Code is law, and also romance

Alexis Madrigal has an interesting column in this month’s Atlantic on the use of algorithms in online dating.  If data mining and algorithms can help people more efficiently find matches, what could be wrong with that?  Plenty, says Madrigal:

The company can quantify things you could guess but might rather not prove. For instance, all races of women respond better to white men than they should based on the men’s looks. Black women, as a group, are the least likely to have their missives returned, but they are the most likely to respond to messages.

I asked Yagan whether OkCupid might try tailoring its algorithm to surface more statistically successful racial combinations. Such a measure wasn’t out of the question, he said. “Imagine we did a lot of research, and we found that there were certain demographic or psychographic attributes that were predictors of three-ways. Hispanic men and Indian women, say,” Yagan suggested. “If we thought that drove success, we could tweak it so those matches showed up more often. Not because of a social mission, but because if it’s working, there needs to be more of it.”

So perhaps it’s a bit tricker than we might think.  Moreover, it’s hard to disagree with his basic point:

Algorithms are made to restrict the amount of information the user sees—that’s their raison d’être. By drawing on data about the world we live in, they end up reinforcing whatever societal values happen to be dominant, without our even noticing. They are normativity made into code—albeit a code that we barely understand, even as it shapes our lives.

We’re not going to stop using algorithms. They’re too useful. But we need to be more aware of the algorithmic perversity that’s creeping into our lives.

Quite so.  This point is in line with Lawrence Lessig’s argument that “code is law”, and I certainly agree that we need to care, as a society, about the values underlying our code.

That said, Madrigal points out that dating algorithms are 1) not transparent and 2) can accelerate disturbing social phenomena, like racial inequity.

True enough, but is this any different from offline dating?  The social phenomena in question are presumably the result of the state of the offline world, so the issue then is primarily transparency.

Does offline dating foster transparency in a way online dating does not?  I’m not sure.  Think about the circumstances by which you might meet someone offline.  Perhaps a friend’s party.  How much information do you really have about the people you’re seeing?  You know a little, certainly.  Presumably they are all connected to the host in some way.  But beyond that, it’s not clear that you know much more than you do when you fire up OkCupid.  On what basis were they invited to the party?  Did the host consciously invite certain groups of friends and not others, based on who he or she thought would get along together?

Is it at least possible that, given the complexity of life, we are no more aware of the real-world “algorithms” that shape our lives?

None of this takes away from the salience of Madrigal’s point: we should want to know more about the algorithms that dictate our online behavior.  Not because we aren’t used to the opaque complexity of circumstance, but because we are.

(FWIW, I highly recommend OkCupid’s blog, OkTrends.  They put the scary amount of data to which they have access to consistently interesting use.)

Facebook and face-to-face

I’ve blogged about this before, but I wanted to share a great post from Ed Glaeser at NYT’s Economix on how social networking – in this case Facebook – supplements in-person interaction, rather than replacing it:Facebook-icon

it isn’t clear if Facebook will increase or decrease the demand for face-to-face interactions.When theory is ambiguous, we need to turn to the data, and it seems empirically that Facebook supports, rather than replaces, in-person meetings. For example, surveys of Facebook users have found that the use of “Facebook to meet previously unknown people remained low and stable” and that “students view the primary audience for their profile to be people with whom they share an offline connection.” In other words, Facebook seems to be typically used to connect people who have connected through some other medium, like being in the same class or meeting at a party, which seems to suggest complementarity between meeting face-to-face and connecting on Facebook.

Another paper looks at whether people who are good at face-to-face interactions made greater use of social-networking sites. The study examined a group of 13- to 14-year-olds in 1998-9 and rated their ability to connect well in person with a close friend. In 2006-8, those same people were asked about their involvement with social-networking sites.

The people who were better at interacting face-to-face in adolescence had more friends on social-networking sites as young adults. Again, electronic interactions seem to complement face-to-face connections.

Transparency or objectivity? Yes.

