CLM Insights Interview with Josh Chin and Liza Lin
Josh Chin and Liza Lin. Surveillance State: China’s Quest to Launch a New Era of Social Control. St. Martin’s Press. Sept. 2022.
What are some of the popular misconceptions about the Chinese surveillance state? In what ways have we inflated or underestimated its capabilities?
There are two big misconceptions about China’s surveillance state. The first is that digital surveillance is about oppression. There is no question the Communist Party has deployed its tracking tools in dystopian ways. Instances such as the ethnic reengineering campaign in Xinjiang go beyond what even Orwell imagined. But the party’s pursuit of security is unfolding on multiple levels. In wealthy cities like Hangzhou, where tech companies and local officials work closely together, the same surveillance tools are used to optimize how the government delivers services and solves problems. Chinese leaders often conjure a techno-utopian vision of life made safer, more convenient, and more predictable through the harvesting of data – a potential new source of legitimacy for the party as GDP growth slows.
The second major misconception concerns how state surveillance works in China. We tend to think of it as a technological project, but it is at least as much about propaganda. Chinese state-run media like to portray the Chinese state as being all-seeing – that’s not exactly true. Several of the capabilities that the Chinese surveillance state claims, or has ascribed to it, turn out to be either exaggerated or fabricated outright. Yet, that doesn’t matter. What matters is that people absorb enough discussion of those capabilities to internalize the idea of being watched, then they adjust their behavior in ways the party wants.
Based on the description of the operations of the surveillance state in your book, China also relies heavily on labor-intensive methods. How does the tech-intensive approach complement or reinforce the labor-intensive approach to surveillance?
The new wave of surveillance technology is sold as a way to automate tracking tasks that used to be crushingly labor-intensive. In Hangzhou, we gained access to the command center of an AI-powered camera system called “City Eye” that is used by the local chengguan to automatically spot and record street-level infractions, such as litter or unlicensed vending. The local commander told us that in one six-month period his human street patrols had singled out 2,600 potential violations. Over the same period, the AI system (working literally around the clock) had flagged 19,000. The local commander had a team of people watching screens in the command center, but he said the system did not really need them. He only kept them around because he would have felt bad about firing them.
But AI has not eliminated human labor so much as it has shifted it. To work properly, AI-based surveillance systems need to be fed reams of data. That data must be collected, labeled, and entered into the system – a job that, in many cases, is still best performed by humans. You see this most clearly in Xinjiang, where the government has hired armies of low-level security personnel and has dispatched thousands of civil servants to the countryside, tasking both groups with collecting information on Uyghurs that is then loaded into the region’s centralized data platform. The authorities there have also installed innumerable digital checkpoints – at bazaars, hotels, banks, markets – where Uyghurs have to scan their faces and ID cards. All these points must be staffed with security personnel in case someone the system deems to be “unsafe” tries to pass through and should be detained. Interestingly, human beings often end up being the weak link in the Chinese surveillance state, whether it is a security guard falling asleep at a checkpoint or a rogue police employee selling surveillance data on the black market.
Some foreign countries, in particular in the developing world, are eager to import China’s surveillance model. What makes the Chinese surveillance model easy to copy? What makes it difficult to copy elsewhere?
China has made digital surveillance accessible to other governments by making it cheap and relatively easy to get started. It used to be that only the world’s wealthiest governments could afford high-definition security camera networks with facial-recognition capabilities. Thanks to Chinese manufacturers, almost any government can now buy them, often with favorable loans from Chinese banks. China also offers good customer service. Prior to the Ugandan police purchase of a $126 million surveillance system from Huawei in 2019, the Chinese ambassador in Kampala flew a group of Ugandan police to Beijing to witness how China’s Ministry of Public Security puts the equipment to use. Later, Huawei stationed employees at the Ugandan police headquarters to help them troubleshoot any issues that might pop up. This is a pattern that repeats itself in other countries.
