How fake automated Chinese social media accounts are being used as a Trojan horse for dissent – Bai Mingcong
On October 21, 2018, an account named ‘People’s Daily bot’ (@人日bot) posted this message on Weibo:
They fear the empowerment of the people, fear that the people shall see the true face of our era, and further yet, they fear that their vice shall be exposed in front of the masses! (他们害怕人民翻身，害怕人民认识大时代的真面貌, 更害怕他们自己的丑恶暴露在人民大众面前!)
Taken at face value, the account appears to directly and forcefully target the Chinese regime. Puzzlingly, by the time of publication, the post has yet to be removed, and the account has not been banned, as usually happens to dissenting social media in China. Yet a closer look reveals that this is a repost of a 1946 editorial from the People’s Daily, the central mouthpiece of the Chinese Communist Party, that criticized the treatment of journalists in the Nationalist “occupation zone” as contrasted with the communist “liberated zone” under CCP control. The survival of the post in the face of hardening censorship is not a loosening of the cords. Instead, it is representative of a new trend on the Chinese internet, in which Weibo accounts purporting to be bots hide their criticism of the government behind prominent and often politically unassailable figures of modern China.
Since its inception in August 2009 as the first fully open social media platform in China (Facebook clones such as Renren preceded it, with closed access between contacts), Weibo – an equivalent to Twitter – has become home to over 400 million active users, with hundreds of thousands of posts uploaded daily. Of these countless interactions between millions of netizens, there is no shortage of political discussion and debates, even on subjects highly critical of the Communist Party. There were even officially-sanctioned “democracy sessions” on Weibo, especially during the annual central government meetings of spring 2011, where the people’s representatives went online to discuss policies with Weibo users. In the summer of the same year, the high-speed train collision in Wenzhou drove civil discourse on Weibo into high gear, with news from the collision site, calls to support rescue and reparation, and anger towards the government for covering up casualty statistics. By then, Weibo had become a platform for civic discourse and open criticism. In contrast, after a continuing effort to filter, censor and monitor its users since Xi Jinping’s rise to power in 2012, political discourse on Weibo has fallen precipitously. Politically active netizens were either driven to foreign social media platforms or were forced to use more creative means to express their views.
Before 2012, Weibo had become a platform for civic discourse and open criticism”
Bots – that is, non-human accounts which post automatically – have a significant presence on Weibo, as they do on Twitter and Facebook. Often, they repost inspirational quotations from books and magazines, among other content. Since Weibo provides its users and developers with a powerful open Application Program Interface (API), it is possible for any technically qualified individual to create third party applications, perform data analyses, or create bot accounts using the API client. As I discovered, however, the “bot” who posted the critical quote above is not a bot at all. It is in fact a real netizen, who guarded themselves from censorship and government harassment by hiding behind the façade of a lifeless bot account.
A detailed analysis of the account exposes this reality. Behind the account is not an automated bot application operating on the Weibo open API, but a human user. We know this because Weibo stamps every post with information of the web “client” on which the posts were made. By default, the stamps should read “微博 weibo.com.” In the case of @人日bot, however, it is clear that there was at least some degree of human intervention in the posting process, as this web client stamp changed frequently. While most of its posts were (supposedly) sent via weibo.com, one post on October 21 was sent from “360安全浏览器” (360 Secure Browser, a security-focused browser developed by Qihoo); on the same day and the day before, however, several posts were made from “iPhone X.” Therefore, even if the account did indeed partially use a bot function to automatically report quotes, there was a human operator behind it.
By taking the 1946 quote out of context, this “bot” was highlighting the irony of poor press freedom under the same Communist Party that once decried censorship under the Nationalists. The same account, which has nearly 8000 followers, has also posted other ironic quotes from People’s Daily, such one from a 1988 editorial which criticizes both the public’s reluctance to speak “frank words” and the tendency of “some leaders” to retaliate against those who dare to speak out. Similarly, it reposted a 1948 news report on the banning of “liberated zone” radio stations by the Nationalist government in Nanjing. The comments section of the account, while relatively quiet, suggests that its readers are able to see the irony.
@人日bot, however, is not the first, or only, of its kind. One of the earliest prominent Weibo “bots” is ‘Lu Xun bot’ (@鲁迅bot), which reposts quotation from Lu Xun, a highly regarded writer and political activist of the Republican era with whom every Chinese schoolchild is familiar. The People’s Education Press’s literature textbook for middle and high school students once defined him as a “fighter for democracy” who spoke the truth and daringly criticized the corrupt Chinese society and government of his time. In 2013, the newest edition of the same textbook removed some of Lu Xun’s works, sparking online debate about tightening censorship of education and literature. Yet Lu Xun’s work is far from banned in China, and @鲁迅bot takes advantage of the ambiguity to highlight quotations from his work that offer veiled criticisms of the current Chinese government. Take this seemingly innocent quote from Lu Xun’s novel Forging the Swords (铸剑):
And now, the whole city is fussed up over the King’s tours, the ceremonies, and his majesty, how they were honored to have seen the King in person, how low they’d bowed, and how they’d presented themselves as a proper example of all citizens, like bees dancing and buzzing in a courtyard. (这时满城都议论着国王的游山，仪仗，威严，自己见得国王的荣耀，以及俯伏得有怎么低，应该采作国民的模范等等，很像蜜蜂的排衙.)
