Oops! The input is malformed!
Originally published December 17, 2013
We’ve been hearing a lot lately about extracting business intelligence (BI) from big data, especially from the social media such as Facebook, Twitter and YouTube, and from the analysis of these networks themselves. In an interesting new twist, we are now being told that we might be able to learn a lot about the intent and plans of anyone that attempts to censor social media. This adds another interesting dimension to the business intelligence discipline that we should understand better and explore.
First of all you may ask, “Who censors social media?” Aside from the attempts of most private sites to exclude non-members and restrict offensive postings, many public sites often try to block out maledicta – obscene, blasphemous or vulgar language.
More important, however, is that many governments engage in that practice also. And now it seems that censoring social media has a price.
There have always been rumors about the many regimes that engage in such actions. But there is one organization, the Paris-based Reporters without Borders, that documents and tracks this aspect of cyberspace life. In its 2012 report, they have put together a list of “Enemies of the Internet.” It includes: Bahrain, Belarus, Burma, China, Cuba, Iran, North Korea, Saudi Arabia, Syria, Turkmenistan, Uzbekistan and Vietnam. China, in particular, gets a fair amount of attention as dedicating a significant amount of resources to this endeavor.
Internet monitoring, by the way, takes place primarily with software and hardware made in the USA; and, in general, there are no impediments for U.S. industry to export such products to most countries. Exceptions to this are exports to nations currently under some form of U.S. Treasury sanctions, such as Iran or Sudan. Nevertheless, some of these tools have already been detected in both Iran and Sudan, as well as in other countries that are widely suspected of being abusers of its citizens’ human rights.
Some of the vendors identified by researchers are Blue Coat Systems, whose software has been detected in Iraq, as well as Narus and even Cisco. Narus, a company now owned by Boeing, was suspected to have been used by Egyptian security forces to track activists during the 2011 protests that toppled Mubarak. Cisco was sued by the Falun Gong for having provided the Chinese government with the surveillance technology to track its members. (See “Regimes’ Web tools made in the U.S.A.,” Washington Post, July 9, 2013, p1.)
China has been one of the principal offenders. Beyond the Falun Gong incident, we have surely read about the Chinese hacking into Google, or into U.S. industry and government accounts, or of People’s Liberation Army Unit 61398 and their other elite hacker teams. But now we have some interesting news about the potential price that censors in general, and Chinese censors specifically, may pay for blocking social media.
Prof. Gary King and colleagues at Harvard University have determined that what the censors block and when they block it is intrinsically linked to happenings inside the internal political structures in China, and hence it provides predictive intelligence into that ruling body politic. (See “How Censorship in China Allows Government Criticism but Silences Collective Expression," by Gary King, Jennifer Pan and Margaret E. Roberts, Harvard University, American Political Science Review, May 2013, pp1-18.)
One example of how this works deals with the timing and manner in which the names of dissidents or opponents of the government start being censored from the social media: frequently a few days before they are picked up by the police and arrested. This behavior has now been clearly observed on several occasions.
An early example surrounded the bizarre episode of Bo Xilai, a member of the Central Politburo and head of the Communist Party in Chongqing. Bo was one of China’s princelings and slated for top positions in both national government and the Communist Party. (His father, Bo Yibo, had been a revolutionary leader alongside Mao Zedong and then one of the “Eight Immortals” around Deng Xiaoping.) He managed to fall into disgrace after a top lieutenant and chief of the Chongqing police, Wang Lijun, accused Bo’s wife, Gu Kailai, of involvement in the death of an Englishman over a business deal gone awry.
Prof. King’s censorship-tracking software counts numbers of posts being blocked by censors in “weibos,” the Chinese Twitter-like microblogs. He determined first that five days before Bo fired Wang Lijun, the number of posts mentioning the latter’s name that were blocked by the censors increased dramatically. This happened again prior to the actual downfall of Bo himself and of Gu Kailai’s disgrace after being linked to the murder.
The way King’s software works is that it looks for posts related to any one of a number of sensitive topics, identifies them and then revisits the site repeatedly to measure the number of censor blocks to posts on that topic on a daily basis. (King and his colleagues note that the censors are not shy and simply leave a message “Sorry, the host you were looking for does not exist, has been deleted or is being investigated” and even leave police cartoons on the site.)
In any case, they were able to identify similar censorship trends days before the arrest of artist/activist Ai Weiwei and likewise patterns that preceded politically sensitive events like the 2011 peace agreement with Vietnam that decreased the tensions in their confrontation over oil rights in the South China Sea.
The amount of censorship of these posts is massive. Another study from researchers at the University of New Mexico and Rice University estimates that 12% of all posts on Sina Weibo, one of the two largest microblogs in China, are censored. (See http://news.unm.edu/2013/03/unm-rice-researchers-document-the-velocity-of-censorship/.) This study attempts to understand how the Chinese government attempts to frame or direct public conversation in the social media through censorship, often by teams inside the weibos themselves. And they are fast; the study calculated that it takes Sina Weibo censors about 8 minutes to find and block an offending post (http://www.dailymail.co.uk/news/article-2300674/Astonishing-speed-China-censors-information-internet-30-deleted-half-hour-5-deleted-FIVE-minutes-appearing-online.html).
Monitoring the Internet is done all the time to insure operational effectiveness and efficiency as well as to protect what is usually an enterprise’s most important virtual real estate. When public or private actors take it upon themselves to censor postings in public social media sites, whether it is done for the right or the wrong reasons, they had better be careful since their techniques leave behind a breadcrumb trail of their intentions.
For BI practitioners, this just means one more potential application of our toolbox. When this allows the development of predictive analytics around actual political events in a country that is a human rights violator, then score one for freedom thanks to business intelligence at work.
Recent articles by Dr. Ramon Barquin