Jay Rosen linked on Twitter to this post by Terry Heaton, a consultant and journalism professor, on new media ethics that frames the subject in a damaging manner:

There are basically two forms of ethical conduct in the press today. One espouses a traditional set of canons and exists with self-restraint as a guide. In this world, objectivity — or attempts at objectivity — are the norm, for balance and fairness are the goals. Here, truth is presented as existing between two or more “sides” to stories. In the second world, however, transparency replaces objectivity in the belief that the audience can determine bias and figure out where the writer is coming from. In this view, objectivity is a farce and truth determination is up to the reader.

I don’t have much to say about the specific example Heaton examines: a gripe against American Express by TechCrunch’s Michael Arrington.

I have to object, though, to the general frame of objectivity vs. transparency.  I do take issue with the press’s misguided preoccupation with objectivity, and I do support greater transparency.  But I’d hate to see the two unnecessarily juxtaposed.

Not either or

For all the problems with today’s impartiality in journalism – and Rosen’s “Cult of Innocence” post is the place to start on that – there is real value in factually accurate reporting that aspires to a certain kind of objectivity.  (Here’s how I think that could work in the internet age.)  And transparency is aiding experiments in how a new kind of objectivity might look.

The best example I’m aware of here is Wikipedia.  Its community aspires to a certain kind of objectivity, based on a neutral point of view, and is heavily reliant on transparency.  While its process is by no means perfect, it offers some insight into the potential interaction between objectivity and transparency.

Transparency has real advantages if your goal is successfully transforming journalistic objectivity.  First, it offers outsiders a chance to help improve the process.  It’s much easier for the average reader to discover shortcomings in Wikipedia’s process, and to suggest improvements, than it is to do the same with a traditional media outlet.

Second, transparency helps build trust.  And producing accurate journalism that aspires to some version of objectivity is meaningless if no one trusts your process.

Transparency is not enough

Transparency may offer journalists the opportunity to reinvent objectivity within the press, yet transparency alone is no cure-all for journalism.  Just as we’re beginning to see innovative experiments in transparency, I hope we’ll continue to see experiments in objectivity building on successes like Wikipedia.

UPDATE: Also via Rosen, I just came across this 2009 David Weinberger post “Transparency is the new objectivity”.  He makes a good case, but I continue to think the juxtaposition will do damage in the long run, if only via those who fail to read past the subject lines.

Who will feed me my vegetables?

Here’s a snippet from a post imagining a news aggregator built into Facebook, which the author refers to as “inevitable”:

Suddenly, Facebook will funnel news to you from a variety of sources based on data it already knows about you and your friends. Whereas Google News (theoretically) knows little about you until you personalize it, Facebook knows your demographic, your interests, stories and pages you’ve liked, your friends and news they’ve read, liked and commented on.

From the perspective of the user, the potential benefits are obvious.  If Facebook can determine from your profile, your friendships and your conversations that you want to read news items about digital music, cars and technology startups it can save you the time and effort required to customize a news diet as you would through Google News or Google Reader.

But what about the negatives?

Readers won’t realize they’re consuming news from an echo chamber designed by Facebook’s feed algorithm.

This might not matter for certain types of news items, but it matters a lot for others.  Consider politics.  Facebook knows I self-designate as “liberal”.  They know I’m a “fan” of Barack Obama and the Times’ Nick Kristof.  They can see I’m more likely to “like” stories from liberal outlets.

So what kind of political news stories will they send my way?  If the algorithm’s aim is merely to feed me stories I will like then it’s not hard to imagine the feed becoming an echo chamber.

Imagine if Facebook were designing an algorithm to deliver food instead of news.  It wouldn’t be hard to determine the kind of food I enjoy, but if the goal is just to feed me what I like I’d be in trouble.  I’d eat nothing but pizza, burgers and fries.

You might argue that a sophisticated algorithm could identify what we could call “second-order desires”, like wanting eat healthy or wanting to read balanced news coverage.

