How China's Firewall Grew Legs
A weird metaphor, but stay with me
We discussed last week the many ways in which China controls its internet, utilizing the Great Firewall, a network of government workers, and other creative, persistent tactics to monitor and curate the web within its borders.
But this control doesn’t just stay in China, and we’ll see how far it extends today. As with last week, I’ll lean extensively on James Griffiths’ book, The Great Firewall of China: How to Build and Control an Alternative Version of the Internet. It’s extensively researched, and borrows heavily from the author’s own journalism covering technology, politics and protest movements in China, Hong Kong, and other parts of the region for over a decade.
I initially began learning about and planning these posts months ago, long before the US House passed the controversial “TikTok Bill” aiming to force ByteDance, the app’s parent company, to sell the app to US interests. And the bill, controversial as it is, lends new urgency to this question of Chinese internet control.
Let’s conduct a thought experiment. What if the country in question was not China, but Russia — another country which sponsors coordinated disinformation campaigns, and which tends to “detain, silence and sideline” the leaders of private companies it disagrees with.
And what if the technology in question was not TikTok, but Facebook?
I hadn’t planned to weigh in directly on the TikTok Bill, but it’s hard not to as I write these posts. I’ve heard compelling arguments — including from trusted Chinese-American friends who have direct experience with the Great Firewall — that push back on the proposed bill. Primary among these is that the US government shouldn’t “stoop low” to combat Chinese government control, and that we should avoid giving the CCP any more reason to stoke anti-American sentiment. And importantly, any more excuse to further cut off China’s internet from the rest of the world.
I think this is very worth considering. And to be honest, I’m most swayed by Max Read's argument that TikTok’s ownership likely doesn’t really matter:
“More than a decade of reporting on Facebook, Twitter, and YouTube, and the unintended consequences and ripple effects of small and well-intended design and incentive tweaks, should have taught us all that platforms are barely able to be monitored, let alone controlled, by the multi-billion dollar companies who own the platforms and employ thousands of people specifically for the purpose of monitoring and controlling them. The idea that direct access to TikTok’s video-ranking and recommendation algorithms would give you similarly direct access to the brains of American voters is patently absurd.”
But alongside all this, we should also consider that the Chinese government is really, *really* good at the whole “internet censorship and control” thing — they’ve spent decades honing it, after all. As we’ll see, the government has doggedly and creatively coerced, controlled and surveilled through not only internet hacks and social media disinformation campaigns, but also agency newswires and even Hollywood.
Just this week, the New York Times published an extensive look at the many ways the Chinese government already seeks to influence this year’s US election, utilizing a corral of fake accounts posing as devoted Donald Trump fans, disparaging Biden and sowing doubt in the value of US democracy itself.
And these concerns aren’t new. Ahead of the 2022 midterm elections, Mandiant published research on “Dragonbridge,” a PRC-backed influence campaign “aggressively targeting the United States by seeking to sow division both between the U.S. and its allies and within the U.S. political system itself.” The report highlights tactics for achieving this including “nuanced impersonation of cyber actors,” “plagiarism and alteration of news articles,” and “personas posing as members of target audience.”
In 2023, researchers uncovered the promotion of AI-generated deepfake videos by Chinese bot accounts on Facebook and Twitter — videos which “appeared intended to promote the interests of the Chinese Communist Party and undercut the United States for English-speaking viewers.”
Researchers at Graphika, which studies online disinformation, even coined these efforts “Spamouflage.” Canada experienced similar disinformation campaigns that same year, and Meta conducted its “single biggest takedown” of accounts tied to a pro-China influence campaign.
It’s worth quoting the Meta takedown article at length to illustrate its scope:
“On Feb. 27, an article claiming that the United States was behind the bombing of the Nord Stream underwater pipelines in the Baltic Sea was published on the Substack and Blogspot blogging platforms.
Within 24 hours, the article — and other versions of it — had been posted to more websites, including Reddit, Medium, Tumblr, Facebook and YouTube. Translations of the article in Greek, German, Russian, Italian and Turkish also began appearing online…..
In total, 7,704 Facebook accounts, 954 Facebook pages, 15 Facebook groups and 15 Instagram accounts tied to the Chinese campaign were removed by Meta, which owns Facebook, Instagram and WhatsApp. Hundreds of other accounts on TikTok, X, LiveJournal and Blogspot also participated in the campaign….”
That takedown of accounts was the seventh involving China over the course of six years, and the trend isn’t slowing. Indeed, researchers wrote in that Meta “takedown” that many of the online bot accounts were riddled with grammatical errors and sometimes slipped into Mandarin, despite their English-speaking personas. But the effort has become much more disciplined.
If all this is surprising — it was for me! — I encourage readers to simply search for “Spamouflage,” and see the diverse dates and geographies where PRC-led disinformation campaigns have been discovered the past several years.
