supply the text—"Watching The Simpsons"—so you don't have to type anything. Conservation of effort is an understandable criterion for leisure activities.
The tag also has implications for more sustained commentary, where providing standardized tags can reduce the friction of organizing a larger conversation. Let's say you're watching the 2000 X-Men film and want to write about its portrayal of minorities in popular culture. You want to be sure your friends know which film you're talking about while also making it easy for people to find your post when they search for discussions of the film. Consider how the process would play out using existing tools like those on Twitter. A well-crafted hashtag could identify the film in question and put your comment in dialogue with others. But the most obvious hashtag—#xmen—won't work. It has been thoroughly colonized by the 2014 X-Men: Days of Future Past. Facebook's tagging system bypasses this problem by assigning a unique identifier to each film. Or perhaps you'd like to include a link to information about the film, making it absolutely clear which film you're referencing and providing a handy reference for those who missed it in theaters. Doing that means that you'll have to find a suitable link, devote 22 characters to it, and quite possibly break the flow of your prose. Facebook's system streamlines the linking process without adding clutter.
The new feature does of course raise privacy concerns. After all, an errant microphone could record intimate offline conversations. Author and NSA-critic Barton Gellman deleted the Facebook app in response to the announcement, and over half a million users joined a petition urging Facebook to cancel its plans. But these concerns go to the potential privacy harms that might arise from poor implementation, not the new approach to frictionless sharing. The reaction may also indicate users' lack of trust in Facebook. Note that Shazam drew much milder criticism last year when it unveiled a more invasive "Auto Shazam" feature. While Facebook's app records for only fifteen seconds each time the user writes a status update, Auto Shazam records for up to an hour on smartphones and up to four hours on tablets.
For its part, Facebook acknowledges the risks and has taken steps to mitigate them. As noted above, the recording feature is turned off by default, and even when it is enabled it only activates when the user goes to write a status update. In addition, Facebook claims no audio clips are stored after recording, and that the only data retained is the record of a match (or lack thereof) with a particular piece of media. While Facebook plans to aggregate these records in order to gauge the relative popularity of different songs, shows, and films, it intends to disassociate these records from the habits of specific users. Facebook can go a long way towards protecting users by making good on these promises.
The Future: Brave New World?
Facebook's attempt to make voluntary sharing nearly effortless may lead to better communications among Facebook friends about the things they find relevant. But any improvements will be limited. Facebook is reducing the frictions that interfere with the kinds of conversations it wants users to have. But other discussions that users might like to have will continue to face frictions that are deeply rooted in Facebook's advertising interests and its aversion to controversy. As Grimmelmann warns in his article on the advisor theory of search, such conflict is to be expected on platforms like Facebook where search, advertising, and recommendation systems converge. Users may enjoy the new audio-identification tool, but Facebook has designed it to advance its advertising interests: volitional sharing is more probative of users' preferences than passive sharing. The tool is essentially another nudge towards pop-culture discussions that can be readily monetized.
Facebook steers its users away from critical commentary and towards mild praise when they engage with popular culture. Consider Facebook's "likes" system. Like many social networks, Facebook provides users with a "like" button but offers no "dislike" button and little in the way of more nuanced assessment tools. Clicking the "like" button involves minimal effort, and the resulting "likes" become a promotional currency. If you visit the "Orange Is the New Black" page on Facebook, for example, you will learn the series has over 1.7 million "likes." If any of your friends are among that number, you will see their likenesses displayed as though they were personal endorsements. While users can provide other forms of feedback on Facebook, such efforts encounter friction. With no "dislike" button, a user must manually type a critical or otherwise substantive comment and hope someone reads that specific comment. Users who might like to read these comments, however, face additional friction. Unlike more robust user-review sites like Amazon, Rotten Tomatoes, or Yelp, the non-"like" assessments are not aggregated or tallied. At best, Facebook's system encourages users to participate in virtual fan clubs. At worst, it pushes a form of newspeak, limiting users' vocabulary to relative shades of good.
The platform also overtly favors paid content. To be sure, Facebook has developed sophisticated algorithms for determining what content is most relevant to particular users. The "likes" that a post receives, a user's prior interactions with the poster, and the user's other interests all play a role in determining whether an item appears in a user's news feed. Yet posters can pay money to "promote" their posts, causing them to appear in a greater number of news feeds notwithstanding lower scores on these measures of relevance. This quid-pro-quo introduces pecuniary frictions to the system. The business model not only displaces information that users might find more relevant, but also privileges moneyed speakers over mom-and-pop restaurants, local civic organizations, and your actual Facebook friends.
Finally, Facebook polices against materials that might offend or annoy its users. This policy is evident in Facebook's history of censoring controversial content notwithstanding the platform's importance as a news source for growing numbers of users. Andrew Tutt describes the automated removal of posts across the political spectrum, including images of men kissing and anti-Obama memes. Facebook also aggressively screens for nudity, even where it appears in historically significantartor in body paintings meant to raise breast cancer awareness. It even blocked a blog post by Eric Goldman and Venkat Balasubramani on the lawsuit between Facebook and Power.com, an ironic move given that the lawsuit concerned Facebook's allegedly wrongful restrictions on access to its network. These acts of censorship create friction—even where Facebook subsequently reverses itself—impeding speech on the contentious topics that are arguably the most in need of public discussion.
These biases belie Facebook's stated mission of helping people "to discover what's going on in the world, and to share and express what matters to them." They create frictions that steer the system towards the Huxleyan, putting people in constant social contact while they cheerfully promote products to one another. Facebook's new focus on reducing the friction in voluntary sharing is preferable to its prior pursuit of involuntary sharing, hands down. But its payoff for users will be modest in light of the frictions that remain.
BJ Ard is a Postdoctoral Associate in Law and Thomson Reuters Fellow at Yale Law School's Information Society Project. You can reach him by e-mail at bj.ard at yale.edu