HONG KONG (AP) — Videos of hundreds of people protesting in Shanghai began appearing on WeChat on Saturday night. Showing chants about the lifting of COVID-19 restrictions and calling for freedom, they would only stay awake minutes before being censored.
Elliot Wang, a 26-year-old from Beijing, was amazed.
“I started constantly updating, saving videos and taking screenshots of what I could before it got censored,” said Wang, who agreed to be quoted using only his English name, out of fear of government retaliation. “Many of my friends shared videos of the protests in Shanghai. I shared them too but they would be removed quickly.
That Wang was able to glimpse the extraordinary outpouring of grievances highlights the cat-and-mouse game playing out between millions of Chinese Internet users and the country’s giant censorship machine.
Chinese authorities maintain a tight grip on the country’s internet through a complex multi-layered censorship operation that blocks access to nearly all foreign news and social media and blocks topics and keywords considered politically sensitive or harmful to the Party government Chinese Communist. Protest videos or calls are usually deleted immediately.
But images of the protests began spreading on WeChat, a ubiquitous Chinese social networking platform used by more than 1 billion people, in the wake of a deadly fire in the western city of Urumqi on Friday. Many suspected that the lockdown measures prevented residents from escaping the flames, which the government denies.
The sheer number of disgruntled Chinese users who have taken to the Chinese internet to express their frustration, coupled with the methods used to circumvent censorship, has led to a short period of time where government censors were overwhelmed, according to Han Rongbin, an associate professor in the University of Georgia’s Department of Public and International Affairs.
“Censors take a while to study what’s happening and add it to their portfolio in terms of censorship, so it’s a learning process for the government on how to conduct censorship effectively,” said Han.
In 2020, the COVID-19 death of Li Wenliang, a doctor arrested for spreading rumors following an attempt to alert others about a “SARS-like” virus, sparked widespread outrage and an outburst of anger against the health system. Chinese censorship. Users posted critiques for hours before censors moved to delete the posts.
While censors deleted posts related to the fire, Chinese Internet users often used humor and metaphor to spread critical messages.
“Chinese netizens have always been very creative because every idea used successfully once will be discovered by censors next time,” said Liu Lipeng, a censor turned critic of China’s censorship practices.
Chinese users began posting images of blank sheets of white paper, Liu said, in a silent reminder of the words they weren’t allowed to post.
Others posted sarcastic messages like “Good good good sure sure right right right right yes yes yes” or used Chinese homonyms to evoke calls for President Xi Jinping to resign, such as “moss shrimp,” which sounds like the words for “resign ” as well as “banana peel”, which has the same initials as Chinese President Xi Jinping.
But within days, censors moved to hold white paper images. They would use a range of tools, said Chauncey Jung, a policy analyst who previously worked for several Beijing-based Chinese internet companies.
Most content censorship isn’t done by the state, Jung said, but entrusted to content moderation operations on private social media platforms, which use a mix of humans and artificial intelligence. Some censored posts aren’t deleted, but can only be made visible to the author or removed from search results. In some cases, posts with sensitive key phrases may be published after review.
A Weibo search Thursday for the term “white paper” yielded mostly posts that were critical of the protests, with no images of a single blank sheet of paper or of people holding carte blanche during the protests.
The global internet can be accessed from China using technologies such as virtual private networks that mask internet traffic, but these systems are illegal and many Chinese internet users only access their home internet. Wang doesn’t use a VPN.
“I think I can say for all the mainlanders of my generation that we are really excited,” Wang said. “But we are also very disappointed because we can’t do anything… They keep censoring, deleting and even releasing fake accounts to praise the policemen.”
But the system works well enough to prevent many users from seeing them. When protests erupted across China over the weekend, Carmen Ou, who lives in Beijing, initially didn’t notice.
The protests only became known later, after using a VPN service to access Instagram.
“I tried to look at my WeChat feed, but there was no mention of any protest,” he said. “If it weren’t for a VPN and access to Instagram, I might not have discovered such a monumental event had taken place.”
Han, the international affairs professor, said censorship ‘doesn’t have to be perfect to be effective’
“Censorship could work to prevent a large enough population from accessing critical information to mobilize,” he said.
China’s opaque approach to quelling the spread of online dissent also makes it difficult to distinguish government campaigns from ordinary spam.
Searching Twitter using Chinese words for Shanghai or other Chinese cities reveals protest videos, but also an almost constant stream of new posts featuring racy photos of young women. Some researchers have proposed that a state-backed campaign could seek to stifle news of the protests with “unsafe for work” content.
A preliminary analysis by the Stanford Internet Observatory found plenty of spam but no “compelling evidence” that it was specifically intended to suppress information or dissent, said Stanford data architect David Thiel.
“I would be skeptical of anyone claiming clear proof of government attribution,” Thiel said in an email.
Searching Twitter for more specific terms related to the protests, such as “Urumqi Middle Road, Shanghai,” mostly yielded posts related to the protests.
Israeli data analytics firm Cyabra and another research team that shared the analysis with the AP said it was difficult to distinguish between a deliberate attempt to stifle sought-after protest information from the Chinese diaspora and a routine spam campaign. commercial.
Twitter did not respond to a request for comment. It hasn’t responded to media requests since billionaire Elon Musk took over the platform in late October and has cut much of its workforce, including many of those tasked with moderating spam and other content. Musk frequently tweets about how he is implementing or enforcing new rules on Twitter content, but he hasn’t commented on the recent protests in China.
AP Business writer Kelvin Chan in London and AP Technology writer Matt O’Brien in Providence, Rhode Island contributed to this story.