Internet Increasingly Censored

A report released today by the OpenNet Initiative (ONI) concludes that the scale, the scope, and the sophistication of state-based Internet filtering have all increased dramatically in recent years. The survey highlights the tools and techniques used by countries to keep their citizens from viewing certain kinds of online material.

ONI is a collaboration among four leading universities: Cambridge, Oxford, Harvard, and Toronto. The group’s testing was carried out during 2006 and early 2007. ONI used a combination of tools that can remotely test filtering conditions within given countries. The group also relied heavily on local researchers who evaluated Internet conditions from inside certain countries. Some countries, such as Cuba and North Korea, were deemed too dangerous for either remote or in-country testing. But of the 41 different countries tested by ONI, 25 were found to block or filter online content.

“Over the course of five years, we’ve gone from just a few places doing state-based technical filtering, like China, Iran, and Saudi Arabia, to more than two dozen,” says John Palfrey, executive director of the Berkman Center for Internet and Society at Harvard Law School. “As Internet censorship and surveillance grow, there’s reason to worry about the implications of these trends for human rights, political activism, and economic development around the world.”

But it’s not just the sheer number of countries doing content filtering that has grown; it’s also the breadth and depth of material being blocked.

The report discusses three primary rationales that nations have for blocking Internet content. The first is political, which leads to, for example, the blocking of opposition-group websites. The second rationale is social: some countries block pornography and sites dealing with gambling or sexuality issues. The third rationale is national security, which can lead some nations to block online material produced by, for example, extremist groups.

According to the report, China, Iran, and Saudi Arabia remain the top blockers. Each nation filters not just pornography, but also a wide range of political, human-rights, religious, and cultural sites deemed subversive by those countries’ governments.

Other countries are more selective in what they let citizens see or not see. Syria and Tunisia, for example, filter a great deal of political content, while Burma and Pakistan target websites that pertain to national-security issues.

One interesting case is that of heavily wired South Korea, where ONI found Internet filtering limited to one topic: North Korea. “The South Koreans block several North Korean websites,” says Nart Villeneuve, director of technical research at the Citizen Lab at the University of Toronto. “They even tamper with the system so that when you try to access one of those North Korean sites, the URL resolves to a South Korean police page telling you, ‘What you’re trying to access is illegal, and we know your IP [Internet protocol] address.’” (An IP address could be used to locate the computer where the search is conducted, with the ultimate goal of identifying the individual involved.)

South Korea’s approach also speaks to the growing sophistication of the filtering employed by countries. Gone are the days when filtering one blog or one website necessarily meant shutting down, say, all of Blogspot, or an entire domain.

“In the early days, countries used relatively crude blocking mechanisms at the national backbone level, or imposed restrictions upon ISPs that were applied in uneven ways,” says Ronald Deibert, director of the Citizen Lab. “Now we see first and foremost that many countries are using commercial filtering technologies, most of which are made by U.S. companies. That’s providing them with a finer-grain level of service.”

Many countries are also getting better at homegrown filtering, according to the report. Five years ago, most countries would only block English-language material deemed offensive. But as more content has been created in local languages, the report concludes, repressive regimes have had to tweak their filtering technology to keep up.

Deibert also notes that ONI found evidence that filtering has moved beyond websites and into applications. Some nations now block access to programs such as Google Maps and the voice-over-Internet application Skype. Thailand recently blocked access to the video-upload site YouTube.

But most pernicious, Deibert says, is something he calls “event-based” filtering, of which Belarus provides an interesting example. Before the elections in March of 2006, Deibert notes, Belarus wasn’t blocking Internet content by technical means. Instead, the country’s strict laws regarding online content kept many Belarusians critical of the government in check.

Then, at the time of key moments in the election, ONI realized that opposition websites were suddenly inaccessible inside the country. This led Deibert to believe that for just this brief period of time, laws designed to promote self-censorship weren’t enough. The government had indeed started blocking content.

“This is a harbinger of what’s to come worldwide,” Deibert says. “You’ll have filtering just during critical times, such as elections. Countries realize they risk becoming pariahs, and so they’ll find more surreptitious ways of filtering.”

Cambodia recently took this kind of censorship beyond the confines of the computer, when it ordered that cell-phone text-messaging services be cut off during elections. ONI is already thinking of ways to incorporate this kind of filtering into future studies.

“We’re going to have to keep an eye not just on the network, but on the endpoint,” says Jonathan Zittrain, professor of Internet governance and regulation at Oxford University, “because the device you use and how it works, whether it’s a computer or, say, a Blackberry, will have a huge impact on what you can do or not do on the Net, and how easily you can be monitored.”