And this effort isn’t just focused on US, Canada, and Europe, as we’ll discuss more soon. The United States Institutes for Peace, a DC-based nonprofit, reports that
“China’s Africa-focused media propaganda is intended to improve African perceptions of the country and its political system while doing the opposite for the United States. Media outlets controlled by the Communist Party of China (CPC) offer African outlets cheap or free international content in order to amplify their messaging while disguising it beneath a veneer of grassroots legitimacy…. Chinese officials entice, cajole, and intimidate African journalists and editors to produce only positive stories about China, the CPC, and its African partners.”
Relevant to the Tiktok Bill today, the report points out that “China’s propagandists are creating an increasing amount of ‘soft’ content for African audiences, which Chinese firms are spreading to urban elites and youth via streaming services, social media, and phone apps.” And it recommends that the US “work with African partners to identify and expose China’s disinformation and to support independent African journalism.”
Exporting Control
“The key danger of the Great Firewall is that, by its very existence, it acts as a daily proof of concept for authoritarians and dictators the world over: proof that the internet can be regulated and brought to heel.” — James Griffiths, The Great Firewall of China
So China exports propaganda and disinformation, but what about censorship?
As Griffiths writes:
“In the past decade, China has begun exporting the technology used to power the Firewall to other countries. Russia has seen its internet, once freewheeling and dynamic, reined in and filtered, closing off one of the few remaining avenues for dissent against the Kremlin. Across Africa, from Zimbabwe to Ethiopia, China’s allies on the continent have adopted Beijing’s tactics of internet blackouts, cutting off access to whole regions during politically sensitive times, robbing people of their ability to organise or rally for change.”
He goes on to point out how Chinese tech companies have gone on to help autocratic regimes build similar systems of internet surveillance, “as Chinese companies, able to make huge profits thanks to the voracious demands of the domestic security state, have moved further afield in search of clients.” (Though it’s notable that US companies helped China build the country’s original firewall decades ago. Global capital knows few bounds and scruples.)
Huawei earned 15% of its global revenue in Africa in 2017, and while this may not be concerning, Foreign Policy reported a decade ago that the company worked with the Zambian, Ethiopian, Zimbabwean and other governments to build tools to monitor and censor citizens. And today, Chinese companies export facial recognition software — developed in part to monitor the country’s minority Uyghur population — to those same autocratic governments.
Playing Hardball
The efforts by the Chinese government to influence foreign politics discussed at the top of the post are mostly “soft” — deceptive efforts, primarily through social media, to influence politics and spread disinformation online.
But China’s government has taken far more drastic, direct efforts to shape the online environment beyond its borders as well. Just last month, the US Dept. of Justice pressed charges against seven Chinese nationals associated with the government-backed hacking group APT31, alleging the group committed “computer intrusions targeting perceived critics of China and U.S. businesses and politicians.”
And in 2015, China used a technology researchers have coined the “Great Cannon” to temporarily cripple two anti-Chinese-censorship sites hosted by Gihub, in San Francisco. According to Citizen Lab in Canada, “The operational deployment of the Great Cannon represents a significant escalation in state-level information control: the normalization of widespread use of an attack tool to enforce censorship by weaponizing users.”
Hong Kong’s 2019-2020 protests over the newly-introduced extradition bill weren’t immune from this form of Chinese intervention, either. Griffiths highlights how days after the protests broke out, police arrested an administrator of a channel called “International Waters” on Telegram, a security-centric communications app used by protestors. The police charged him with “conspiring to commit a public nuisance,” and forced him to unlock his phone to export all the messages he’d exchanged — along with a list of the group’s more than 20,000 members.
Soon after this, Telegram itself — based in Dubai — was hit by a massive, “state actor-sized” denial of service attack similar to that which struck Github in 2015. Pavel Durov, the app’s founder, wrote on Twitter that the attack likely came from China.
And again in Hong Kong that year, writes Griffiths, the Great Cannon “was wielded against LIHKG, a Reddit-style forum that had been a key platform, alongside Telegram, for organising the ongoing protests. During one sixteen-hour period, the site said it had been hit by more than 1.5 billion requests, crippling its servers as the Great Cannon attempted to force it offline.”
Big Markets Wield Big Power
China’s influence on how foreign countries perceive and portray the country isn’t just technical — and not just constrained to the internet, social media, or even hacking and DDoS attacks.
Last summer, the Barbie movie was banned by Vietnam and censored in the Philippines because of a scene with…. a map. Specifically, a map with contentious boundary lines showing Chinese ownership of a disputed region in the South China Sea. Studio executives claim this was simply an oversight, but according to Harvard professor Hue-Tam Ho Tai, “Aiming for the PRC market, [studio execs] are ready to accept the PRC’s view of geography. Disinformation works by repetition.”
Erich Schwartzel, author and Wall Street Journal journalist, wrote an entire book about China’s “soft” influence in Hollywood, and says the aim of the film industry is now to avoid “including anything in your movie that risks angering the state.”
