Censorship in China

China’s WeChat censoring ‘sensitive’ photos, not just text, study shows

Study the first systematic research into image censorship of one topic on the hugely popular Chinese instant messaging app, according to researchers

PUBLISHED : Thursday, 13 April, 2017, 1:30pm
UPDATED : Friday, 14 April, 2017, 10:13am

China’s popular instant messaging app WeChat is filtering not only keywords, but also images deemed sensitive, without users’ knowledge, according to researchers in Canada.

The findings from the University of Toronto’s Citizen Lab came after its researchers studied how the app removed discussions and posts concerning a nationwide crackdown on rights lawyers and activists in China.

There has been anecdotal evidence of picture filtering on WeChat, but the findings are the first systematic study of image censorship on a particular topic on the app, according to Lotus Ruan, one of the report’s lead authors.

How to avoid getting your WeChat messages censored both in and outside China

WeChat has more than 800 million users in China and is controlled by the technology firm Tencent Holdings. The company has yet to respond to a request for comment about the researchers’ findings.

Citizen Lab looked at the censorship of images related to the “709 crackdown” on the Moments feature on the social media platform, similar to the Timeline on Facebook user’s profiles.

More than 200 lawyers and activists were interrogated or detained from July 2015 in what rights groups said was the harshest crackdown on campaigners on the mainland for years.

The researchers found 58 image search results for the Chinese term for “709 Crackdown” triggered censorship on Moments.

The images blocked included infographics on the crackdown, profile sketches of affected lawyers and their relatives, or images of protesters holding placards accusing the authorities of torturing one of the attorneys.

WeChat allegedly censoring photos from Hong Kong protests

If an image posted by a user is censored, it does not appear on the feeds of their friends and both will have no idea that the content has been censored as no notification is given.

The researcher said censorship of images on Moments was significantly higher than in group chat on the social media platform.

Users who registered their accounts with a mainland China phone number were subject to the censorship. People with accounts using numbers outside the mainland avoided the filtering, the researchers said. This is consistent with the research team’s previous findings on keyword filtering on the app.

The Chinese government has repeatedly said it must have the right to manage online content deemed harmful.

Censors take new tack, WeChat study finds

The authorities issued a regulation last September specifying that messages and comments on social media platforms such as Moments can be collected and used as “electronic data” in legal cases.

State-run media have reported since that WeChat users have been arrested for insulting the police or threatening to blow up a government building after using the Moments platform.

“The authorities are paying more attention to [Moments] because of its influence and ability to reach a large audience,” said Ruan, a research fellow with the University of Toronto’s Munk School of Global Affairs. “I wouldn’t be surprised if image filtering has been applied in other politically sensitive cases.”

Ruan said the filtering and censorship of images did not seem overly sophisticated, although the report said that if pictures were slightly amended, they would still be filtered.

How China’s highly censored WeChat and Weibo fight fake news ... and other controversial content

“For example, while a couple of images of people holding placards saying, ‘Oppose Torture, Pay Attention to [rights lawyer] Xie Yang’ were censored, we didn’t experience censorship when we produced and submitted our own image,” said Ruan.

The report said more work was needed to identify the kind of algorithms WeChat was using to identify and filter images deemed sensitive.

The team identified 44 keywords that were banned in group chat on the app, including the names of 13 people affected in the crackdown, including lawyers and their wives.