To understand the Chinese censorship system, Gary King, a professor of political science at Harvard University was tricky to intrude inside. According to the Technology Review website, this professor and two colleagues created a fake Chinese social networking site in order to understand how censorship really operates in this country.
We thus discover that the Chinese e-censorship is not based on a centralized system, but rather on a very dynamic capitalist market: companies are actually competing to offer the best technology and the best tools for removing content. There is actually no censorship written guide by the Chinese government, which explains the forbidden words, such as Tibet or Tiananmen. Rather these companies are setting for themselves their own guidelines, not to fall within the scope of government repression, which can lead to the closure of a site.
The kit provided to the researchers to manage their site had some sophisticated tools that can automate online censorship. New posts to the site could even be filtered based on keywords upstream of a manual control. They could be managed specifically according to their length or position where they appear on the site. This software also allows you to stop the publication of posts so that anyone could read it before you publish it or not. It also allows a specific person to be further spotted for censorship, according to his or her IP address and internet reputation, said Gary King.
According to the latter, a company needs to hire two to three censors for 50,000 users to satisfy the Chinese government’s cybercensorship. It considers that in China between 50,000 and 75,000 people used to censor websites.
Reuters has also recently published information devoted to the Chinese social network Weibo use of censorship. The software used by Weibo would scan all posts on its social network before publication. Only a fraction should be considered sensitive to be read by a censor, who will decide whether or not to validate the post . In 24 hours censors thus treat approximately three million posts.