Within the scope of the novel, this refers to nothing more than a nation whose people sought personal gain by flattering the king. Yet within a contemporary context, as the comments section clearly shows, @鲁迅bot is ridiculing official Chinese media who fawn over Xi Jinping and other high officials. These news reports use the same set of phrases to describe official actions. For example, Xi’s speeches are always referred to as “important speeches” (重要讲话), while his interactions with party officials and students are characterised as “warm” (亲切). As well as using Lu Xun to target contemporary Chinese media coverage of the CCP, the quotation can also be interpreted as critiquing the increasingly prominent cult of personality around Xi Jinping.
By directly quoting from prominent, politically unassailable figures, the accounts became shields against censorship”
The @鲁迅bot account was created on September 9, 2018, and stopped posting just over a month later, leaving behind at least 280 posts (some of which are now missing, including the announcement to cease operation on October 20, screenshots of which can be found online). Since then, it has been renamed ‘Ruins of Hotel Xianheng’ (@咸亨酒店遗址), the fictional hotel – supposedly located in Lu Xun’s home town of Shaoxing – mentioned in many of his works. The death of the account sparked open discussion, on more formal open forums such as NetEase and Zhihu, of how it tested the boundaries of Weibo. By the time of writing, @咸亨酒店遗址 has nearly 220,000 followers, making it the most popular among this kind of fake bot account, and likely the inspiration for many other similar accounts. During @鲁迅bot’s month of activity, other “bots” such as the ‘Deng Xiaoping Selected Works bot’ (@邓选bot), ‘Mao Zedong Selected Works bot’ (@毛选bot), Xinhua Daily bot (@新华曰报)”, and various reincarnations of the original Lu Xun bot, were registered and began posting in a similar manner. By directly quoting from prominent, politically unassailable figures, the accounts became shields against censorship, and even against arrest or questioning of the “bot” owners – keeping themselves relatively safe despite their criticism of the government and Chinese society.
Despite this dissent, its impact remains small in a heavily censored and often dangerous political environment. The almost 220,000 followers of the new Lu Xun bot is relatively small, considering that Weibo is home to nearly 2000 times more active users. And many of its reposts and replies to original posts have been filtered, hidden or removed. Its first incarnation, which ceased on October 20, came under heavy criticism for its use of “out-of-context quotations.” Some asserted that these accounts were not voicing legitimate criticism, but were just venting their anger – dismissing their activities as acts of the “angry youth” (愤青), a term used to refer to radical and outspoken young people. Subtle censorship has prevented popular “bot” posts from hitting the “most read” trends on Weibo, making it unlikely that these accounts will see further growth in public awareness in the foreseeable future.
These dissenting “bots” are left with a small and silenced audience, and a looming threat against their existence and personal security. To protect themselves, they fall back on their status as supposedly non-human bots. Yet their cover falls apart on closer examination. More vulnerabilities in their tactics could be found upon further inspection: the inconsistent web client stamps of @人日bot, the @毛选bot posting a screenshot from the iOS Notes application. Meanwhile, in multiple instances the @鲁迅bot posted screenshots of hand-written text in an attempt to dodge Weibo’s digital text-filtering algorithm. The efforts of these individuals to obscure their identities is, in the end, moot: effective September 15, 2017, all Weibo users are required to supply their cell phone numbers, which are in turn linked to their government-issued ID.
This ingenious way to avoid Weibo censorship is, therefore, all too imperfect. Not only are the messages muffled by government censors and questioned by Chinese netizens, but the personal safety of the posters is under threat. Weibo has given birth to yet another flawed Chinese art form, one that meshes social ingenuity and technological knowledge – obscuration with Chinese characteristics. Practitioners of this art distinguish themselves from the “open” activists on Weibo, the Big Vs (officially verified cyberstars and organizations) such as Xue Manzi, who openly spoke out against the regime, and found himself swamped by government prosecutions. Instead, they use technical know-how to disguise themselves from being identified as dissidents. The “bot” accounts treasure the free flow of information online. Yet given the circumstances, they are forced into a delicate dance with censorship, and a struggle to defend truth on a fragile and volatile internet. ∎