Despite all that, no other country has become remotely close to replicating the scope or sophistication of the Chinese surveillance state. Why is that? First, digital surveillance systems require large numbers of relatively well-trained people to run them. Few countries have an immense bureaucracy capable of keeping a modern surveillance machine humming. Second, a proper surveillance state cannot function without a large amount of well-organized data. At the heart of China’s systems is a centralized ID database, with the name, age, address, and biometric photo of every adult – something not even the U.S. can boast.
Chinese citizens appear to have a great deal of ambivalence about state surveillance. What causes such ambivalence? Does such ambivalence have the potential of curtailing the reach and intensity of state surveillance?
It is important to note that the degree of ambivalence varies widely. In huge swaths of the country, even today, you would be hard-pressed to find someone who cares deeply about privacy. You do find a noticeable level of privacy awareness in the bigger cities, like Beijing and Shanghai. But what is interesting is the way the government has managed the rising tide of interest in privacy. You might expect the authorities to censor it. Instead, they have encouraged it, but only when it targets the private sector. The government recently passed a personal information protection law that is among the strictest in the world when it comes to how companies like Alibaba and Tencent handle user information. But it has huge carve-outs for the government.
For the most part, Chinese people seem to be content to direct their anxieties about surveillance at the tech companies rather than at the government. That started to change somewhat with the pandemic, as people are now chafing at the intensity of government scrutiny and control under Xi Jinping’s Zero-Covid policy. But we will have to wait to see whether the outrage achieves escape velocity.
What are the key differences and similarities of state surveillance in Xinjiang and that in the rest of China? Is today’s Xinjiang tomorrow’s China?
The differences are growing smaller every day. Prior to the pandemic, Xinjiang was unique in applying intensive real-time surveillance to an entire population. In other parts of the country, police reserved that level of scrutiny for certain categories of people: criminal suspects, known drug users, dissidents, and a small number of other “persons of interest.” Since the arrival of Covid-19, however, we have seen surveillance approaches that were pioneered in Xinjiang spreading everywhere. They started with the lock-downs of residential compounds, a tactic known in Xinjiang as “closed-style management” (fenbishi guanli). Then there are the smartphone-based health codes, which track and categorize everyone in the country according to their risk of exposure to the coronavirus – strikingly similar to the way authorities in Xinjiang categorize Uyghurs according to their alleged exposure to the “ideological virus” of radical Islam.
Han Chinese are still not subject to the same intensity of tracking, privacy invasion, or digital phrenology about their future trajectories as are Uyghurs or other Turkic Muslims in Xinjiang. But similar to Xinjiang’s minority populations, they are experiencing the hard edge of AI-powered state control on a broad scale.
How effectively has China deployed its surveillance technology during the Covid pandemic?
During the early phase of the pandemic, the Chinese surveillance state performed about as well as Beijing could have hoped. It was so successful, in fact, that Xi Jinping dialed in to the Group-20 Summit in November of 2020 and proposed that the rest of the world adopt the Chinese system of digital health certificates. At the time, China reported it had recorded fewer than a total of 87,000 Covid-19 cases, with just over 4,600 deaths. The United States, meanwhile, was on the cusp of 12 million cases and a quarter million deaths. Nearly every Chinese person with whom we talked said they were happy to sacrifice a little privacy in exchange for being able to live relatively normal lives, while other countries were floundering in the grip of Covid anxiety.
The story started to change with Omicron, which moved too quickly for even China’s surveillance state to track. That led the government to turn the machinery of digital surveillance on the people rather than on the virus. Suddenly, you started to see nightmarishly dystopian video footage of robot dogs and drones tracking down people who broke the quarantine. You saw questions and pushback online about the utility of Zero-Covid and the competence of the government. We are at an interesting moment now in which the government’s Covid policy is generating real public anger – to the point that someone dared to hang a banner in Beijing calling Xi Jinping a “traitorous dictator” just days before he was due to embark on his third five-year term as leader. But even if China has failed to keep Omicron under control, the surveillance state has proven itself effective at dealing with dissent. Perhaps that is why Xi feels confident in sticking with Zero-Covid in the first place. It no longer matters if people disagree with Xi. He has the tools, both technological and political, to enforce his will.