Perhaps.  But human will power is weak.  Just as we’re bad at sticking to our diets, we’re bad at seeking out perspectives with which we disagree.

For the sake of the public sphere, we need news diets that insist on feeding us our vegetables.

Hierarchies and/or networks

The world really doesn’t need another response to Malcolm Gladwell’s article on Twitter and social revolutions so instead of offering my full thoughts, I’ll just make one point.  Gladwell:

Facebook and the like are tools for building networks, which are the opposite, in structure and character, of hierarchies…

There are many things, though, that networks don’t do well. Car companies sensibly use a network to organize their hundreds of suppliers, but not to design their cars...

Because networks don’t have a centralized leadership structure and clear lines of authority, they have real difficulty reaching consensus and setting goals. They can’t think strategically; they are chronically prone to conflict and error.

This brought to mind a similar point from a recent National Journal piece, How Tea Party Organizes Without Leaders:

Headless organizations have other problems. They are much better at mobilizing to stop a proposal or person they dislike than at agreeing on an alternative. They are bad at negotiating and compromising, because no one can speak for them, and many of their members regard compromising as selling out.

Successful open source projects clearly utilize networks effectively.  But that doesn’t mean that they are paralyzed in the face of decisions.  That’s because they employ alternative decision-making structures – including hierarchies – to tackle tasks for which networks are ill-suited.  Take Linux, for instance.  While a global network of programmers contributes to the project, the community employs a hierarchical decision-making structure to handle changes.  (For more on how that works, I recommend this book.)

When considering the implications of networked political movements, it’s worth remembering that this in no way precludes the use of hierarchies for certain tasks.  Gladwell’s own example of car companies demonstrates the viability of a blended approach.  But he somehow chooses to ignore it when criticizing the usefulness of networks in political change.  Successful political movements will increasingly utilize a blend of the two as we’ve seen in the context of open source software.

(If you’re looking for other reactions here’s Tyler Cowen, Henry Farrell, and a bunch more at Nieman Lab.)

Facebook’s “Photo Memories” and filter failure

If you’ve scanned your friends’ photo albums on Facebook recently, you may have noticed a new feature on the right sidebar labeled “Photo Memories.”  This raises an important issue that I’ve been meaning to blog about for a while: our collective digital memory.  It’s the subject of a fairly new book titled Delete: The Virtue of Forgetting in the Digital Age.

I’ve not yet read the book, but I listened to a talk by the author, Viktor Mayer-Schonberger, at Harvard’s Berkman Center, as well as his conversation with Farhad Manjoo of Slate on Bloggingheads TV.  (While the Berkman talk is a more thorough discussion of his ideas, in some ways the Bloggingheads talk is clearer and more illuminating.)

Here’s how Berkman describes the book:

DELETE argues that in our quest for perfect digital memories where we can store everything from recipes and family photographs to work emails and personal information, we’ve put ourselves in danger of losing a very human quality—the ability and privilege of forgetting. Our digital memories have become double-edged swords—we expect people to “remember” information that is stored in their computers, yet we also may find ourselves wishing to “forget” inappropriate pictures and mis-addressed emails. And, as Mayer-Schönberger demonstrates, it is becoming harder and harder to “forget” these things as digital media becomes more accessible and portable and the lines of ownership blur (see the recent Facebook controversy over changes to their user agreement).

Mayer-Schönberger examines the technology that’s facilitating the end of forgetting—digitization, cheap storage and easy retrieval, global access, and increasingly powerful software—and proposes an ingeniously simple solution: expiration dates on information.

The cataloging of our lives online is a relatively new phenomenon, so we haven’t had much time to consider its impact.  But it’s going to be interesting.  Here’s Facebook VP Christopher Cox explaining the potential impact of Facebook Places:

Too many of our human stories are still collecting dust on the shelves of our collections at home…Those stories are going to be pinned to a physical location so that maybe one day in 20 years our children will go to Ocean Beach in San Francisco, and their little magical thing will start to vibrate and say, ‘This is where your parents first kissed.’

