Category Archives: Cass Sunstein

Internet showing you what they think you want, not what you need (UPDATED)

Flickr photo by antoonsfoobar

I recently saw an interesting TED talk by Eli Pariser on the next wave of cyberbalkanization.  [Read his fascinating new book “The Filter Bubble” here.]

Background: Marshall Van Alstyne predicted 15 years earlier that users would self-segregate on the net and choose to get exposed to ever more narrow communities of interest.

We’re now onto the “The Daily Me” 2.0.  Some news sites originally let users click on their interests a user could limit his/her news to say sports and entertainment news.  Cass Sunstein and Nicholas Negroponte predicted that it would lead to stronger news blinders and expose us to less and less common information, what they called “The Daily Me”.

Well, it turns out that users actually choose to subject themselves to more diversity in opinions and networks on the net than people predicted.

But the latest onslaught, what Eli Pariser calls “The Filter Bubble”, is more invidious.  More and more user sites (Facebook, Google Search, Yahoo News, Huffington Post, the Washington Post) now automatically tailor your stream of results, facebook feed, and news feed based on your past clicks, where you are sitting, what type of computer you use, what web browser you use, etc.

Unlike in the past, this is not “opt in” cyberbalkanization but automatic.  And since it happens behind-the-scenes, you can’t know what you’re not seeing.  One’s search of Tunisia on Google might not even tell you about the political uprising if you haven’t expressed interest in politics in the past.  Eric Schmidt of Google said “It will be very hard for people to watch or consume something that has not in some sense been tailored for them.”

Pariser notes that we all have internal battles between our aspirational selves (who want greater diversity) and our current selves (who often want something easy to consume).  In most of our lives or Netflix queues we continually play out these battles with sometimes our aspirational selves winning out.  These filter bubbles edit out our aspirational selves when we need a mix of vegetables and dessert.  Pariser believes that the algorithmic gatekeepers need to show us things that are not only junk food but also things that are challenging, important and uncomfortable and present competing points of view. We need Internet ethics in the way that journalistic ethics were introduced in 1915 with transparency and a sense of civic responsibility and room for user control.

It’s an interesting talk and I clearly agree with Pariser that gatekeepers should be more transparent and allow user input to tweak our ratio of dessert to vegetables, to use his analogy.  But I think Pariser, in forecasting the degree of our Filter Bubble, misses out the fact that there are other sources of finding about news articles. Take Twitter retweets.  Even if my friends are not that diverse — and many of us will choose to “follow” people we don’t agree with — as long as one of the people I’m following has diverse views in his/her circle of followers and retweets their interesting posts, I get exposed to them.  Ditto with e-mail alerts by friends of interesting articles or social searches using Google.  We live in far more of a social world where information leads come from many other sources than Google searches or Yahoo News.  So let’s work on the automatic filters, but the sky is not falling just yet.

See “The Filter Bubble.” (Feb. 2011 TED talk)

Social Capital Nudges [UPDATED 3/27/13]

In Cass Sunstein and Richard Thaler‘s wonderful book Nudge (2008), they note that individuals’ choices can be influenced by countless factors that people are unaware of (where food is positioned in a cafeteria or a grocery store, what the default settings are on an iPhone, whether programs like retirement contributions are opt-out or opt-in, etc.).  The authors advocate libertarian paternalism: that American leaders be choice architects, trying to arrange society so that Americans maximize good outcomes, but enabling Americans to opt out of these arrangements if they disagree.

Reading the book made me wonder how “choice architects” trying to maximize social capital (social interaction) or civic engagement in society would arrange affairs.

1) Physical space: Architects, as Nudge points out, often are choice architects.  Architects can influence anything from how comfortable we are in our homes or communities to how efficiently we move, to how much we interact.  On the latter, think about front porches or stoops, walkable streets, vibrant public squares, or farmers’ markets as only a few examples of this.  The Forum at Harvard Kennedy School (where public and private leaders come to give talks) was designed based on the Roman Forum and Greek agorae.  It has little nooks along staircases that take you to higher floors, maximizing the chances for interacting, encountering others and conversing.  Of course there are limits to this, as can be seen in “Social Capital and New Urbanism: Leading a Civic Horse to Water?

2) Asking people their intentions in advanceNudge points out that asking people about their intentions (“priming”), increases their chance of doing so.  See this earlier blog post about asking people of their intention to vote to increase voting.   One can also increase the likelihood of voting by having them draw a map of how to get from their home to their voting location.  One wonders what other civic actions we could effectively prime: volunteering, attending a public meeting that discussed town affairs, joining a group, attending a block party, etc.

