Preface. China hopes to nip rebellion in the bud before a revolution takes place by excessive monitoring and then respond quickly with improvements to prevent one. And show the world their system is better than democracy.

But then there is the dark side. It has enabled China to oppress the Uyghurs with their “all-seeing” surveillance of over 15 million people.

And so creepy: The volume of personal data theoretically available to the Party strains comprehension. By the start of 2020, close to 350 million cameras recorded the comings and goings in Chinese streets, in public squares, in subway stations, and around commercial buildings. More than 840 million smartphones bounced around in the purses and pockets of Chinese pedestrians, sending a steady stream of location data to telecom operators. Mobile payment systems logged millions of transactions a day in databases that offered searchable, ever-evolving portraits of human activity rendered in breathtaking detail.

Government officials can scrutinize your private chat history, reading and viewing habits, internet purchases, and travel history and can crunch the data to judge how likely you are to help or harm public order; in which artificial intelligence companies work hand in glove with police to track down fugitives, find abducted children, and publicly shame jaywalkers; and in which public services, rewards for good deeds, and punishments for misbehavior are all delivered with mathematical precision and efficiency.

The history of surveillance goes way back. Nations have inventoried their people and resources for millennia (for taxation and more). The extent that China can monitor people and how they use the information is scary though.  Perhaps a glimpse of our own future.  By 2019, Chinese surveillance systems had expanded across the globe. Huawei had installed Safe City systems in 700 cities across more than 100 countries and regions. ZTE in 160 cities spread across 45 countries.

When the grid goes down permanently, so will electronic surveillance. Meanwhile it is amazing how much information is being gathered. This is about government spying, I imagine that corporations collect more than we know as well.

What follows are some of my kindle notes to give you an idea of what the book is about (or at least excerpts from the parts I was interested in).  But there’s so much more, buy the book if you find this of interest.

Alice Friedemann  www.energyskeptic.com  Author of Life After Fossil Fuels: A Reality Check on Alternative Energy; When Trucks Stop Running: Energy and the Future of Transportation”, Barriers to Making Algal Biofuels, & “Crunch! Whole Grain Artisan Chips and Crackers”.  Women in ecology  Podcasts: WGBH, Financial Sense, Jore, Planet: Critical, Crazy Town, Collapse Chronicles, Derrick Jensen, Practical Prepping, Kunstler 253 &278, Peak Prosperity,  Index of best energyskeptic posts

Chin J. Lin L (2022) Surveillance State: Inside China’s quest to launch a new era of social control.  St. Martin’s Press.

Surveillance in history, in China, and elsewhere

State surveillance has been with us as long as there have been states. As far back as 3800 BCE, Babylonian kings in what is now Iraq pioneered an embryonic form of mass data collection, using cuneiform and clay tablets to keep a constantly updated record of people and livestock. The ancient Egyptians, Romans, and Persians followed suit, conducting regular surveys of their populations that grew more detailed and sophisticated over time—efforts to render their societies “legible,” in James C. Scott’s memorable description. For any state, being able to read the populace—being able to see who lives where, how many people are in a given household, what they own, and how much they earn—is critical for basic acts of governing like levying taxes, conscripting soldiers, and doling out grain.

In “The Soft Cage”, a book on the history of American surveillance, Christian Parenti describes how colonists in Virginia invented an embryonic form of the modern ID card with rules that required Native Americans, indentured Irish servants, and, later, plantation slaves to carry passes when traveling in or out of the territory.

One of the more appalling episodes in American corporate history: IBM’s collaboration with the Nazi regime in the lead-up to and during World War II. In the early days of Adolf Hitler’s ethnic-cleansing campaign, the Nazis struggled to identify the large proportion of European Jews who had assimilated and could effectively “hide” their Jewishness. To overcome this, Nazi demographers would have had to conduct detailed surveys that traced the lineages of entire populations back several generations.

As journalist Edwin Black documented in his book IBM and the Holocaust, IBM had technology perfectly suited to satisfy Hitler’s “data lust”: a machine that automated the tabulating of census data by reading and sorting punch cards marked to record an individual’s traits. Nazi Germany leased hundreds of them and bought punch cards in the hundreds of millions. IBM’s involvement in the Holocaust didn’t stop at helping the Third Reich count Jews. As Hitler’s project moved from identification to ghettoization, forced labor, and eventually genocide, the American corporation adapted its machines to each new task. Through a German subsidiary, IBM helped automate the scheduling of trains that crisscrossed Europe, carrying supplies to German troops and Jews to concentration camps. The machines became critical to the operations of the camps themselves and, toward the end of the war, to the enormous logistical challenge of gassing and shooting as much of Europe’s Jewish population as possible. “The technology had enabled Nazi Germany to orchestrate the death of millions without skipping a note. At the same time, the company touted its patriotic leasing of other machines to the US government, which used them to organize its own troop movements and corral Japanese-Americans into internment camps.

As Hitler’s horrifying endgame came into clearer focus, IBM executives in the corporation’s New York headquarters shielded themselves from the details of how their machines were being used in Europe. Long after the war ended, the company continued to deny responsibility for the way its machines were used in Germany. IBM established the American template for using strategic corporate ignorance as a lubricant to the collection of profits from activities that violate democratic values.

Throughout history, social control has been inseparable from the harvesting of personal information.