Google CEO Eric Schmidt puts it this way:

I don’t believe society understands what happens when everything is available, knowable and recorded by everyone all the time.

According to the Wall Street Journal, Schmidt  “predicts, apparently seriously, that every young person one day will be entitled automatically to change his or her name on reaching adulthood in order to disown youthful hijinks stored on their friends’ social media sites.”

In case all of this wasn’t tricky enough, applying expiration dates to information, should we want to take Mayer-Schonberger’s advice, is somewhere between difficult and impossible.  Luckily, deleting information isn’t the only way to control our collective memory.  As Clay Shirky says, “There’s no such thing as information overload – only filter failure.”  What we view online is only partially determined by what’s online, because everything’s online.  What we view is determined largely by our filters.  Facebook is a filter, as is Google.  My RSS reader is a filter, as is my email inbox.

Blogging for Foreign Policy, Evgeny Morozov is thinking along the same lines:

So what else could we do, given that expiration-date-technology capable of destroying all copies is not an option? This is an easy one: make offensive information harder to find. After all, it’s the fact that our data is findable – most commonly through search engines – that makes us really concerned.

To apply Shirky’s maxim to the question of digital memory, let’s return to Facebook’s Photo Memories.  The albums of photos of my friends from college have been on Facebook for years.  But, until recently, I had to dig to look at them.  They only registered on my feed if someone tagged someone or commented on them. Typically, as time passed, this happened less and less and then not at all.  And so those photos, though available, were no longer viewed.  And then the filter was changed.  Suddenly, albums from years ago are being thrust in front of me and I’m looking through some of them again.  The point is that digital memory is about more than availability.  In practice, it’s about filters.

What do we want to remember and what do we want to forget?  I’m not sure.  I find several of Mayer-Schonberger’s examples of the dangers of remembering to be quite compelling.  But, in general, the availability of more and more information also has plenty of upside.  Obviously, the question is about balancing the two, and I think it’s clear that we don’t yet have any idea where that balance should be struck.

In the meantime, perhaps we should focus on improving the design and governance of our filters.  We should be in favor of openness, transparency, democracy and individual autonomy.  This isn’t the same as saying we want information to be available and transparent.  But if our filters are open, transparent, and democratic, we’ll at least have an easier time evaluating and improving them.

Perhaps we won’t miss forgetting as much as we think.  (Will we wistfully look back at old photos and fondly remember forgetting?)  Yet, there’s reason to think that if we do ever want to forget, the solution lies in filtering the past rather than deleting it.

Apple’s music social network, Ping, and a followup on Facebook feeds

After my last post, I had a few conversations with friends about a categorized or sortable Facebook feed.  The point of my post was as much about how Facebook could have better managed the transition from profiles to feeds as it was about categorizing updates, but the latter was, for whatever reason, what I ended up debating.

What I heard from multiple people was, basically, “I don’t use Facebook that way.”  As far as I can tell, my friends mostly use Facebook to post pictures, view pictures, and write on walls.

Perhaps this is the wrong conclusion to draw, but I think this is just more evidence that Facebook is continuing to miss out on an opportunity.  The people I spoke to aren’t thinking about Facebook as a tool for discovering new music, books, interests, etc. because Facebook hasn’t made it easy for them to do so.

If my friends are any indication (and perhaps they’re not?), Facebook users view the platform as a place to see what their friends are up to, and to stay in touch.  They don’t see it as an unprecedented social graph with massive potential to inform and recommend various aspects of their lives.

But that’s what it is.

Perhaps the introduction of a music social network by Apple will spur innovation at Facebook.  After all, is a sortable/categorized news feed really all that much to ask?  Maybe I’m nuts, but I think people would learn to appreciate it quickly.

UPDATE: Apparently, Facebook is testing a feature to “subscribe” to certain friends, to be sure you don’t miss any of their updates.  I have nothing to say about this now, and it’s not directly related to what I’m talking about, but I’ll count it as news feed innovation.

