OSI – Social media in closed societies

My friend and colleague Evgeny Morozov is spending a year as an Open Society Institute fellow, working through some of his ideas about cybernationalism and cyberwarfare, and organizing events to discuss the future of the Internet at OSI. I was lucky enough to be included in the first of these events, a presentation by Columbia University and Berkman Center researcher John Kelly and a panel discussion on the role of the blogosphere in closed societies. Good fun, though a 9am event after a long night out on the town the evening before is no one’s idea of a good time.

(Not my fault, really. Joi Ito was kind enough to invite several Global Voices folks to join him at WITNESS’s benefit gala. We spent much of the time trying to figure out what a Global Voices gala would look like – it probably wouldn’t be hosted by Peter Gabriel and Angelique Kidjo, for one thing… not that we’d complain if they wanted to help us fundraise…)

Darius Cuplinskas, head of OSI’s Information Program, framed the discussion by outlining three stories we tell ourselves about online media and their effect on society:

– A picture of sunny optimism, articulated by writers like Don Tapscott, who see the opportunity to contribute and collaborate online as creating a generation of citizens who are more involved and creative than a previous generation of passive media consumers

– A dystopian vision advanced by folks like Andrew Keen, suggesting that the unedited blather of user-generated content will cause us to devalue and neglect expert content and may decrease meaningful participation

– A nuanced view, advanced by thinkers like Cass Sunstein (perhaps more in “Infotopia” than in Republic.com 2.0″) that suggests that new media likely enhanced democracy, but entails new risks, like the isolation and polarization that might come from ideological echo chambers.

These theories, Darius argues, are largely based on research in open societies, especially on the US. But there’s lots less work on the effects of new media in other parts of the world, especially in closed societies, and much of the work that’s done is incomplete and sometimes inaccurate.

John Kelly has been advocating some interesting new methods to explore the blogospheres of open and closed societies – he’s best known for his work visualizing clusters of blogs in the Iranian blogosphere. His method creates fascinating maps of the connections of blogs, clustering blogs together when they link to the same media sources. (If you and I both link to the Christian Science Monitor on a regular basis, we’ll appear close to each other in his maps.) John’s PhD research at Columbia focused on the English-language blogosphere, where he was able to cluster blogs into four major “haystacks” – left and right-leaning political blogospheres, a tech blogosphere, and a UK cluster. (The blogs considered were the 8,000 top blogs as ranked in terms of incoming links, so the patterns displaed are likely very different than from a random sample of blogs.) Smaller clusters exist around science, environmentalism, law, international security and parenting.

One of the major insights Kelly was able to offer in the Iranian blogosphere was the idea that there were lots of blogs that didn’t feature the voices of pro-Western reformers. In fact, the reformers celebrated in Western media were a small cluster, significantly smaller than a pop culture community, a community focused on Perisna poetry, and a number of conservative clusters. One conservative cluster included political bloggers strongly supportive of the Iranian state, but often highly critical of Ahmedinejad. Kelly refers to another conservative cluster as “the 12ers”, adherents to a branch of Shia theology which is awaiting messianic appearance of the 12th imam. It makes sense that western media focused on the liberals, as they’re a group that frequently wrote in English and engaged with Western media, but we’ve got an inaccurate picture of the Iranian blogosphere if we concentrate on that sector.

A new generation of Kelly’s maps overlaws data from the Open Net Initiative, showing which blogs get blocked by Iranian authorities. While lots of liberal blogs are blocked, there’s a decent number of conservative and poetry blogs that get blocked as well – Kelly explains that some of the love ghazals get pretty passionate, and that fervent support of the institution of “temporary marriage” occasionally gets some conservative religious bloggers into trouble.

The maps can “pivot” around a term – it’s possible to see which blogs refer to a term like “America”: religious blogs drop out and political ones will stay. Certain terms – “the 12th imam” – show up only in certain blog communities and are essentially invisible in others. It’s also possible to use the maps to show who’s linking to what – there’s far more linkage to international media like the BBC’s Persian service from pop culture and left-wing blogs than from the right blogs, for instance.

Each language community Kelly has studied shows different patterns and clustering. Arabic blogs appear to cluster geographically, with large, identifiable Saudi, Egyptian and Kuwaiti communities, and some small “trading zones” where there’s lots of cross-linking between globs from different nations. Other blogospheres are harder to explain in terms of clustering – a map of the Russian blogosphere looks incredibly isolated and separated – this is likely a result of the fact that most Russian bloggers use LiveJournal, and the service makes it easy for communities to have small, closely linked circles of friends, which lead to different graphs than blogospheres that evolve on traditional blogging tools.

Kelly is cognizant of a possible critique of his work – these maps are pretty and fascinating, but so what? He suggests that the maps help show the conditions necessary for a healthy online public sphere. We need bloggers, memetrackers and search engines, but we also need rewards within the community to participate. These rewards are generally social capital – attention from other bloggers. High-profile bloggers in the US are motivated by the network of peers they’re relating to online. It’s likely that these motivations are constant across blogospheres and that mapping can help show how social capital flows in different online spaces.

I’m fascinated by John’s work, but I always wonder how to resolve his data-based generalizations with the sorts of observations I make based on our work with Global Voices, looking at a comparatively smaller number of blogs in some detail. I talked briefly about:

– the Ethiopian blogosphere, and how a government-led crackdown on online speech turned off a critical part of the blog ecosystem – the attention bloggers were able to get from readers. With no readers, bloggers gave up, even if they were able to edit their blogs through proxy servers.

