http://www.scmp.com/tech/apps-gaming/article/1840933/rumours-politics-and-corruption-most-likely-topics-be-censored

Rumours, politics and corruption most likely topics to be censored on Chinese messaging app WeChat

James Griffiths

PUBLISHED : Monday, 20 July, 2015, 12:02pm

Messages that spread rumours regarding sensitive topics, or those that comment on corruption among Chinese officials or mention Chinese politics, are among the most frequently censored posts on Tencent’s WeChat mobile messaging platform.

A new report by the University of Toronto’s Citizen Lab illustrates how censorship takes place on the most popular real-time messaging app on the Chinese mainland.

It suggests a change in tack by Beijing, as previous research on other Chinese messaging and social media platforms has shown that censors focused more on posts encouraging protests and other forms of collective action.

Past studies showed protectors of China’s so-called “Great Firewall” were not as interested in comments on government policies, or the ongoing anti-corruption campaign of President Xi Jinping, as they appear to be when it comes to WeChat.

Keywords relating to corruption made up five of the top 50 most sensitive keywords in the latest report, according to its author Jason Ng, a research fellow at Citizen Lab.

The results also hint at the impact of Beijing’s “anti-rumour campaign,” launched in 2013, which threatened rumourmongers with gaol time.

Previous studies have largely focused on microblogging platform Sina Weibo. Ng said he chose to examine WeChat as it has been overlooked despite exploding in popularity in recent years.

The messaging platform boasts over 500 million monthly users globally. Ng’s study focused on WeChat’s official accounts feature, which enables businesses and individuals to post public messages.

Registering for an official account is somewhat labourious. Users must provide a copy of their identity card, for example. But as of the end of 2014, there were over 8 million registered accounts, many with hundreds of thousands of subscribers.

“This turned WeChat from a private communication tool into a media platform,” wrote media scholar Hu Yong last year.

Ng’s study examined 36,000 posts by public accounts between June 2014 and March 2015, monitoring them over time and tracking whether they were deleted.

While it is known that WeChat censors certain keywords, preventing users, for example, from sending messages about the banned Chinese religious movement Falun Gong, Ng’s study is the first to look into how public account posts are censored.

In January, more than 100 public accounts were shut down by the authorities for “spreading distorted historical information”. The move came after tech giant Tencent promised to take action to clean up its platform in mid-2014.

Around 4 per cent of all the posts examined by Ng were censored.

“About 40 per cent of the deletions … were deleted while the author’s account was still active, meaning the content of the post itself was flagged for deletion,” rather than the user being banned, he wrote.

The content of the posts, particularly those that focused on rumours, were not easily discernible by an algorithm or keyword search, suggesting that a human hand was at play, Ng said.

In other words, either other users were flagging posts as offensive, or censors employed by WeChat were deleting them manually.

“WeChat appears to tread in ambiguous waters [regarding censorship], forcing us to ask questions like: are those who object to a given post regular users, or are they employees of WeChat or the government?”

Yet the results suggest that censorship is less common on WeChat than on Weibo.

A series of studies in recent years found that between 12.8 per cent and 16.3 per cent of Weibo posts are deleted.

However, Ng said it is more difficult to capture WeChat posts in real-time, adding that his survey missed posts that were deleted shortly after being published.

Tencent did not respond to multiple requests for comment.

In 2014, it told the Wall Street Journal that it took unspecified measures against “offensive and abusive activities,” based on user feedback and “in line with relevant guidelines on illegal and pornographic content”.