3) Social norming:  I’ve written about this earlier, but publicizing that most people do things differently than individuals expect can nudge people toward the actual norm. For example, administrators  found that describing the actual infrequency of  college binge drinking (lower than what most college students had believed), lowered the rate of binge drinking further among those who were binge drinking. What are the opportunities to capitalize on this for civic engagement.  We could observe that 40% of Americans attend a house of worship weekly or more frequently and close to three quarters attend at least monthly. We could tell Americans that 62 million fellow Americans volunteer, contributing 8 billion hours of volunteer service worth $162 billion.  And there are undoubtedly other facts we should bring to their attention to increase civic behavior.  [One does have to worry that the folks who are doing a civic behavior far more than the average may be induced to lower their behavior.]  [For an article on social norming, see Tina Rosenberg, ‘The Destructive Influence of Imaginary Peers“, New York Times, 3/27/13]

Another example of this is OPower that is trying to induce people to use less electricity by pointing out how energy conscious their neighbors are.

4) Offer rewards for civic behavior?  We might provide a deduction on health insurance rates for those participating in community groups (a la Samaritan Ministries’ self-insurance or discounts for health club memberships). Similarly, neighborhood stores might offer discounts or coupons for students engaged in regular community service.  Tax deductions are likely more controversial.

5) Civility check: Sunstein/Thaler recommend in Nudge an invention that detects an uncivil e-mail that a user is considering sending and asks the user if he/she is willing to have a 12 hour cooling off period and see if he/she still wants to send the message as is.  [This is reminiscent of Google Mail Goggles which was set up to avoid people sending embarrassing e-mails when drunk.]

6) Get civic commitments up front: Nudge highlights the effectiveness of getting individuals to commit to take action and depositing monthly checks with a third party.  For every month that this commitment is not honored, a group reviled by this individual gets a contribution.  So for example, an advocate of women’s right to choose would deposit checks with this third party made out to the National Right to Life.  She would  vow to attend a book group every month or conversely to attend a church committee’s meetings each month.  Any month that she did attend, her check for that month would be ripped up.  Any month that she didn’t attend, her check would be sent to the NRL.

7) Using groups to reinforce promises:  Many studies have found that commitments made to groups are more likely to be honored.  That is why AlAnon or Weight Loss Groups or Self-Help groups have groups at their core.  We should pursue opportunities like the New Year’s Nudge to encourage people to stick to commitments and might have a Pledge Day where people pledged to take certain pro-civic or pro-social capital activities over the next year.

8) Hanging around other civic types: In James Fowler’s  and Nick Christakis’ Connected, they detail how many social practices are spread through social networks (drinking, depression, obesity, etc.).  It appears from their research that hanging around others who are civic works in the same way;  part of this effect may be the “social norming” effect described earlier, that being around lots of civicly active people changes one’s reference group for citizenship in the same was as being surrounded by overweight people changes one’s sense of how heavy or thin one is.

9) Behavioral placement.  Public health advocates like Jay Winsten were very successful in getting television shows to popularize behavior like “designated drivers.”  More recently, television programs have interwoven wellness behaviors (like exercise or eating right) into show themes.  These may be more effective since a viewer observes a character that the viewer likes or respect undertaking these behaviors.  (Note: it may not be advisable to have Homer Simpson displaying the civic behavior you want to encourage.)

10) metering people’s health and happiness benefits from civic engagement.  Not sure that the studies of the links between social capital and health and social capital and happiness are finely enough calibrated, but it is clear that people alter their behavior in response to more immediate feedback (e.g., the instantaneous MPG feedback on Toyota Priuses or on some GPS devices, the smart meter readouts in some homes of how much electricity they are using instantaneously).  What if there were an iPhone app where you daily plugged in your amount of socializing, volunteering, attendance at public meetings  on town affairs and club going? The iPhone app would generate an instantaneous readout with a face with differing expressions of how well you were doing (with a scowl showing the most displeasure and a smile showing the most) or a stylized body that looked healthier or less healthy depending on your habits.  [A Brazilian bar had an interesting experiment to make people aware of the hidden costs of drunk driving.]

11) Positive deviance: Learn from outliers so others can do the same.  We might do more studies of places that are unusually civic (like Minnesota) to understand what it is that they do differently from communities where there is far less civic behavior and whether nudges could help spur people towards the positively deviant (e.g. very civic community).

12) Setting up defaults.  Here is a graphic on how best to choose a default (not focused on civic engagement).  I haven’t thought of good ideas of how defaults could be used to boost civic engagement, but I’m sure this could be used more.