– the tolerance for opposition speech in the Zimbabwean blogosphere (indeed, the near absence of speech supporting the current government), which might reflect a commitment to public debate, an inability to effectively filter the web, or the belief that online speech isn’t going to reach beyond an elite audience that’s already unlikely to support the ruling government.

– Michael Anti’s concern that widespread surveillance and pervasive censorship in the Chinese online world is sending activists away from Web2.0 tools and towards an earlier generation of tools which are less powerful, but harder to control.

– the encouraging examples of Kenyan bloggers ability to become an alternative to traditional journalism systems during political crisis, where bloggers who rarely addressed political issues previously became engaged in reporting the news. It’s possible that healthy, thriving blogospheres like Kenya’s are resources that can be activated for social and political purposes when social conditions dictate.

Porochista Khakpour, an Iranian-American journalist and novelist based in NYC, offered a reaction to Kelly’s talk as a regular reader of Iranian blogs and social media. She points out that the book sharing site Goodreads has become a huge social space for Iranians – roughly 20% of the site’s users are Iranian. The site’s managed to stay off the radar of Iranian authorities because it’s not explicitly political, though it’s now starting to be blocked by some Internet service providers.

Khakpour’s traces her personal fascination with the Internet to her conservative upbringing. “I wasn’t allowed to sleep over at other kid’s houses or go to school dances. So I got obsessed with chat rooms” as a space in which she could socialize online, even if she couldn’t be social offline. She’s been obsessed in the past with Friendster and with blogs, and now with Goodreads, and is preparing to check herself into a residential facility for internet addiction… but only because she thinks it will make a good magazine article.

She tells us that the Iranian obsession with the Internet is a recent manifestation of a general fascination with communication technology. During the 1979 revolution, modern technologies – cassette tapes and fax machines – brought the Islamist government into power. Now the Internet has provided spaces not available in modern Iranian society. When a crackdown on the independent press put 1500 reporters out of work in 2001, it helped spark Iranian interest in blogs as an alternative space for reporting and political discussion. Now services like Yahoo 360 – enormously popular in Iran – are serving as a public space for youth who have no spaces where they can congregate. (This resonates with danah boyd’s observations about the internet as an alternative public space for American teens.)

One of the blogs Khakpour is most fascinated by is Life Goes on in Tehran, a photoblog put together by an Iranian determined to challenge stereotypes about his nation and people. Shooting primarily with a cameraphone, the author – “A” – is able to document spaces like house parties where contemporary Tehranis carve out social spaces in what can be a very constrained society.

“The internet is a tool for combatting cultural isolation,” she suggests, explaining that Iran now has higher internet penetration than any other nation in the region, including Israel. She points out the irony that Iran is 9th in the world in terms of blogs hosted, but is also on the list of 15 enemies of the Internet. The truth is that the Internet is perfect for an Iran “taken hostage by fundamentalists” – the anonymity of the medium “is good for passive aggression… or just aggression. And it reflects Iran’s obsession with safety, which precedes the events of 1979.

I’d love to hear Khakpour speak at more length at some point – she drops an amazing wealth of details in her talk. Two that struck me in particular – “Tehran is filled with grafitti that includes URLs – it’s probably the nerdiest grafitti ever.” But it’s important to remember that cultural values in Iran can be very different than in the US. She tells us that an intern with Goodreads started receiving a flood of complaints about the profiles of women with Persian names. Over time, the intern figured out that the complaints were coming from Iranian men – the women in the profiles were photographed without scarves over their hair. Goodreads concluded that this didn’t constitute an inappropriate profile by their rules and left the profiles in place.

Evgeny Morozov, whose wide-ranging interests center on the transformative power of the internet – suggests that in observing communications in closed societies we need to consider government use of technologies as well as activist uses. Referencing the “50 cent party” in China, he suggests that governments are finding ways to use the same tools as activists to support government ideology, creating astroturf campaigns and clogging spaces for dialog with propoganda.

He suggests that, just as Goodreads has emerged as a social space in Iran due to the fact that it’s not obviously political or social, we can expect much of what’s interesting in closed societies to be hidden from easy view or analysis. This is a possible shortcoming in Kelly’s analysis – it’s easy to study blogs, but in countries where blogs are regularly censored or blocked, the interesting conversations are going to be carefully hidden and may defy easy analysis.

Governments, Morozov warns, are developing more subtle and sophisticated ways of discouraging people from blogging than the ham-handed Ethiopian approach I described. The most credible voice in the Ossetian war, he tells us, was a Georgian blogger who’d fled Abkhazia for Russia. His LiveJournal account was highly critical both of Moscow and of Sakashvili, and was widely read in the Russian blogosphere. But a flurry of denial of service attacks, launched by a set of zombie computers likely controlled by Russian hackers, disabled LiveJournal for an hour, and forced the owners of LiveJournal to ask the blogger to leave the service so that future attacks wouldn’t take down the platform. He moved to WordPress, but had the same experience. If governments are able to unleash attacks that can disable whole platforms, it’s likely that they’ll successfully silence many online voices.

Darius summarizes a lively discussion and question and answer period with the observation of two major themes in our discussion:

– Alongside the emergence of explicitly political and activist behavior online, there’s a much larger set of banal, “hedonistic” form of online behavior, which might serve as “dark matter”, capable of becoming political or journalistic if there’s a demand for such behavior

– State responses to social media are getting increasingly subtle, moving beyond simple censorship and blogger intimidation to more nuanced responses, like targetted DDoS attacks. Truly sophisticated approaches are trying to marginalize political speech and suggest that the appropriate use for online tools are these more banal uses, making dissidence socially deviant and less desirable online.