WeChat monitors foreign accounts to know what to block in China
Research from Citizen Lab found that sensitive content shared with international WeChat users wound up blocked for Chinese accounts
WeChat, the Chinese messaging app, is systematically monitoring the content sent by international users to build up its censorship algorithms applied against accounts registered in China, a new study has found.
Researchers at Citizen Lab, an academic research lab at the University of Toronto, determined that WeChat screens images and documents shared by accounts registered outside China after they are sent, then adds the digital signature – or “hash” – of any files deemed sensitive to a blacklist. Those files then cannot be sent or received by China-registered users.
Numerous studies have identified WeChat’s use of censorship tools against China-linked accounts, but this research provides proof for the first time that non-China registered users are also swept up in its surveillance apparatus.
Published Thursday in a report called “We Chat, They Watch,” the Citizen Lab findings are likely to add fuel to existing concerns, particularly in Washington, about data security and the international reach of information control tools used by Chinese tech companies.
Earlier this year, Citizen Lab found evidence that WeChat blacklisted more than 500 keyword combinations relating to the coronavirus outbreak since the start of January, including text that referred to Dr Li Wenliang, the Wuhan whistle-blower.
“If users [of WeChat] weren’t concerned before, they should be very concerned now and re-evaluate the risks of using this application,” Ron Deibert, the Citizen Lab director, said of the new findings.
Beyond matters of data security, the issue was a moral one, said Deibert: “I would urge international users to consider that, as you use this platform, you’re actually helping to strengthen digital repression in China.”
Neither WeChat nor Tencent, the Shenzhen-based technology giant that owns the app, responded to requests for comment.
Researchers at Citizen Lab ran a number of tests to prove the existence of surreptitious content surveillance, which is more difficult to identify than outright censorship.
In one, researchers sent a sensitive image – a cartoon of the late human rights advocate and Nobel laureate Liu Xiaobo – from a non-China registered account into a chat group with other international users. When the researchers sent a file with the same “hash” into a group chat that contained China-registered users moments later, they found the image was not received.
To ensure that the file had not been blocked in the second group due to WeChat’s real-time censorship mechanism – which it deploys against some images – researchers completely altered the image’s appearance to remove any sensitive elements while manipulating the file so as to keep its “hash” unchanged.
The researchers carried out that experiment 20 times, each one using a unique “hash” to ensure it was not already in WeChat's database.
The tests, which also included an experiment based on statistical analysis, ran between November and January. The researchers could not determine when the surveillance mechanism was introduced.
As is the case with many internet platforms in China, it is all but impossible to ascertain what – if any – restrictions are enacted at the explicit direction of government or Communist Party officials and what come from companies taking steps on their own based on their reading of cyberspace regulations.
Within China, WeChat’s commonly used functions extend far beyond messaging: it serves as a multi-tool for consuming news, paying for goods and services, hailing taxis and more.
As of 2019’s third quarter, the app had more than 1.15 billion monthly average users worldwide, according to the company, which does not disclose usage data by country.
On Google’s Play Store, which is largely inaccessible within China, the app has been downloaded more than 100 million times. The app is also available both within and outside China on Apple’s App Store.
Elsewhere, however, language regarding how user data can be used by the app is more vague. Citizen Lab researchers noted that WeChat International states in its terms of service that the app reserves the right to use user content “for the purposes of providing, promoting, developing and trying to improve WeChat and our other services.”
Deibert, of Citizen Lab, said, “If one were to examine closely those terms of service, you’d have no idea that this type of political surveillance is going on.
“That's a pretty serious violation that should trigger the attention of government regulators.”
Sarah Cook, an expert in Chinese media and information control, said the Citizen Lab findings would “trigger a lot of red lights” among national security circles.
“This is going to amplify calls for greater scrutiny, especially in the United States, of WeChat,” said Cook, a senior research analyst at Freedom House, a democracy-focused think tank that receives US government funding.
While Citizen Lab had identified one application of WeChat’s content surveillance – training its censorship algorithms – Cook said it could also be applied to a number of other functions, such as “identifying certain users and creating a portfolio about them, [or] feeding other aspects of the [Chinese Communist Party’s] transnational repression apparatus.”
Despite those concerns, Cook predicted that usage of the app among the Chinese diaspora would largely continue, given its crucial role as a bridge to those in China.
“They don’t have the liberty to stop using WeChat,” she said. “Because for a lot of people it’s a lifeline.”