censorship

State of Play: Internet Censorship in China

The Great Firewall is unique in its scope and sophistication, but the true strength of China’s censorship system lies in its government, not its technological prowess.

Chengdu downtown in China
Amy Hawkins
By Amy Hawkins

Last summer, Hua Yong’s sister tried to post him a picture frame. The painter was under house arrest, holed up with a friend in the mountainous region of Shangri-La. He was allowed to pass the time by painting, but needed frames to keep his work in good condition. Hua Yong repeatedly tried sending his address on WeChat – China’s ubiquitous messaging app – but it was never received. Only when he removed his name did the message deliver.

Hua Yong couldn’t contact his sister because his name was blocked on WeChat, meaning messages that referenced him would not be delivered, though it would look to the sender as if they had. He was under house arrest for protesting against the government, and authorities had taken extra precautionary measures to ensure no one could easily talk about his case – or post him picture frames. While Hua Yong’s life as an artist and activist is unusual, his experience with censorship is far from unique. China is a country where privacy, if it exists at all, can only be experienced offline. Online, not only can the government see all, it goes out of its way to control what others can see.

China’s internet is largely a closed space. Content from the outside world is blocked and filtered by the Great Firewall; content within China’s borders is subject to censorship, and users to perpetual surveillance. “Without web security there’s no national security, there’s no economic and social stability,” President Xi Jinping said last year. “Web security” in China means limited access for foreign companies and a highly curated domestic internet, giving rise to what James Griffiths, author of The Great Firewall of China, calls “an alternative version of the internet”. In the past decade, this alternative version has been defined by the increasing power that the Chinese Communist Party (CCP) exerts over domestic and foreign companies, as well as over its own citizens.

The walled garden: Keeping foreign companies out

For outsiders and visitors to China, the most obvious manifestation of censorship is the blocking of ubiquitous Western websites such as Google, Facebook, Twitter, and any related products such as Google Maps, Instagram or WhatsApp. This is inconvenient, and for those doing international business with China, a serious barrier to productivity. The only way to access these websites is via a virtual private network (VPN), which encrypts your connection and directs your request through a proxy server, delivering you to the censored website.

While most VPNs are forbidden and enforcement of their ban is escalating, the CCP is aware that most businesses would be crippled without communication with overseas partners. As a result, authorities have tried to establish business licences allowing companies to use state-sanctioned VPNs for pre-approved means.

The most common internet experience in China is found within the ‘walled garden’ that the CCP has carefully cultivated in the decades since the internet’s arrival in the country. Nearly every Western internet company has a Chinese equivalent: internet searches are made on Baidu, status updates are shared on Weibo, and messages are sent on WeChat. The latter platform, owned by internet giant Tencent, has more than one billion active users, thanks in part to its multifaceted service offering. As well as messaging, the mobile payment function is ubiquitous; users can order taxis, play video games, rent bicycles and much more, all within one platform. “The sophistication of the domestic internet means that the trade-off one makes to stay within the firewall is less onerous than ten years ago,” says Griffiths.

The indispensability of domestic internet services to the Chinese netizen was no accident, nor was it an inevitability. As Griffiths notes in his book, within two years of launching in China in 2002, Google controlled 25% of the market, “a level of success never achieved by other foreign search engines”. Despite agreeing to the Chinese government’s demands to censor search results, albeit with a disclaimer to alert the user to the manipulation, Google was forced out of China in 2010. Robin Li, CEO of Baidu, celebrated this as a ‘win’ for his company, but as anyone who has used Baidu can tell you, it was not Baidu’s higher quality offering that forced Google into retreat. Rather, faced with cyberattacks, mounting pressure in the US and increasing censorship demands in China, Google was forced to withdraw for the next eight years.

Indeed, inside Baidu, Google’s departure was seen as a negative development. Despite the government’s claims that the ‘walled garden’ would promote domestic companies, Kaiser Kuo, former director of international communications at Baidu, Google’s retreat “was one of the worst things to happen to [Baidu]”. There was “a very real awareness that losing a major competitor was a big problem, and all these things that [Baidu] worried about did come true: no fire in the belly and it was too easy to hit targets”. Kuo says that the result is that the Baidu brand is “in crisis”. “Google was by far the most successful internet company in China…[it] created a brand in peoples’ minds that people equated with international connectivity”.

