How Censorship Backfires on Repressive Regimes

By pulling the plug on the Internet, Mubarak hastened his own demise.

As the digital elements of the Arab Spring continue to be parsed, it seems that at least one conclusion has been reached: Internet censorship can seriously backfire for repressive regimes. Turning off the Internet forces activists into the streets, while selectively filtering popular platforms like Facebook unifies even the non-political and motivates citizens to acquire anti-censorship skills.

A previous post by Alix Dunn on this site explained how the “the Internet kill switch didn’t kill Egypt’s protests” because offline and television transmission of key message was possible [UPDATE: and just released a short case study on the topic]. However, we are now seeing that not only were activists able to mitigate the damage of significant digital censorship, it actually ended up being to their advantage.

1) Digital Service Shut-Down Forces Offline Engagement

Rather than limiting participation in the protests, turning off the Internet (the most dramatic form of digital censorship) actually increases offline participation. This conclusion has been drawn repeatedly with regards to the Egyptian Revolution, most recently by the participants at a recent Danish conference, Cyber Activism Changing the World?: A Conference on Social Media and Women in the Uprising. Courtney Radsch, a PhD candidate studying cyberactivism and Arab media, was in attendance and reports on comments by influential Egyptian digital media commentator Mona Eltahawy:

in Egypt, when Mubarak cut off the internet he effectively forced people to the street – a major tactical mistake in Mona’s perspective. By shutting down the internet activists could no longer tweet or SMS each other to see what was going on so they had to actually go into the streets to find out.

At first, forcing citizens to engage offline might sound like good news for repressive governments. Out in the open, activists are easy to identify, torment, and detain. However, when offline activism takes the form of public protests (rather than back-room meetings), the public nature of the protest creates its own cascade effects. Seeing more people on the street makes it more likely that fence-sitters will join in the protests.

2) Censorship Itself Becomes a Cause

A similar conclusion was drawn by digital activist Noha Atef at the re:publica conference in Berlin in April, with the addition that censorship not only forced people to go into the street to see what was going on, but that censorship became part of what motivated them to do so. Jillian York reports:

When Egypt turned off the Internet and scrambled mobile signals (there are 70 million mobile subscriptions in Egypt, ~80% of the population), Noha explains that it drew more people into the streets. She explains that censorship became a part of the cause, and that despite the Internet shutdown, people were taking photos and videos, knowing that at some point the Internet would be available again.

3) High-Profile Censorship Incites Circumvention Skills Acquisition

This idea of censorship as a unifying grievance and a motivation for protest was also described in the Tunisian context. In a recent event at MIT Ethan Zuckerman reports on the comments of Sami Ben Gharbia, Tunisia’s most prolific digital activist:

Facebook became central to the Tunisian media ecosystem because all other sites that allowed video sharing – YouTube, Daily Motion, Vimeo and others – were blocked by the Tunisian government, along with hundreds of blogs and dozens of key twitter accounts. This censorship, Sami argues, drove Tunisian users towards Facebook, and made it hard for the government to block it. The government tried in 2008, but the outcry was so huge, they reversed course. The main reason – usage of Facebook more than doubled during the 10 days of blockage as Tunisians found ways around the national firewall and onto the service. [emphasis added]

Here again a new element is added. It is not only that blockage of digital technologies forces people to engage with other dissidents IRL and that censorship is a unifying grievance that can motivate participation in protest, high-profile cases of censorship, especially of popular services, also motivates users to learn censorship work-arounds that will limit the government’s ability to censor in the long-term.

4) The Dictator’s Dilemma: Increased Pressure from Economic Elites

The dictator’s dilemma is a theory posited by Christopher Kedzie in 1997 that dictators risk commercial and financial repercussions by limiting the Internet and are thus less likely to do so. Though China has proved that it can selectively block cultural content without negatively affecting economic development, when the question is whether to shut down the entire Internet, the dictator’s dilemma is very much back in play. The fact that Mubarak decided to allow the ISP Noor to continue functioning for a while after the shut-down because it provided Internet access to the Egyptian stock exchange is a direct illustration of the dictator’s dilemma.

The theory of change behind the dictator’s dilemma is that the negative financial repercussions will add further motivation to protesters. But this is not specific enough. It is unlikely that the average Tunisian or Egyptian will feel a direct economic impact from loss of Internet access. However, economic elites will. Economic elites, who are often the beneficiaries of government patronage in non-democratic regimes, are most often regime supporters. That is because these elites are essentially conservative in that they benefit financially from the current regime. (If they are opponents, as in the case of Mikhail Khodorkovsky, they are quickly dispatched).