13) Take advantage of fact that people are  more willing to increase their civic behavior in the future. Sunstein and Thaler capitalize on research that shows that people give more when asked to make a donation in the future than if asked to make it today; they propose a project called “Give More Tomorrow“.  I haven’t seen any research on this but one wonders whether “Volunteer More Tomorrow” or “Coach Little League Tomorrow” or other civic or community engagement  would show similar patterns.

14) Publicize non-civic behavior.  “In a 2005 study, Alan Gerber of Yale got Michigan voters to increase their turnout an amazing 8.6% with a single peer-pressure mailer that listed the previous voting records of their neighbors and noted that a follow-up would be sent indicating who voted this time. (The Obama [presidential] campaign actually priced out a similar mailer but decided not to risk a backlash.)” (From Nudge Blog)

We welcome your thoughts about what nudges might have the biggest impact on increasing social capital or civic engagement/volunteering.   Sunstein and Thaler note that nudges are most helpful where: a) the consequences of bad decisions are delayed (true for many social capital decisions); b) the decision is complex or confusing; c) where you face overwhelming number of choices (true of how we spend our free time); d) where people would benefit from greater feedback from their decisions (true of many social capital decisions); and e) where such decisions are infrequent.

Read  long interview with Thaler here (Fall 2009 Yale School of Management), BBC interview (2009), a To The Point roundtable and Firedoglake conversation.

See “Cass Sunstein Wants to Nudge Us” (NYT Magazine section, 5/16/10)

Read the Nudge Blog.

Facebook as Big Brother (UPDATED 4/10/12)

1984-by-view-askewFacebook has morphed dramatically from their college and pre-college base.  Now only a quarter of users are 18-24 year olds (college and pre-college) and the fastest growth is coming from mature age groups.     Facebook is up to 845 million active users and their IPO capitalized on all the private information that users have inadvertently revealed.  And an Austrian student (Max Schrems), through the Austrian right to discover what information Facebook is collecting, learned that Facebook had 1,222 pages on him including posts he had deleted and his physical location when he posted.

“A Wall Street Journal examination of 100 of the most popular Facebook apps found that some seek the email addresses, current location and sexual preference, among other details, not only of app users but also of their Facebook friends. One Yahoo service powered by Facebook requests access to a person’s religious and political leanings as a condition for using it. The popular Skype service for making online phone calls seeks the Facebook photos and birthdays of its users and their friends.”

The fundamental equation of Facebook is that it provides a free service, funded by Facebook freely distributing the reams of personal information that users reveal about themselves and which Facebook makes available to application developers, advertisers and the like.

Discussing the social implications and privacy, Jeesi Hempel writes in Fortune magazine:

“At times it may seem hard to reconcile Zuckerberg’s lofty aspirations for Facebook with the utterly commonplace content that users create on the site. Consider 25 Random Things, a new take on the chain letter that has grown so popular it was written up in the New York Times Style section. You list 25 supposedly random things about yourself and send the note on to 25 of your friends (who are supposed to do the same), but your randomness also ends up on display to any gawker who may be surfing your profile. The items range from the banal (No. 17: I never, ever, ever throw up. Like five times in my adult life) to the intimate (No. 2: I knew I was gay in the sixth grade but didn’t tell anyone until I was 19). The feature is high profile – some 37,500 lists sprang up in just two weeks – but taken as a whole it just seems like a lot of user-generated babble.  [Note: Slate had a recent post about how 25 Random Things spread in a style approximating a natural virus.]

“Yet it is that very babble that makes Facebook so valuable to marketers. Imagine if an advertiser had the ability to eavesdrop on every phone conversation you’ve ever had. In a way, that’s what all the wall posts, status updates, 25 Random Things, and picture tagging on Facebook amount to: a semipublic airing of stuff people are interested in doing, buying, and trying. Sure, you can send private messages using Facebook, and Zuckerberg eventually hopes to give you even more tools to tailor your profile so that the face you present to, say, your employer is very different from the way you look online to your college roommate. Just like in real life. But the running lists of online interactions on Facebook, known as feeds are what make Facebook different from other social networking sites – and they are precisely what make corporations salivate.”

Facebook users get to “curate their stream” – the flow of information about changes individuals have made to their Facebook page that goes to their social networks on the site.Individuals on Facebook have two feeds: a personal field that logs changes you have made to your own site (a photo, a status update, a video post) and a second feed that tracks all the

“interactions your friends are having (and alerts friends to updates you’ve made on your personal feed). If your brother RSVP’d to a dinner party, for example, you might be notified about it, even if you weren’t invited to attend. And if you change your profile photo, it may let your brother know. Like Facebook itself, the feeds are subject to the network effect: The more data you share and interact with, the more robust your news feed becomes….