Chinese man buys drink with WeChat

WeChat’s mobile payment function is almost ubiquitous in China. Image: nagarjun from Flickr

Pruning the walled garden: Inside China’s internet ecosystem

 Competitive concerns aside, the ubiquity of Chinese platforms means that censorship and surveillance is unavoidable. Companies in China have always been politically obliged to share user data with the government as requested, and this was formalised in 2017 with the Cybersecurity Law. This law also made platforms directly responsible for any content they host, making them even more proactive in scrubbing their services of any sentiments that could be perceived as anti-government. The law was part of the increasing authoritarianism that China has experienced since Xi Jinping became president in 2013. The tools for censorship have certainly become more sophisticated in recent years – detecting images as well as text, for example. But the defining change of the past decade has been the increasing power that the government exerts over domestic and foreign companies alike.

Weibo, for example, used to be a forum where critical views were often shared. This was particularly apparent in 2011, after a fatal high-speed rail crash in Wenzhou killed 40 people. Netizens flocked to Weibo to voice their outrage at the way the government handled the crash, arguing that the authorities were concealing the true damage of the tragedy. This was the “golden age of Weibo,” says Kuo, when it was “hard to go a week without some glaring instance of malfeasance being exposed by zealous Weibo watchers”.

The idea of any similar conversation happening on the Chinese internet today is unthinkable. “When censorship first started, it was limited to a narrow list of subjects – Tiananmen, Tibet, sex,” says Charlie Smith, founder of GreatFire.org, an anti-censorship project. “Now it’s just so broad,” he adds, citing examples as wide-ranging as bans on celebrity gossip to discussions of stock prices. Recently, topics that have been targeted either directly by the government or by overzealous company censors include vulgar memes on the now-shuttered app Neihan Duanzi, sex education videos, and Winnie the Pooh. Each of these examples is instructive in the differing kinds of content that the government wants to keep out.

“the defining change of the past decade has been the increasing power that the government exerts over domestic and foreign companies”

For Neihan Duanzi fans, the problem wasn’t that their memes were often based on dirty jokes, although the censors are often puritanical in their opprobrium. Rather, it was that the app was becoming so popular that fans were meeting up in real life to discuss their favourite jokes. Neihan Duanzi users even had secret signs, such as flashing car headlights in a certain sequence, to greet other meme-lovers. This emergence of an independent social identity is exactly what the CCP rejects in all its forms – fearful of competing loyalties and collective instincts, real life gatherings of people connected by a common interest are closely monitored and rarely tolerated.

Perhaps surprisingly, hate speech remains rampant on the Chinese internet, consistently evading censorship or reprobation. In 2013 Lu Wei, the since-disgraced tzar of the Chinese internet and head of the Cyberspace Administration of China (CAC), declared that “in cyberspace, people with different skin colors, nationalities, cultures and languages should be equally entitled to participation, free speech and development”. “If everyone were to spread positive energy on the internet, the world would be a much better place,” he continued.

The reality is that while political dissent is verboten, discrimination and abuse are at least as common as they are in the West. After the Christchurch shootings in March 2019, Chinese social media was awash with pro-shooter comments which were not deemed by censors to counteract the philosophy of “positive energy”. Tony Lin, a media researcher at Hong Kong University who studies hate speech on Chinese social media found one article on WeChat that described the attack as “heroic revenge”. The article quickly surpassed 100,000 views (at which point the view count stops being public) and hosted a poll in which 76% of respondents expressed sympathy with the shooter. Elsewhere, after writing on Weibo about her experience of discrimination in China, the transgender advocate Chao Xiaomi was told that people like her should be exterminated “just like in the Second World War, when they killed the Jews”. “They were the most terrible words I have ever heard from people,” she remembers.

Lin argues that the structure of Chinese social media promotes sensationalist views. Weibo is often compared to Twitter, but unlike the American platform, posts with more likes are more visible in a user’s newsfeed, which Lin says “promotes controversy”. Furthermore, while censorship is “completely internalised for some people,” hate speech is seen as less sensitive; no public figures have faced professional repercussions for spouting vitriol against minorities, while they have for expressing support for Hong Kong or Taiwanese independence.