As noted in Vox, Hollywood’s desire to appease China’s whims “explains why Doctor Strange changed the ethnicity of the Ancient One, a spiritual leader living in a mountainous monastery, from Tibetan to Caucasian, why Bohemian Rhapsody removed all references to Freddie Mercury being gay, and why Top Gun: Maverick originally removed the Taiwanese flag from Pete Mitchell’s flight jacket. (It reappeared after protests from fans and the departure of investor Tencent; as a result, the film never played in China.)”
And China’s threats against Hollywood aren’t without teeth. In 2022, John Wick and other Keanu Reeves movies were pulled from streaming platforms in the country after the actor expressed public support for Tibet.
I’d be remiss if I didn’t mention the many ways the Chinese government — as it’s begun to do with Hollywood — has repeatedly forced US tech companies to compromise their stated values, too. The Chinese market represents ~1.4 billion potential customers, and neither Hollywood nor Silicon Valley are immune to the siren’s song of all that sweet, sweet money.
Google, for instance, struggled on numerous occasions to balance its “Don’t Be Evil” ethos with hopes to entire China — entering, then leaving, then entering again. This most recently led to employee protests (in which I participated 💪) against “Project Dragonfly,” a China-specific Google search product that would comply with the government’s censorship demands.
And Mark Zuckerberg made numerous entreaties toward the CCP. He visited the country several times, ran a marathon in smoggy Beijing, and even reportedly asked Xi Jinping to name his first child. More insidiously (but less bizarrely), the company years ago “quietly developed software to suppress posts from appearing in people’s news feeds in specific geographic areas…. created to help Facebook get into China.”
Back to the TikTok Bill
Given what we’ve learned, it’s worth again considering our feelings around proposed US legislation to force TikTok to sell to US owners, or be banned. This flavor of protectionism feels to many of us I think strange, and distinctly un-American.
But it’s also worth remembering that the case of TikTok, or any other global social media platform, is itself extremely strange. Never before in human history has the majority of people in the world been able to interact with — and influence, for both better and worse — everyone else in the world. Proposed regulations of social media content even by US companies regularly butt up against First Amendment concerns, often forming a question of whether Facebook, Twitter and others represent a public square, or private space.
And just as these regulations must take creative forms to be effective — perhaps Twitter, Facebook YouTube and the rest are both public squares and private spaces; both broadcast networks and platforms for speech outside owners’ hands — we must, I think, also realize the unprecedented questions raised by the scruples and alliances of the entities controlling the technologies we use. And we must think creatively about how to respond to those.
As others have argued, I don’t believe any of this excuses the need to better regulate homegrown US-owned social media. But as discussed last week, that seems like a much larger, thornier question with a small likelihood of passing through Congress anytime soon. (Though maybe we — for once, through gritted teeth — can be happy about something Ron DeSantis did?)
Again, I’m most convinced by Max Read’s argument that fears the Chinese government could manipulate TikTok to influence international politics “rely on the stubborn belief that ‘platform manipulation’ is (1) possible, (2) coherent, or (3) effective.”
But we now know a lot about the truly expert ways China has found to manipulate the internet — and maybe all this should at least give us pause.
The case of TikTok and China today is controversial, so let’s conduct a thought experiment. What if the country in question was not China, but Russia — another country which sponsors coordinated disinformation campaigns, and which tends to “detain, silence and sideline” the leaders of private companies it disagrees with.
And what if the technology in question was not TikTok, but Facebook?
Consider what we’ve learned over the past decade about Russia’s vast efforts to weaponize social media — and more recently, fabricate entire online publications — to undermine faith in democracy, sway opinion over the war in Ukraine, and influence elections in the US and beyond.
For one, Russia’s influence on our politics would likely be more difficult to unearth, as Facebook, YouTube, and others regularly work in concert with US intelligence to discover and protect against foreign disinformation and hacking.
And for another, this influence would be much harder to protect against. The US and other democratic governments can at least pressure US companies to behave. Meta put a lot of resources into taking down the Chinese disinformation campaign last year for a reason, and I’m guessing it sounded a lot like Mark Zuckerberg hoping to avoid another Senate Judiciary Committee hearing. I could be wrong, but I don’t know how US or European policymakers might exert such control over ByteDance and TikTok if the app’s owners reside in a foreign country — and are beholden (which they are) to the very governments spreading that disinformation.
Would we feel comfortable with the Russian government leveraging so much power — much of it inscrutable — over Facebook, Instagram, and WhatsApp?
Or would we feel at least a little apprehensive?
Song of the Week: Ekber Turdi — Wherever I May Roam, a Metallica cover played on a truly *badass* guitar. Turdi comes from Xinjiang, a far-west province of China home to many in the Uyghur ethnic minority. China has focused many of its repressive technology policies there, including deploying facial recognition software and at times shutting down the region’s entire internet.