However, economic elites are often not ideological supporters of the regime. If they see that the regime’s actions are putting their profits at risk, they are likely to put pressure on leaders – whom they often have direct access to – pushing to quickly resolve the political unrest. They may not sympathize with demonstrators – in fact, they probably do not because they do not want political change – but they are one more source of pressure on the regime: “Crush them or leave, we don’t care, just open the banks again.” You can bet that when Mubarak was forced to shut down Noor he was getting a lot of angry phone calls from wealthy investors.

Lessons for Activists

Clearly, censorship still has value to repressive governments, and it can still be effective if it does not bother most citizens. In China, most citizens don’t care about the blockage of Facebook because they have RenRen and Kaixin001. Losing Facebook was not a great loss because it had few users. This was not the case in Tunisia in 2008.

It does not even matter much if elites are circumventing censorship, so long and the information they access is not transferred onto the national internet. As Zeynep Tufekci notes:

Thus, the effect of selective filtering is not to keep out information out of the hands of a determined public, but to allow the majority of ordinary people to continue to be able to operate without confronting information that might create cognitive dissonance between their existing support for the regime and the fact that they, along with many others, also have issues. Meanwhile, the elites go about business as if there was no censorship as they all know how to use work-arounds. This creates a safety-valve as it is quite likely that it is portions of the elite groups that would be most hindered by the censorship and most unhappy with it.

From the perspective of activists who wish to challenge censorship, the best option is to create “dilemma actions” (lose/lose situations for the government) in which either:

  1. Citizens are incited to seek access to a service that is currently censored or
  2. A popular and currently accessible service is shut down.

Put another way, I suggest that Chinese activists post the next Hong Kong film star sex tape on Facebook and start posting hilarious political satire on RenRen. Censorship may not be the solid foundation of Internet control that regimes thought it was.

Semantic Censorship Evasion: Example from Libya

As Gaddafi’s scope of influence shrinks to the capital, news out of Libya is that the regime is still not giving in. A video of the dictator’s son has emerged, beating the drums of war and saying to supporters “I am bringing you reinforcements, resources, food, weapons, everything you need. We are doing well.” The violence is likely to continue.

So far, digital technology has not played the prominent role it played in Egypt’s revolution, but at least one interesting digital activism case study has emerged: using online dating sites for activist networking. ABC news reports:

“We used to call it the digital black hole,” said Nasser Wedaddy, a civil rights outreach director for the American Islamic Congress and longtime cyber activist who has worked on cyber outreach efforts in the Middle East for years. “It’s not that they don’t use the Internet. They’re very afraid.”

Activists in Tunisia and Egypt adopted social media on a mass scale, but “for all intents and purposes, in Libya, there isn’t much cyber activism going on,” Wedaddy said….

To avoid detection by Libyan secret police, who monitor Facebook and Twitter, Mahmoudi, the leader of the Ekhtalef (“Difference”) Movement, used what’s considered the of the Middle East to send coded love letters to rally the revolution.

It was “for the freedom, not for the marriage,” he told ABC News.

Not only are Libyan activists showing their flexibility in switching platforms, they are displaying the same “semantic work-arounds” to censorship that have previously been more visible in China, where images of the cartoon green dam girl were used to critique new censorship software and the terms “grass mud horse” and “river crab” are used to poke fun at online filters to freedom of speech.

In Libya we are now seeing the same innovation in co-opting apolitical words to discuss political topics undetected, including switching genders to allow male-to-male messaging and using codewords like “Jasmine” to refer to the revolution in Tunisia, and “love” for liberty. Once a connection is made, activists can continue their conversations on less public channels like SMS and Yahoo Messenger.

The conservative [dating site Mawada] doesn’t allow men to communicate with other men, so other revolutionaries posed as women to contact him, assuming aliases like “Sweet Butterfly,” “Opener of the Mountain,” “Girl of the Desert” and “Melody of Torture.”

….They also communicated in code the number of their comrades supporting the revolution. The five Ls in the phrase “I LLLLLove you,” for example, meant they had five people with them. If a supporter wrote, “”My lady, how I want to climb this wall of silence. I want to tell the story of a million hurts. … But I am lost in a labyrinth. … Maybe we can meet on Yahoo messenger,” it told the writer to migrate the chat to Yahoo Messenger so as not to raise the suspicion of the monitors, Mahmoudi said.

In the language of social movements, this is a classic example of “tactical innovation,” responding the an opponent’s counter-measures with a creative alternative.

