Google, Censorship and China: An Experiment

Back on March 7, 2007, I made an oblique reference to China and censorship in this post. What followed was a remarkable series of events that raise troubling questions about Google, censorship and freedom on the internet.

To start, I should point out that this blog is hosted by Google through its free blogger service. In addition, I use Google Analytics to track the traffic to it. This also is a free application through Google that allows a user to analyze site traffic in a variety of ways, including the geographic location of visitors.

In the original March 7 post, at the bottom, is this sentence:

"We would expect this type of ruling from China. But not from France."

It follows a paragraph in which I criticize the French government for passing legislation that makes some forms of citizen journalism illegal.

That's it.

Five days later, on March 12, two hits arrived on my website traffic report from within China. What was immediately interesting to me was that there was no evidence of how they got to my blog. They did not arrive by search or from a link, but directly. This usually indicates that someone was sent a link that brought them into the site from an email or some untraceable source.

On the following day, March 13, eight more visitors came to my site from within China. They were fairly equally distributed across eastern China, all in large cities.

Once again, on March 18, several more hits came to my site from within China.

In every case, the visits from China came directly, without referral. I immediately assumed that I was being monitored by Chinese censors, and decided to see what I could discover by looking back at them using Google's Analytics tools.

Using the mirroring website, GreatFireWallofChina.org, I tested my site to see if it had been blocked from their servers inside the country. The test came back that it had indeed been censored. (Addendum: The GreatFireWallofChina.org site is down, perhaps permanently. Thanks to Michael Shaw at BAGnewNotes for a new link to an even better tool.)



It was my assumption then that Google's operation in China, Google.cn, had notified the Chinese authorities to monitor my blog, that they had done so for a few days, and had decided then to censor my site. Since March 18, 2007, until now there have been no visitors from China visiting my site. At this time I also tested several other blogspot.com urls belonging to friends' blogs and all tested that they were visible from within China, so I could rule out domain level blocking at that moment.

Having seen a sequence of events emerge:

1. That I posted a politically critical post about China and freedom of speech.
2. That observable traffic to my site from within China followed.
3. That my site was then censored from within China's firewall.

I concluded that it would be interesting to deliberately provoke censorship in an experiment and observe the activity that it caused. Using the same Google account, I created another blog, Porcelainwatch, which I filled with Latin text copied from an online version of text by Horace, assuming that ancient Latin would not immediately be offensive to Chinese censors. After a week of posting daily bits of text, I started to insert English words that I thought the Chinese censors would respond to. But no traffic from within China was observable.

At this point I thought I should check the availability of my site to the Chinese to set a baseline for Chinese traffic, so I ran the url through GreatFirewallofChina.org and found that it had already been blocked. So far no visitors other than myself had visited the blog.

Since there had been no traffic to my blog at all, it appeared that PorcelainWatch had been censored from the moment that it was created because it was created using the same account as this censored blog.

To test this, I created eight more blogs using the same Google account. Each blog tested as blocked from within China.

Carrying the experiment further, using the same computer, I set up a new account with Google using different names and email addresses. From this account I set up five new blogs. All five also tested as blocked from access from within China. One can be seen here.

From this it would appear that the second account, created using the same computer as the first account, was also censored from China from it's outset. I had used entirely different names and email accounts to create this second account. The only thing that it shared with the first was an IP address from the same computer.

Conclusions:

This raises far more questions than it answers. At question here is how much information the Google parent company here in the United States was directly communicating to the Chinese version of Google, and how that information was/is being used to censor information on the web. It seemed apparent to me that as a Google user once censored from access from within China, I was subsequently banned from creating content that could be communicated into China using Google's services. Not only was this blog censored, but any other blogs created with my Google account were also censored, and a different account created on Google from the same computer as the first was also censored.

Currently it appears that most social media sites are banned in China as broad policy, so it may be hard to track how this is working. But the disturbing question remains, at what point did/does Google share private information from users inside the United States with the Chinese government, and what are the clear boundaries and protections that are in place for users here in the US to avoid international censorship and to keep their private information from being exchanged without their knowledge or permission? Google has said that to play in the Chinese market, they have to play by the Chinese rules. After traveling several thousand miles inside of China, using public internet facilities across the country, I assure you that the Chinese rules are incompatible with our standards of free expression and are incompatible with any reasonable view of the future of freedom on the internet.

What is extremely worrisome is that if Google is sharing US user data with Chinese authorities, then there is the option for the Chinese government to exert reverse pressure to start controling western content. Since Google keeps all of their algorithms secret, there is no way for us to know that the content is open and free. It certainly is not open and free within China and Google has already demonstrated that they are willing to censor when under economic pressure from the Chinese government. So far Google would have us believe that the censorship is a one way street. Information going into China is being censored while Google for the rest of the world operates freely and open. But what if China were to exert enough economic pressure to convince Google to start applying some of the censorship elsewhere? Google has already opened the door for that possibility.