Despite the wide berth given to online abuse, China is in fact highly socially conservative. This is especially the case under the CCP of recent years, which has re-embraced the traditional values of Confucianism that the Communists originally rejected. When it comes to censorship, this means that all broadcast content must be appropriate for a 12-year-old audience, with sex scenes edited out of foreign films and clothing photoshopped onto naked actors. These standards extend to internet censorship. The video channel Wei Zai Bu Dong Ai has over 5 million followers of its punchy sex education videos. Every week, the host Chang Mengran puts out entertaining, instructive videos on topics such as menstruation, consent and body image. Increasingly, however, these videos have been taken down for being “vulgar” or “low-level”, according to censors.

Even Winnie the Pooh has come under censure. The honey-loving bear has rarely been accused of vulgarity, unless you’re looking in some particularly dark parts of the internet. But he was targeted by censors in 2017, when netizens started to make joking comparisons between the cuddly cartoon character and Xi.

How Chinese netizens navigate censorship

Whether it’s for abuse, activism, or even just memes, netizens who want to speak out have to find a way to beat the censors. Studies on how the general public use VPNs are limited, but those that exist suggest that the figure is low – less than 3%. “Any hurdle in people’s way has a real effect,” says Griffiths. “Removing VPNs from the App Store and blocking VPN websites – that’s pretty effective. It’s never going to be 100% effective, but it doesn’t need to be”.

In an experiment published last year, researchers from Harvard and Peking universities found that merely providing people with the tools to circumvent censorship did not result in more acquisition of sensitive or censored information. It was only with active encouragement that some subjects used the tools available to them to access foreign media and other banned websites. The researchers concluded: “censorship in China is effective not only because the regime makes it difficult to access sensitive information, but also because it fosters an environment in which citizens do not demand such information in the first place”. It is worth noting that although some foreign media such as the New York Times and the Financial Times publish Chinese translations of their articles, these are not mass market; the vast majority of Chinese language websites are found within the Great Firewall.

For the minority of people who do choose to use VPNs, life is getting harder. While they have always been taboo, for years the CCP turned a blind eye to the relatively small number of people who used them. VPNs have been described as a pressure valve for elite public opinion – without being able to blow off some steam and connect with peers overseas, the frustration would bubble over. In 2017, Chinese regulators announced that VPNs would have to be licensed to operate in China. Since then, Apple has removed all VPN apps from its Chinese iOS App Store, and Google now refuses to sell third-party advertising for VPN services. There are ways round this: iPhone users can change their App Store to an overseas location, but this is only possible with some technical knowhow and a foreign credit card. There have been multiple cases of people being arrested or fined for using or selling VPNs. “The avenues for getting past the firewall are smaller,” says Griffiths, “for an ordinary person who doesn’t leave the country, your ability to access a VPN has shrunk dramatically”.

So what are people within the firewall to do? For slogans, one way of getting around banned words is to use homonyms – a particularly ripe area for wordplay in Chinese, because each sound can have four different meanings, depending on its tone. The most famous example of this used to be 肏你妈 (cào nǐ mā), which means “f*** your mother”. With online arguments debasing as they often do, censors decided to block this insult. The result? 草泥马 (cǎo ní mǎ), which means “grass mud horse” became the new way to offend an online adversary.

“VPNs have been described as a pressure valve for elite public opinion”

More recently, followers of the #metoo movement were forced to make similar creative leaps. The hashtag for people sharing their experiences of sexual assault and harassment was swiftly blocked after the movement came to China in early 2018. As the rest of the world saw, the outpouring of emotion that #metoo heralded amounted to a searing criticism of the social order. This kind of collective identity and cohesion made it a prime target for the censors. WeChatscope, a research initiative at Hong Kong University that monitors censorship on WeChat, found that #metoo was one of the most heavily censored topics of 2018.

While #metoo was a prime topic for censorship, most cases are not so clear. According to Smith, “company [executives] are well aware that if anything ‘anti-government’ goes viral on their platform, they will be severely punished”. As a result, internet platforms generally have to second-guess the government’s priorities. This can sometimes lead to overcompensation, which Chinese netizens respond to by flexing their consumer power. In April 2019, for example, the LGBT hashtag #les was blocked on Weibo, prompting outcry from activists and allies, who then flooded the platform with rainbow symbols and photographs of women with their mouths taped shut. Weibo partially reinstated the hashtag soon after.

