Early last month, some Chinese social media users had an inkling that their discussions about a spreading deadly pathogen were being censored.

Behind the Great Firewall, internet users said they suddenly saw their WeChat accounts suspended for “spreading rumors,” if they were given a reason at all. Now a new report is giving us a glimpse into how China’s most popular messaging app might be blocking online chats about the new coronavirus epidemic.

The University of Toronto’s Citizen Lab created three new WeChat accounts -- two in Canada and one in China -- to test if any keywords were being banned. Researchers used one Canadian account to send out chats with words taken from news stories. They then checked to see if any of the messages reached the other Canadian account but not the Chinese one.

The Citizen Lab also looked into YY, a live-streaming platform that has a keyword blacklist built into the app to filter text chats. This is in contrast to WeChat, which performs censorship on a remote server.

In response to our inquiry, WeChat’s owner Tencent says, “Our operations in China are guided by local laws and regulations related to internet content, and like other companies, we comply with regulations, laws and legal requests in all countries and markets where we operate.” 

The company added that since the coronavirus outbreak, it has launched new tools to combat online misinformation. It says it’s in the company’s interest to “ensure people in China and around the world have access to reliable information on the coronavirus.” 

JOYY, the owner of YY, did not respond to our request for comment. 

A nurse works in an ICU ward for coronavirus patients in Wuhan. (Picture: Xiao Yijiu/Xinhua)

Over the course of six weeks, researchers found that 516 coronavirus-related keyword combinations were filtered on WeChat. All the terms tested were in Chinese.

Predictably, blocked terms include words critical of the government (e.g. Wuhan + Conceal Epidemic) and speculation about the outbreak (e.g. Wuhan pneumonia epidemic out of control). Some contained references to Li Wenliang, a doctor who was detained by police for warning classmates about the disease. He later died of the disease after treating Covid-19 patients.

But surprisingly, neutral and factual information also appeared to be censored, such as messages mentioning human-to-human transmission of the virus in Wuhan, the epicenter of the epidemic.

On YY, 45 coronavirus-related keywords were added to its blacklist on December 31, with five removed later in February.

In China, the onus falls on internet platform operators to remove chatter about illegal or politically sensitive content. Companies who fail to comply could be reprimanded or face other serious consequences, such as temporary or permanent removal of their apps. Experts say the practice can sometimes lead companies to over-censor before scaling down.

In 2018, for instance, popular microblogging platform Weibo reversed a decision to ban LGBT content after criticism from Chinese state media.

UPDATE (March 6, 2020): This story was updated to include Tencent’s response.