Trends in Cybercensorship

A talk with an editor of Access Denied: The Practice and Policy of Global Internet Filtering about the recent rise in Web censorship

When the military government of Burma cracked down violently on protests by tens of thousands of Buddhist monks (, 11/7/07), it wanted to do so in private. So, to keep photos, videos, and descriptions of last November's clashes from leaking out, the government decided to halt the flow of information by cutting off its citizens from the Internet entirely.

Not long after, when a YouTube video said to be offensive to Islam caught the attention of Pakistani regulators, the government there moved to block access by its citizens—not just to the video, but to all of YouTube. This took place only weeks after the Pakistani government had ordered blackouts on cable TV news outlets and newspapers (, 11/27/07).

In China, a country infamous for a "Great Firewall" that strictly controls what its citizens can read online, outside reports and information about the recent crackdown on protests in Tibet have been blacked out (, 3/17/08). Once again, regulators targeted YouTube after clips of protesting monks in the capital city of Lhasa began to circulate.

Blocking access to the Internet has become a standard "policy" option around the world. The editors of a new book, Access Denied: The Practice and Policy of Global Internet Filtering, published this month by the MIT Press, say this type of filtering is taking place in at least 25 countries—in Asia, North Africa, and even Europe. And U.S. technology companies are finding themselves torn between the open principles of their own country and the realistic need to do business in countries with restrictive regimes.'s Arik Hesseldahl discussed the topic with one of the book's editors, John G. Palfrey, a Harvard University law professor who is also executive director of the Berkman Center for Internet & Society. Edited excerpts of their conversation follow:

You and your collaborators on this book conducted a survey of cybercensorship in 40 countries around the world. Can you summarize your findings?

The thesis of the book is that there is an undeniable increase in the extent to which states are making it policy to censor the Internet. The data show that five or six years ago there were a few states that were doing it, China and Saudi Arabia most famously, and today we've found more than two dozen states that are carrying out state-mandated Internet censorship.

On what grounds do they justify these policies? National security? Morality?

It's all of the above really.… In places like Iran or China or Uzbekistan, there are multiple reasons for the extensive filtering going on. One is political. Those in power don't want dissent to spread so easily as it can on the Internet. Another is cultural. They don't want so many people seeing pornographic images. In other cases there are very specific reasons. Like in Turkey, which has just happened, they are blocking entire services like YouTube and WordPress when even a single video is critical of Turkey's founder, Kemal Ataturk. Another example is Thailand. When there is a single Web page or video critical of the king, it's taken down. There are also some western countries that do some censorship. Germany and France block information about Nazi paraphernalia and Holocaust denial.

What is the least restrictive country?

There is no one least restrictive country.… I would put the U.S. on the more open side of things, but there is some censorship that takes place in schools and libraries.

What responsibilities do U.S. tech companies—telecom hardware makers such as Cisco Systems (CSCO), software companies like Microsoft (MSFT), and service providers including Google (GOOG) and Yahoo! (YHOO)—have when they face the question of facilitating or enabling this censorship?

We have a chapter in the book that addresses this. I would distinguish between some of the companies you just mentioned. Google, Microsoft, and Yahoo are involved with us in creating a set of principles to determine some ethical guidelines for American and western companies to respond to these concerns. There are some companies that are not party to these efforts. Cisco, Juniper [Networks (JNPR)], and Nortel [Networks (NT)], for instance, are not.

Tell me about the principles you're working on. How should these companies behave in these cases?

The principles should set a bar that is at least as high as what the best company is doing as far as protecting speech and privacy online, and would include a mechanism to ensure that those companies who are signatories are adhering over time. These are well-meaning U.S. technology companies that want to do business in markets around the world where censorship and surveillance are taking place. But they need an ethical guide for how to respond to requests from these governments to turn over information or block search results. You need to obey local law in the places where you do business, but you also need to remain true to the principles of the place where you are chartered. It's a core problem of doing business in a global economy.

So if company refuses to sell a bunch of routers to the Chinese government because it knows they're going to be used for surveillance or censorship, that would mean a big loss of revenue. China would just switch to a more cooperative supplier, right?

More likely China would simply turn to one of the local companies that have reverse-engineered a Cisco or Nortel router and are selling them to their government.

So it seems that one approach might be to go ahead and sell the routers, but not program them for network surveillance. Instead, they could leave that to a local, third party?

That's true. I think you have to distinguish between two things. One is sales to governments where you're in fact taking a dual-use technology and then not providing any further service, and situations where you get an ongoing series of requests, either to add software to hardware or to take some specific result out of a search result or turn over someone's personal information. We have a chapter in the book where we break down the different categories and create a typology of responses. That's a long-winded academic way of saying you're right.

In your surveys, what was the most blatant or shocking case you came across?

I think the most striking one to me was in Burma. When the monks were protesting in the fall, the military government tried to shut down the Internet entirely. What we saw in that story was a recognition by an authoritarian government of the Internet as the mode by which most people were going to get the story of what was happening on the ground to the outside world. The other thing that happened, as we did a lot of testing around that time, was that the military junta also brought the Internet back on during the middle of the night. They would have it back on from, say, 2 a.m. to 5 a.m. Our supposition was that the Internet was also crucial for the military itself, to keep in touch with the outside world. I think there's tension for these states that are carrying out state-mandated Internet filtering: They want to have the Internet they want and not the Internet they don't. So far that has proven impossible.

Before it's here, it's on the Bloomberg Terminal.