Still, for highly sensitive topics such as #metoo, familiar workarounds have to be found. Campaigners started using the hashag #mitu or #米兔 (mǐtù), which means “rice bunny”, to talk about sexual harassment. Other tactics included sharing screenshots – sometimes distorted or inverted, to further impede detection – rather than text messages, to slip through the censors’ web, or sharing old articles that had a symbolic relevance for the present moment. In one incident, supporters of a detained activist even stamped her censored message on an Ethereum transaction, where it remains, inscribed permanently on the blockchain.

As hiding places go, deep within a cryptocurrency is an effective way of making sure few people ever see your message. But the Ethereum message has a symbolic value that goes far beyond that particular activist. The optimism of blockchain enthusiasts regarding the world-changing, liberating potential of the technology sounds much like the tech idealists who saw the internet as the ultimate democratising force in its early years. Controlling the internet in China would be like trying to “nail jello to the wall,” joked Bill Clinton in 2000. But that is exactly what China did – with the help of companies from Clinton’s own country.

Exporting censorship: The Great Firewall around the world

The internet was an American creation and as such, so were the first internet surveillance tools. Companies like Cisco and Spring developed technology that allowed companies to monitor what websites employees used while at work, and block certain material. This principle was rolled out on a national level once China started taking advantage of that same technology; by 2001, China was buying $20-billion of telecoms equipment each year, mostly from American companies, Griffiths writes. Historians Tim Wu and Jack Goldsmith have said that the Great Firewall was built “with American bricks”. Since then, China’s domestic capabilities have skyrocketed, and the country is now the world leader in supplying other countries with the tools to monitor their own people.

Last year, news emerged that Venezuela was rolling out a digital ID card program developed by Chinese technology company ZTE. The program would collect data on Venezuelan citizens to aid the autocratic government in controlling sedition. A few months earlier, the Zimbabwean government signed a deal with CloudWalk, a Chinese start-up, to buy Chinese facial recognition technology. The deal would allow CloudWalk to train its artificial intelligence on darker skin tones; since then the company has boasted that its facial recognition has become adept at recognising Chinese Uyghurs, a persecuted minority in China who have darker skin than the majority of Han Chinese people.

One way in which the Uyghurs are targeted is by the transformation of their homeland into a digital prison. The Uyghurs are a Muslim minority living in historically contested territory in the north-west of China (known by the CCP as Xinjiang and by independence activists as East Turkestan). In the past three years, the CCP has detained over one million Uyghurs and Kazakhs in prison camps across the region under the guise of anti-terrorism measures. For those who are not detained, daily life has become a testing ground for the most extreme facets of China’s digital dictatorship. In just one example, residents are forced to install “web-cleaning” software onto their mobile phones, which is in fact spyware. Since 2016, the video surveillance companies Dahua and Hikvision have won over $1 billion in government contracts in Xinjiang.

Censorship itself is becoming big business. While the government does maintain some aspects of monitoring and censorship itself, the scale required and the expectation of media companies to take independent responsibility for their content means that the demand for outsourced censorship services is booming. Whether it’s by employing masses of human censors or developing artificial intelligence that can track keywords across vast swathes of content, start-ups and established media institutions are cashing in. The state-run newspaper People’s Daily reported a 166% rise in revenue from “third-party censorship services” in 2018, while Baidu and the gaming company NetEase are reportedly offering similar services across text, video, and images. Kuo says that historically, Baidu pushed back on government censorship requests, knowing that “consumer-facing companies do not labour under the illusion that people prefer censored material”. Baidu would “try to grow the whitelist” and “shrink the blacklist”. Whether or not that remains the case, the internet landscape in 2019 means that censorship can be a lucrative way for companies to diversify their income.

Not all censorship has to be paid for. In recent months there has been an increase of patriotic netizens brigading pages that post what they perceive to be anti-China content. Talk to East Turkestan and the World Uyghur Congress, two groups that promote awareness of Uyghur issues, were recently targeted by this phenomenon. Chinese netizens flooded the groups’ Facebook pages with thousands of pro-China comments such as “We are Chinese! We reject Xinjiang’s independence, and we will never stop!” CNN reported. The attack came from members of a Reddit-like forum whose moderators encourage to be as patriotic as possible, lest the forum itself attract scrutiny from the censors.

