‘What they’re doing is called distraction’: Chinese spam dilutes Twitter feeds on protests

A Twitter search on November 30, 2022 using the hashtag locator “Beijing #北京” brings up a provocative tweet (left) of a young woman as the first search result – but no sign of spreading protests against the country’s zero-tolerance Covid restrictions as shown in a photo taken of Beijing (right).
A Twitter search on November 30, 2022 using the hashtag locator “Beijing #北京” brings up a provocative tweet (left) of a young woman as the first search result – but no sign of spreading protests against the country’s zero-tolerance Covid restrictions as shown in a photo taken of Beijing (right). © Twitter / @GFWfrog

As protests spread across China over the country’s tight anti-Covid restrictions, Twitter users searching for information on the affected cities were instead flooded with provocative tweets showing young women in suggestive poses advertising “dating” services. 

Advertising

Starting Sunday, November 27, anyone searching for tweets to find information about protest hotspots like Beijing, Shanghai and Hangzhou will have scrolled through pages and pages of spam content, making it almost impossible to find useful information and share it. As a result, images of the protests were spreading more slowly.  

The FRANCE 24 Observers team analysed the Chinese posts, which were still flooding Twitter feeds on December 2, even after the Chinese authorities eased restrictions in cities including Shanghai and Guangzhou.

The tweets were posted every couple of seconds from hundreds of different Twitter accounts, most of them created in the last two months, with few or no followers – common signs of artificially created or fake Twitter profiles. 

>> Read more on The Observers: How to investigate a Twitter account or suspicious tweets

The tweets, tagging different cities, had similar, sometimes identical, wording, often random words or meaningless sentence fragments in Chinese and English, and the photos posted appeared countless times.

These posts come from different accounts that carry the same Twitter name  ‘小女人22’ but are tagging two different cities: Chengdu (left) and Beijing (right).
These posts come from different accounts that carry the same Twitter name ‘小女人22’ but are tagging two different cities: Chengdu (left) and Beijing (right). © Observers

The same photos are posted over and over again by different accounts, clogging up the Twitter stream. The accounts shown below may have different Twitter handles but they all have the same profile picture suggesting that this is spam content. 

This photo of a woman sitting on an outdoor couch has been tweeted by five different accounts all with the same profile picture. The posts tagged Xuzhou (top left), Shanghai (top right), Changsha (bottom left), Guangzhou (bottom right) and Beijing (right) - all cities that were seeing protests against Covid restrictions.
This photo of a woman sitting on an outdoor couch has been tweeted by five different accounts all with the same profile picture. The posts tagged Xuzhou (top left), Shanghai (top right), Changsha (bottom left), Guangzhou (bottom right) and Beijing (right) - all cities that were seeing protests against Covid restrictions. © Observers

Experts suspect that the Chinese government could be involved with the tweets, but there is no evidence in recent findings to show any direct link. Further investigation and collaborative effort with the platforms would be required to identify the source of this network and whether it is coordinated activity.

‘It appears that some of the previous accounts have been suspended’

Digital investigator Benjamin Strick of the Centre for Information Resilience collected more than 3,000 tweets using the location tags of three cities spelled in Mandarin: Beijing #北京 , Hangzhou #杭州 and Chongqing #重庆. 

He kept a log and posted his findings on Twitter, calling the accounts “dens of spam”, some of which were interacting with one another, and others that were simply “lone accounts”. In his Twitter thread, Strick shows that the tweets have a minimum audience engagement on the platform and that there are common traits which are found among them.

A twitter thread posted by online investigator Benjamin Strick (@BenDOBrown) showing the activity and interaction of 3000 collected spam tweets tagging different Chinese cities, and the common traits amongst the tweets.

On the morning of Friday, December 2, he noted that there had been over 1,200 tweets published in just 30 minutes under the hashtags of #上海 (Shanghai) #南京 (Nanjing), # 杭州 (Hangzhou), which had also been the case throughout the week.

In an update on his research, Strick noticed that “some of the previous accounts have been suspended”. He explains that this is the systematic nature of spam networks: accounts will be suspended and then new ones created. This distraction campaign shows a breach of Twitter’s policies around spam and hashtag hijacking and Twitter is known for taking down such posts. 

Two posts by Benjamin Strick (@BenDoBrown) showing the 2,000 uploaded tweets using the hashtags of Hangzhou #杭州 (above) and a graph demonstrating that the tweets are gaining minimum interaction.

‘Anyone with a sense of literacy can see that these are just spammy accounts’

When asked whether or not this method of distraction is effective, Strick said that it didn’t tend to work well.

This method of distraction isn’t really successful, as anyone with a sense of digital literacy can see that these are just spammy accounts but it is a hindrance to human rights investigators like me that rely on footage circulated on platforms like Twitter. When I go to those cities and look at the latest content, it’s all dating spam obscuring the videos of protestors being beaten and detained by police in China, so it’s a pretty serious issue, and Twitter should be able to mitigate the spread of the spam.

The FRANCE 24 Observers team took a closer look at who exactly is tweeting spam content. The majority of these Twitter accounts were created in the last month (November 2022), had zero followers and were either following a handful of randomly chosen accounts or none at all. 

These four accounts all joined Twitter in November 2022 and have 0 followers. Two of them (top left and bottom left) have the exact same bio.
These four accounts all joined Twitter in November 2022 and have 0 followers. Two of them (top left and bottom left) have the exact same bio. © Observers

Sex-related images from China rarely appear on social media since a Chinese government crackdown on vulgar and pornographic content in the 2000s, making it unlikely that tweets advertising sex services using Chinese hashtags are sent by individual people. 

Professor Gary King, director of Harvard’s Institute for Quantitative Social Science and an expert on computational social science, says the tweets are a form of “distraction” designed to divert attention from the protests.

What they’re doing is called distraction rather than engaging with the people that have said bad things. It’s very hard to persuade somebody to change their view of things. It is not so hard to get them focused on a new thing, to get them to change their object of attention. The idea is to dilute the bad message and the Chinese government does that at a scale.

‘The idea of controlling information is not going to go away’   

On November 30, a tweet posted by Singaporean journalist Melissa Chen reported that Chinese social media users were complaining about their China-made Huawei phones automatically deleting videos of the protests in their gallery.

Tweet by Melissa Chen @MsMelChen saying that Chinese social media users reported having videos of protests automatically getting deleted on their Huawei phones without their consent.

The FRANCE 24 Observers team could not independently verify this claim but King noted that such a possibility would not come as a surprise.

It would be technically possible and it is highly likely to happen in the future, it’s going to keep happening in different guises, maybe it’s the phone deleting videos, or maybe it’s something that we have not thought of yet. The idea of controlling information is not a new thing and it’s not going to go away.