(hat-tip to Patrick Meier for tweeting the ABC article)

Chinese Censorship and the Philosophy of Language

Chinese censors are probably not familiar with the “AAA framework,” but they should be. AAA, which stands for “Aristotle, Augustine, and Aquinas”, is a theory of linguistic meaning that states that “‘meaning is a relationship between two sorts of things: signs and the… things they mean.” The Great Firewall of China is very good at tracking and blocking signs (words), but is getting worse and worse as tracking and blocking meaning. This is because Chinese dissidents and politically-motivated pranksters are getting better and better as creating new signs – new words and images – for critiquing the government that bypass machine-based censoring programs.

In her talk on networked authoritarianism for the conference on Liberation Technology in Authoritarian Regimes last month, scholar of the Chinese Internet Rebecca MacKinnon reported that, when she tried to post a comment on imprisoned Nobel prize recipient Liu Xiaobo on three local Chinese sites, including Baidu and Sina, she was blocked from posting with a moderation message. On the micro-blogging site Sohu the name Liu Xiaobo was removed from the post when it was published.

Yet she might have might have been able to post about Liu Xiabo using Latin letters or a homonym. The creation of alternative signs allows dissident to freely express a censored meaning. Reports the Economist:

News on October 8th that an imprisoned activist, Liu Xiaobo, had been awarded the Nobel Peace Prize spread quickly through domestic microblogs despite the authorities’ best efforts to block it. Users wrote homonyms for Mr Liu’s name, or abbreviations in Latin characters.

These alternative signs are the new proxies, routing meaning around censors. As the graphic below shows, by creating alternative signs for a censored meaning, dissidents can post – and thus communicate – ideas the censors are trying to repress. Examples abound, from images of Green Dam Girl and Grass Mud Horse to discussion of “harmonized” blogs.

Just as the Internet does not solve the free speech problem of activist, automated censorship software does not solve the massive censorship problem of user-generated content. In the end, in a massive system like China’s Internet censorship regime, it is the developers vs. the dissidents.

Unfortunately, for censors there is an extra step – updating the software. Users can iterate new signs in a matter of seconds, whereas it takes hours or days to update censoring software to block new signs. Because it is easier to create new signs than find and block them across an entire system, I believe censors will always be playing catch-up with the Internet’s creative linguistic activists.

Watchdogs of the Mobile Network

The big mobile story of 2002 was that the number of mobile subscriptions had finally exceeded the number of land line subscriptions worldwide. In 2007, the story was that mobile subscriptions had hit 3 billion, dramatically narrowing the digital divide. Since then, the story has been about the backlash. Repressive governments like Iran, China, and Egypt have shifted their censorious gaze from the Internet to mobile phone networks, tracking data on the mobile web and shutting down SMS services before elections and during periods of mass protest.

Fortunately, activists are becoming more aware of these activities, not only by governments, but also by the private firms that help them. The Internet freedom organization Access Now has started a No to Nokia! petition, asking Nokia Siemens, the second largest telecommunications equipment supplier in the world, to “completely end all sales, support, and service of tracking and surveillance technology to governments with a record of human rights abuses”. The petition page notes that the company has admitted that their technology was used to “suppress dissent” in Iran, and Access Now alleges that at least one Iranian journalist, Isa Saharkhiz, was tracked down and imprisoned using Nokia Siemens’ technology. The American Islamic Congress also took action against Nokia Siemens last year by staging a protest outside their Manhattan store.

Researchers are also shining a light on how telecommunications companies are collaborating with governments to reveal subscribers’ private communication. The Citizen Lab at the University of Toronto and the SecDev Group, an Ottawa-based think tank, have teamed up to create RIM Check (, a research project that allows Blackberry users to check how the phone’s manufacturer Research in Motion (RIM) is routing their phone’s encrypted data. The results are retained by the RIM Check project and will be later made public. The homepage states:

Recently a number of governments have threatened to ban Research in Motion’s BlackBerry services if the company does not make encrypted BlackBerry data and other content available to state authorities . A major concern of these regimes is that BlackBerry data can be encrypted and routed through servers located outside of their jurisdictions. Unconfirmed reports have circulated that RIM has made data sharing agreements with India, Saudi Arabia, and United Arab Emirates. Other countries are also requesting the company locate data centres within their jurisdictions.

The RIM case also highlights the convergence of mobile services with the Internet, since much data traffic on the Blackberry is actually Internet traffic, rather than simple mobile service that carries SMS messages. It looks like we are headed into a second front in the war for freedom of digital expression.

