Researchers Made a Fake Social Network to Infiltrate China's Internet Censors

It's the first time anyone has had a clear understanding of what China's government censors.

Aug 21 2014, 8:05pm

Image: Tim Quijano / Flickr

It's no secret that China has been censoring and controlling the information its citizens can send and receive, especially on the internet. But, until Harvard researchers recently broke into the system, no one knew exactly how it worked.

Today, researchers from Harvard and the University of California San Diego released a report in Science that reads more like a spy novel than a scientific paper. 

In order to get inside China's notorious filter, researcher Gary King and his team created dozens of shill accounts and posted hundreds of messages on China's most popular social networks to see what would be filtered. But then, the team went one step further, creating its own fake social network in order to gain access to the programs used to censor content, so it could reverse-engineer the system.

"From inside China, we created our own social media website, purchased a URL, rented server space, contracted with one of the most popular software platforms in China used to create these sites, submitted, automatically reviewed, posted, and censored our own submissions," King wrote. 

the government is promoting innovation and competition in the technologies of censorship

"We had complete access to the software, documentation, help forums, and extensive consultation with support staff; we were even able to get their recommendations on how to conduct censorship on our own site in compliance with government standards," he continued.

These are the two basic methods of censorship. Image: Science

After running a series of tests on both Weibo (China's Twitter) and its own fictitious social network, King and his team came to a surprising conclusion: It's generally OK to criticize the government in China, as long as you aren't inciting others to act.

The country, he says, is worried about uprisings, protests, and anything that could spur real-life action, not government criticism. Many of those posts are caught in an auto keyword filter or are deleted by manual censors, who have dozens of options for banning IP addresses, users, and deleting or hiding social media posts.

While China is essentially autocratic, King says that it's a "responsive" autocracy, meaning the government is cool with people criticizing their local leaders. In fact, it actually serves the government well if they do so.

"The knowledge that a local leader or government bureaucrat is engendering severe criticism—perhaps because of corruption or incompetence—is valuable information," he wrote. "That leader can then be replaced with someone more effective at maintaining stability, and the system can then be seen as responsive."

So, that's why criticism is allowed, but calls to action aren't: The most commonly censored posts include words like "masses," "incident," "terror," "go on the streets," and "demonstration."

A list of the various options available for censorship. Image: Science

What makes King's report truly notable is sorting out how China censors content, and on that, China has found, perhaps, a Silicon Valley-esque solution: Allowing each social media network to attempt to disrupt censorship itself. 

"We conclude that the government is (perhaps intentionally) promoting innovation and competition in the technologies of censorship," he wrote. "Such decentralization of policy implementation as a technique to promote innovation is common in China."

It makes sense, if you think about it. China can maintain the illusion of not censoring posts if it's implemented on a slightly different basis by hundreds of companies throughout the country. At the same time, companies know that they have to censor content that incites people to action, allowing the Chinese government to keep people held down without the illusion that it's actually doing so.