“censorship can be a lucrative way for companies to diversify their income.”

The targeting of social media beyond the Great Firewall is a relatively recent development in China’s campaign to control online discussion. “It’s connected to the broader political trends of controlling how China is perceived overseas,” says Griffiths. The Chinese police have detained people for liking tweets critical of the CCP, and Chinese Twitter users have been forced to delete their accounts or content, or handover their login information to police for the same result. In March, the anonymous @AirMovingDevice, an account which posted data analysis about China, tweeted: “I will be deleting all of my tweets and will no longer be tweeting or responding to DMs. All of my tweets were entirely based on my personal analysis using publicly available data, and did not involve other individuals. It is not my intention to subvert state or Party authority” – a tweet that was itself later deleted.

Xinjiang Province in China

Xinjiang is a historically contested region home to China’s Muslim minority. The area has become a testing ground for extreme surveillance and censorship. Image: Anagoria, Wikimedia Commons.

Silicon Valley’s support of the Great Firewall

Just as the Chinese state is reaching into foreign social media, foreign tech companies are becoming increasingly compliant with the Chinese government’s demands. In February last year, Apple migrated Chinese users’ iCloud data to servers in mainland China, and began hosting the encryption keys for this data in China rather than the US. The company also updated its Chinese terms of service to state that both Apple and its Chinese partner company would have access to user data. In August, the news leaked that Google was developing plans to re-launch in China with a censored search engine codenamed Dragonfly. The project was ostensibly dropped following internal and public outcry, but there have been rumours since that the company remains determined to gain access to China’s 800 million internet users.

For the most part, Chinese people are beholden to domestic technology companies who not only provide the government with access to user data, but often leave it open for anyone with the technical knowhow to access. Earlier this year, the Dutch hacker Victor Gevers discovered a database of over 2.5 million people living in Xinjiang, detailing their names, ID card numbers and locations tracking data. It had been left open online by the Shenzhen-based facial recognition company SenseNets.

Speaking to Top10VPN, Gevers says that Chinese companies are particularly bad at exposing personal data: as of March, there were 29,808 open databases in China – “that number should be zero”. The problem is particularly acute because many of the Chinese databases relate to security and surveillance. When these databases are left open, anyone can interfere with it, meaning “the value of that data is zero,” says Gevers.

Considering how punitive the government’s response can be to digital missteps, it’s not hard to imagine malicious manipulations of vulnerable databases. The biggest database that Gevers found, for example, was linked to WeChat. It stored one billion private messages per day, which were flagged due to “sensitive” key words such as “Winnie the Pooh”. The messages were linked to ID card numbers and geo-locations.

The netizens fight back

One database that can be used against censors is GitHub. Considered essential for developers around the world, the open-source software database is widely used in China’s tech industry. Although the company is owned by Microsoft, who generally comply with CCP censorship requests, it has still hosted tools that allow people to circumvent China’s digital diktats. Most commonly this is by providing the code for VPNs or other anti-censorship software, but there are also projects such as “f***-xueixiqiangguo”, which contains software to cheat the government’s propaganda app. Elsewhere, tech workers have used it to voice protest against the “996” culture that is common in their industry – 9am to 9pm, six days a week. “China is not going to censor GitHub for one simple reason: they need it,” believes Gevers, who sees democratic potential in the database.

In reality, it is unlikely that white collar tech workers will be the downfall of the CCP. Given the close relationship between internet platforms and authorities, more tech workers are working in support of China’s surveillance state than against it, whatever their personal misgivings may be.

The Great Firewall is unique in its scope and sophistication, but the true strength of China’s censorship system lies in its government, not its technological prowess. Restrictions on VPNs at politically sensitive times such as the annual National People’s Congress meeting demonstrate that even the tools of liberation are under the thumb of the CCP. With a leader bent on consolidating more power than anyone since Mao Zedong, these restrictions are only likely to tighten. But despite the state’s Orwellian inclinations, Chinese netizens are not without a voice. As #metoo’s rice bunnies and the #les rainbow backlash have shown, communities online have a way of talking back.