health + tech in law = mixing chalk + cheese
If you are familiar with the popular platform WeChat, then you’d know that there are differences in the app depending on whether you registered your account with a phone number from China or elsewhere. It was previously known that, due to Chinese censorship regulations, WeChat content was subject to surveillance and moderation. But it was thought that surveillance and content moderation applied only to China-registered accounts.
However, recently U of T’s Citizen Lab released a report that demonstrates this not be the case. The Citizen Lab is a research lab known worldwide for monitoring spyware used against journalists, activists and other individuals.
The Citizen Lab researchers found that data sent from non-China-registered WeChat accounts are seemingly used to train the censorship system.
The researchers based this conclusion on some experiments that involved sending documents and images containing politically sensitive material. As a baseline, documents that had not previously been sent over the platform were not censored to China-registered accounts in real time, although images sometimes were. However, documents and images containing politically-sensitive material that were first sent among non-China-registered accounts would be censored in subsequent attempts to send those same documents and images to China-registered accounts. Further, the researchers found that this occurred even with data from recalled messages.
From a privacy perspective, these results raise a whole slew of questions for non-China-registered accounts. The Citizen Lab researchers conducted three activities to answer at least some of these questions. First, they obtained public documents (e.g. privacy policies) on how Tencent, the company behind WeChat, asserts it handles user information. They then used a series of structured questions to analyze these documents. Finally, the researchers contacted the company’s international data protection office with specific questions. These questions were, in part, to determine whether the company would confirm or deny the collection or generation of non-China registered user data for censorship purposes.
In conducting this policy analysis, the researchers found Tencent’s privacy communications to be inadequate:
Tencent has not only failed to explain to its international users how their communications content is being used to facilitate the censorship apparatus that is applied to China-registered WeChat accounts, but the company has also failed to explain, or clarify, whether international users’ communications content are subject to surveillance that is not associated with the censorship of content that is deemed sensitive in China.
Disturbingly, the researchers note that social media surveillance and content moderation is not limited to WeChat. It occurs across social media platforms, and “[s]urveillance constitutes a fundamental feature of all mainstream profit-oriented social media businesses.”
This report shows that, even with a privacy framework, Tencent is not transparent with how it handles user data. No one reading its policy documents would expect that their data were being used for training a censorship system. Yet one might interpret those privacy terms to include such use after finding out about it. And Tencent’s privacy office did provide a response to a request, but it was incomplete.
What is also concerning is that Tencent may be using the data for other uses, uses that can’t be discovered by experiments like the ones conducted in this case. Further, as noted by the Citizen Lab researchers, surveillance is likely a part of all other social media platforms. And in those cases, there may also be uses of user data (including personal information) that can’t be easily discovered. That laws around the world may provide for privacy complaints to be brought to authorities is of little comfort if one doesn’t even know that such complaints should be brought.