China's Precision Censorship Machine Allows Some Controversial Keywords, But Blocks Combinations Of Them

from the politically-problematic-images-also-a-no-no dept

China's censorship of the Internet is both impressively thorough, and yet surprisingly subtle at times. For example, we've already written about ways in which the boundary between censored and non-censored is often vague, which paradoxically encourages people to be even more cautious than they would be with well-defined limits. But hidden among all the uncertainty, are there perhaps some fixed rules about when posts will definitely get censored?

A team of researchers at the University of Toronto's Citizen Lab decided to find out by investigating one of the topics considered most controversial by the Chinese authorities, the so-called "709 Crackdown." This refers to a major government clampdown that began on July 9 in 2015, when more than 250 Chinese rights lawyers, law firm staff, activists, and their relatives were detained by public security agents across China. Internet users are understandably keen to discuss this important event, and many of those conversations take place on the main blog site in China, Weibo, and using the messaging service WeChat, which is even more popular. But as the researchers discovered, those online conversations were subject to subtle but consistent interference:

as our experiments show, a good portion of that discussion fails to reach Chinese users of WeChat and Weibo. Our research shows that certain combinations of keywords, when sent together in a text message, are censored. When sent alone, they are not. So, for example, if one were to text Mainland China or Wang Quanzhang's Wife or Harassment on Relatives [all written in Chinese characters] individually, the messages would get through. Sent together, however, the message would be censored.

Moreover, for the first time the researchers discovered censorship not just of text, but of images too:

In addition to a large number of censored keyword combinations our tests unearthed, we also discovered 58 images related to the 709 Crackdown that were censored on WeChat Moments for accounts registered with a mainland China phone number. (For accounts registered with a non-mainland China phone number, on the other hand, the images and keyword combinations go through fine).

Neither of these observations is earth-shattering in itself, but they do add usefully to our knowledge of the intricate clockwork of China's mighty censorship machine.

Follow me @glynmoody on Twitter or identi.ca, and +glynmoody on Google+


Reader Comments

Subscribe: RSS

View by: Time | Thread


  1. identicon
    Anonymous Coward, 18 Apr 2017 @ 10:56am

    "Microsoft notes FISA orders are on the rise. Of course, its reporting is limited to useless "bands," so the only thing that can definitely be determined is Microsoft's FISA interactions have at least doubled."

    From a different post about a different country and a different thing but why censor when you can just collect it all?

    reply to this | link to this | view in thread ]

  2. icon
    CypherDragon (profile), 18 Apr 2017 @ 7:52pm

    Data classification

    Sounds like they are using some variant of a data leakage protection (DLP) product for the censoring. One of the key features with most DLP products is that you can set thresholds for what triggers the rule. Eg, I want to block anything with the words "TechDirt" "Censorship" "Moody" and "China" but only if it has all 4 of those words in it. Simple to do with a DLP policy. Alternately, I could have a list of keywords, and have it trigger the policy once it hits a certain count.

    These systems are fairly robust, but they aren't without their flaws. Also, the system will only be as good as the policy makers can target their policies.

    reply to this | link to this | view in thread ]

  3. identicon
    Anonymous Coward, 19 Apr 2017 @ 12:22am

    A system that blocks specific combinations of keywords... smells like My_Name_Here's involvement.

    reply to this | link to this | view in thread ]

  4. identicon
    Anonymous Coward, 23 Apr 2017 @ 4:56pm

    Trump is thinking: "Why can't we do this?"

    reply to this | link to this | view in thread ]


Add Your Comment

Have a Techdirt Account? Sign in now. Want one? Register here
Get Techdirt’s Daily Email
Use markdown for basic formatting. HTML is no longer supported.
  Save me a cookie
Follow Techdirt
Techdirt Gear
Shop Now: Copying Is Not Theft
Advertisement
Report this ad  |  Hide Techdirt ads
Essential Reading
Techdirt Deals
Report this ad  |  Hide Techdirt ads
Techdirt Insider Chat
Advertisement
Report this ad  |  Hide Techdirt ads
Recent Stories
Advertisement
Report this ad  |  Hide Techdirt ads

Close

Email This

This feature is only available to registered users. Register or sign in to use it.