The Unsung Heroes of Circumvention

[UPDATED] Most people who use circumvention tools (and there aren’t many of them) use generic simple web proxies, rather than brand-names tools. That was the most interesting take-away I got from the Berkman Center‘s new “2010 Circumvention Tool Usage Report” (PDF), prepared by Ethan Zuckerman and Hal Roberts, along with Rob Faris and Jillian York of the OpenNet Initiative and John Palfrey. Here are some quotes:

First, even though much of the media attention on circumvention tools has been given to a handful of tools – notably Freegate, Ultrasurf, Tor, and Hotspot Shield – we find that these tools represent only a small portion of overall circumvention usage and that the attention paid to these tools has been disproportionate to their usage, especially when compared to the more widely used simple web proxies.

Of the 11 circumvention tools with at least 250,000 monthly users (Ultrasurf, Freegate, Tor, Hotspot Shield, and SWP #s 1 – 7), 7 are simple web proxies. Those 7 proxies together appear to serve close to half of the combined unique users of the 183 simple web proxies whose usage we were able to estimate.

We were surprised to discover that several widely-used simple proxies remained unblocked for very long periods of time in highly censorious nations that aggressively block the more well-discussed blocking-resistant tools. This difference in the treatment of the different types of tools may be the result of the difference in press coverage of these tools. Unlike Freegate, Ultrasurf, and Tor, the more widely-used simple web proxies have not been lauded much if at all in the U.S. press as agents of political change.

Support for circumvention technology has been a major element of the US State Department’s Internet Freedom initiative, and brand-name tools seem to be their focus, with Secretary Clinton’s public support for Haystack being the most unfortunate example.

The report indicates that the importance of these brand-name circumvention tools may be overstated, at least in terms of user volume, which might convince the State Department to change their focus as well. However, if simple web proxies are able to “remain unblocked for very long periods of time in highly censorious nations” because they “have not been lauded much if at all in the U.S. press as agents of political change” then I hope that the State Department refrains from publicly aligning these tools with US interests.

It’s a catch-22 of digital activism research: sometimes a tactic works precisely because it is not publicly known.

Better in the Dark

by Mary Joyce

Famed sociologist Jürgen Habermas likes to do civic participation with the lights on. According to his theory of the public sphere, societies benefit from having a space where citizens can “discuss matters of mutual interest and, where possible, to reach a common judgment”. The classic analog example of a public sphere is the coffee house. Professor Yochai Benkler of Harvard Law School then applied this idea to the Internet with his theory of the networked public sphere, in which this civic space exists online and is mediated through digital tools. This public ideal of digital participation and activism is born out in fact: most of the digital tools used for activism, like blogs and social networks are extremely public. Recent design changes, like the loosening of privacy by Facebook, have made digital activism on those platforms even more public.

In repressive societies, the public nature of these popular technologies soon brought grief to activists. Once authorities figured about that activists were using these tools they could simply check an activist’s Facebook profile page (as they do in Iran) to see who their likely co-conspirators are or arrest visible online organizers (as happened in Moldova) when seeking the leader of a protest. The Russian government has gained increasing influence over the popular social blogging platform LiveJournal, probably also in an effort to monitor member’s actions. Vietnam and China have gone so far as to hack seemingly private applications, like Gmail.

Not surprisingly, the increasingly danger of using public technologies to mobilize in repressive regimes has drawn responses from activists, such as Sami Ben Gharbia of Global Voices Advocacy, who has cautioned activists in repressive states not to use Facebook. Jillian York of the Berkman Center for Internet and Society has also been analyzing the effect of reduced privacy on digital activists. Critics like Evgeny Morozov see the increasing savvy on the part of repressive governments as further evidence of the danger of digital activism.

Yet this is overstepping. Public digital platforms are becoming ever more dangerous to activists in repressive regimes, yet they are still useful to activists in freer societies. Also, to use the danger of public platforms to condemn all digital activism tools as dangerous is inaccurate. There are many digital tools that actually help digital activists to operate in the dark.

Rather than eschewing digital tools entirely, it may be safer for them to use encrypted digital tools rather than offline alternatives. Pidgin is an encrypted chat client, that allows activists to communicate without the government listening in. Guardian is an encryption client for Android mobile phones that allows for encrypted SMS. Wikileaks passes uploaded documents through several national jurisdictions to protect its right to publish and the leaker’s anonymity. Tor and Ultrasurf route Internet traffic in a similar (though more complex) system, to hide the Internet activities of users and allow they to bypass censorship. Proxy servers, though less secure, perform a similar function. Though its security system requires several steps, Hushmail, allows users a less hackable form of email.

Critics who use the danger of public platforms as an excuse to denounce digital activism are doing activists a disservice. It is better to inform activists of more secure online options than to simply push them back into the offline communication tools that repressive regimes have spent generations learning how to crack.

image: xJasonRogersx / Flickr

Proudly powered by WordPress
Theme: Esquire by Matthew Buchanan.