The information that pops up is partly a result of controls you establish in your privacy settings and feedback you provide to Facebook. But Facebook also can track your behavior, and if the site notices you’re spending a lot of time on the fan page of a certain movie star, for example, it will send you more information about that celebrity.

Kind of Big Brother-ish, and a marketer’s wet dream.The irony is that despite the use of tracking this personal information to sell you things, users sense that they are not being watched because there is not so much advertising currently on Facebook.It’s almost like baiting a bear by getting it comfortable feeding nonchalantly at a location before one drops the trap. And on Facebook there is no retracting all the personal information that users have left on Facebook about how they know Jane, or their e-mail chatter with friends, or who is in their inner and outer circle based on number of shared friends or who they share their personal feeds with.  That’s all stored on Facebook servers somewhere deep within the enterprise.

Moreover, users’ desire for privacy and Facebook’s desire to know with whom they are dealing often collide.  Facebook has recently actively fought the right of users to use pseudonyms (even for Arab Spring activists or Salman Rushdie).  As someone interested in social capital, I do think there is another side to this story (although I’m not at all sure this is what is motivating Mark Zuckerberg).  Online interactions that are anonymous are far more likely to be vitriolic and interfere with users investing heavily in preserving their online reputation.  If one can lie, or cheat, or flame, and no one knows that it is you, many studies have shown that lying or cheating or flaming is more widespread.

We may be six degrees of separation from anyone else in the world, but only only degree of separation from Big Brother wearing the mask of Facebook.

Facebook is walking a fine line as much of their market value will go up in smoke if they lose user trust.  It is for this reason that several years back they put new changes in the site to a vote in which 30% of their then user base (or 60 million users) had to approve the changes.

This loss of privacy is more concerning, considering  just how many people this affects and how widely users use Facebook to post pictures, posts, links, and friendship patterns that reveal lots about themselves that they probably wouldn’t feel comfortable posting to the world.

Fortune magazine notes how Facebook dramatically shrunk the amount of time to reach 150 million users or sell 150 million units.  Phones took 89 years, televisions took 38, cellphones took 14, iPods took 7 years and Facebook took only 5.  [Fortune doesn’t focus on the fact that the population is larger now so getting to 150 million users or units is easier and that Facebook is aided by the fact that it is free, but Facebook’s growth was impressive nonetheless.]

In 2009, Facebookers spent 169 minutes a month on average (or almost 3 hours) on the site and this increased rapidly.   Fortune doesn’t present a graphic but I assume that there is a group of manic Facebook users that spend 3-5 hours a day on the site or more and some users who use it very rarely.  Facebook acknowledged several years back that less than 10% of users, although still a sizable 15 million folks, do update their status every day (and this is up almost 400% from 2010 while over the same period the number of users was up only 75%, so the growth wasn’t just coming from more users ).

Zuckerberg’s vision is to have Facebook be a “social utility” where “one day everyone would be able to use it to locate people on the web “ David Pogue has an interesting story showing how this is starting to come true: a woman who found a wallet in a NY cab was able to track the wallet’s owner down on Facebook when she couldn’t through 411.  But equally the reality is that it can and is being used not just to locate people on the web, but to sell everything about them to others.

Read other posts about the social implications of Facebook.

See New York Magazine’s, “Do You Own Facebook? Or Does Facebook Own You?” (4/5/09) (describing how Julius Harper’s group on Facebook, protesting Facebook’s privacy policies, swelled to almost 150,000 members

Read the interesting Fortune Cover Story, “How Facebook is Taking Over Our Lives”  by Jessi Hempel  (3/2/09)

Facebook Is Using You” (Lori Andrews Op-Ed, NYT, 2/4/12) that notes that information revealed on Facebook can hurt you down the road in mortgage applications, job interviews, etc.

Read Lori Andrews’ I Know  Who You Are and I Saw What You Did: Social Networks and the Death of Privacy (2012).  [NYT review of book here.]

Read “Selling You on Facebook” (WSJ, 4/8/12 by Julia Angwin and Jeremy Singer-Vine)

And if you want to laugh about it, see the Onion’s satire “Mark Zuckerberg Is a CIA Agent “.  Laughing aside, the CIA has purchased a stake in Q-Tel and Visible Technologies to actually listen in on social media (including YouTube, blogs,  tweets, etc.) and the  CIA has admitted to using social media software in recruiting operatives.

Obama Information Czar and Nudge co-author Cass Sunstein suggested in a 2008 paper “Conspiracy Theories” with co-author Adrian Vermeule that government might “cognitively infiltrate” social networks to help unveil conspiracy theorists and change their minds.