Facebook Profiles and the News Feed

Richard McManus has a post at ReadWriteWeb on the declining significance of profile pages on Facebook.  This is as good a jumping off point as any for a post I’ve been wanting to write about what I consider a missed opportunity for Facebook.  But first, a little background from McManus:

As Facebook becomes more and more popular, the social network giant is putting more emphasis on the real-time feed. In other words, the activities of your friends displayed in reverse chronological order on your Facebook homepage. In the old days of Facebook – and indeed traditionally with social networks like MySpace and Friendster – you’d visit a person’s profile page to see what they’re up to. Facebook changed this paradigm in September 2006, when it introduced the news feed as the primary way to keep track of your friends. In October 2009, that feature was re-named the “live feed” and Facebook introduced a more filtered news feed for your homepage.

Now on to my gripe…*

It’s easy to imagine why Facebook would want to push the feed.  The more often content is updated, the more often you’re likely to check the site.  When a huge percentage of the site’s content is relatively static – profile pages – there is less reason to visit.

Yet, profile pages were central to users’ conception of the site.  This may explain at least a small part of the anger over the introduction of the News Feed and the more recent “Connections”.

And my feed on Facebook is pretty uninteresting.  Though Facebook has taken some steps to improve the relevance of the feed, there’s more work to be done.

Which is why I wish they’d built the feed around the existing structure of the profile.  Facebook already had divided my life up into a surprisingly useful, yet simple, set of categories: music, books, TV, movies, activites and interests.  Why not structure a news feed around these categories, plus a Twitter-esque “what are you up to?” (or Facebook’s “What’s on your mind”)?

For each category, imagine you replaced “Favorite” with “Latest” and posted updates by category.  What are you listening to these days? What’s your favorite TV show this season? What book did you just finish?

Now imagine a feed divied up by these categories.  You could see all the updates at once, of course.  But when looking for new music I could click the Music tab on my Feed to see all my friends’ updates on “Latest Music.”

I suspect that this adaptation of the profile structure would have provoked less rage than the original News Feed rollout.  And though that opportunity is missed, it may not be too late to introduce some sort of basic tag/category structure that accomplishes the same thing.

I know it’d make me check Facebook more often.

*I hesitate to second guess these decisions as I generally think users’ reactions against changes to Facebook – the News Feed being the most prominent example – represent a disappointing bias against change of any kind.  Most users didn’t think much about the changes, nor did they give themselves time to grow accustomed to them; they simply protested something new.

Population and the Senate

Admittedly, this has nothing to do with media or the web – the usual topics of this blog… but the U.S. Senate has been on my mind recently for many reasons, including this excellent New Yorker piece by George Packer.

Since each state has two senators regardless of population I started to wonder what the least representative possible Senate majority would look like.  In other words: what percentage of the U.S. population lives in the 26 least populated states (since the senators of these states, acting together, could technically form a majority)?

It turns out that a Senate majority could be reached that represented only 16% of the country. A filibuster could be broken (60 votes, 30 states) by a coalition representing 22% of the country.

(To reach these figures I used the Census Bureau’s 2009 state population estimates.)

I don’t want to make any normative claim here.  The Senate wasn’t designed to reflect state population and when the nation was founded senators weren’t even directly elected.  I don’t have firm opinions about how I’d reform the Senate so my purpose here is merely descriptive.

That hypothetical majority – 52 Senators from the 26 least populous states – would be comprised of 23 Republicans, 27 Democrats and 2 Independents.

The map below shows the 26 least populous states in DARK GRAY, the next 4 least populous (needed to reach a 60 vote coalition) in LIGHT GRAY, and the 20 most populous in WHITE.

Least representative possible Senate majority

Of course in the 60-vote Senate blocking legislation is significantly easier than passing it.  As David Roberts of Grist notes, senators representing a mere 8.3% of the population could successfully filibuster legislation.