Chinese dynasties stumbled down a wide variety of paths to collapse, but in many instances they first lost their way as a result of the same problem: breakdowns in the flow of information. The oldest complete census still in existence dates to 2 CE, when the court of the Han Dynasty—from which the country’s dominant Han Chinese take their name—dispatched officials to compile a written record of all 57.7 million people then living in what is now considered the Chinese heartland. Historical records suggest Chinese emperors ordered regular censuses over the centuries that followed. But between surveys, they were typically cloistered in palaces, dependent on vast bureaucracies to tell them what was going on outside. Corrupt or incompetent officials often distorted the picture rulers received, which in turn led to paranoia and bad decisions, and sometimes to collapse.

A student of history who reportedly slept in a bed more than half covered in books, Mao was steeped in tales of emperors who had been dethroned after losing touch with what was happening outside their palaces. When he came to power, he made certain the Party had eyes everywhere. Armies of journalists for Party-run newspapers filed confidential “internal reference” reports to top leaders from every province. A 300,000-strong Ministry of Public Security (MPS) scoured the country for spies and other internal threats. Party cells in communes and urban work units tracked every aspect of residents’ lives, from meals and medical care to education and entertainment. Mao insisted on implementing the Great Leap Forward, a series of delusional industrial policies that led directly to the Great Chinese Famine.

The final death toll is estimated at around 30 million people. As the famine began to take hold, leaders in Beijing should have known what was unfolding. But rather than reflect the truth, the Party’s informers, like their predecessors over the course of history, told Mao what they thought he wanted to hear. Local officials suppressed reports of starvation, and newspapers staged photos of grain crops so dense that children could stand on top of them. Networks of domestic agents run by the MPS, meanwhile, were hobbled by bureaucratic infighting, inexperience, and incompetence, which on occasion led to them spying on each other without knowing it.

A free press and the prospect of election defeat, he argued, are enough to scare leaders in democracies into taking action to head off mass starvation before it happens. In China, though, censorship and propaganda made it impossible for regular people and officials to grasp the nationwide nature of the disaster.

“While the famine was going on, there was also a starving of information,” Sen said in a 2001 interview. The censorship “had the effect of hoodwinking not only the public but ultimately hoodwinking the state.

In the 1960s, the state of California invited bids from local aerospace companies to use systems engineering to inject scientific efficiency into its management of areas like transportation, criminal justice, and social welfare. The hope, Governor Pat Brown said, was to use “know-how that will get a man to the moon to get Dad to work on time.” Nearly all of the efforts failed to produce improvements. Conducting a review of how they went wrong, University of California sociologist Ida R. Hoos argued the projects proved that the effectiveness of social systems is measured in values that can’t be reduced to math.

Hiring a systems engineer to solve malfunctioning social systems, she wrote, was like calling “a hydraulic engineer to cure an ailing heart because his specialty is pumping systems.

Word of California’s ill-fated experiments appears not to have reached Qian. Or if it did, he dismissed it. In 1981, he worked with his protégé, a scientist named Song Jian, on an update to his book, which made an explicit case for applying systems engineering to social problems. “Under the conditions of socialism, a new science will eventually be born: That is social cybernetics,” the pair wrote

It was true, they wrote, that social cybernetics wouldn’t work in capitalistic societies, which in Friedrich Engels’s terms were too vulnerable to the “anarchy” of market-based production. Socialistic societies were different, however. They could be designed in such a way as to be automatically self-correcting. Qian and Song, the rocket scientists, were proposing a guidance system for society based on the same cybernetic feedback concepts that missiles used to automatically home in on their targets.

Having loosened the reins on Chinese society, the Party began to worry it might be bucked out of the saddle. On the evening of June 3, leaders in Beijing ordered troops back into the city, authorizing them to “use any means” to clear the square. China’s government has never accounted for how many people were killed by the volleys of bullets the soldiers unleashed. Estimates by survivors range from a few hundred to several thousand. Regardless of the death toll, the protests and the violence that ended them carved a wound in the Party’s psyche that forced a change in its calculus. From that point on, it would continue with market reforms but end any serious experiments with political liberalization.

An essay by Qian in a journal run by the Chinese Academy of Social Sciences would later seem eerily prophetic. In it, Chinese society was described as an immense supersystem made up of economic, political, and ideological subsystems. It was folly to expect changes in one not to have ripple effects in the others. If the development of these three facets of civilization isn’t harmonized”—Party-speak for made to work smoothly without conflict—from the point of view of systems science, it will push the entire social system from order to disorder, chaos, and collapse.

The following year, Qian took his ideas for managing society a step further. In an article for the Chinese Journal of Nature, he described human societies as an example of an “open complex giant system,” a category that also includes the human nervous system and galaxies like the Milky Way. These massive systems are made up of millions of subsystems that interact both with each other and with the outside world. Compared to smaller closed systems, like those used to control rockets and satellites, they are vastly more difficult to predict and control with math alone. “Not even a supercomputer is up to the task,” Qian wrote, “and there won’t be enough computing power in the future to do this work either.” He

As China limped into a post-Tiananmen world, accolades for Qian continued to flow. In 1991 he was given the First-Level Model Hero Prize, China’s highest honor for a scientist. Around the same time, propaganda officials launched a “Learn from Qian Xuesen” movement. His ideas began to be taught at the Central Party School, the country’s top training academy for political leaders in Beijing, and were credited by scholars with influencing the political theories of then-president Jiang Zemin.

The ideas Qian had brought to China about how to manage society offered a way not just to fight fires but possibly to prevent them from breaking out in the first place. Signs of Qian’s influence in this sphere started to become apparent in the early 2000s, when the Party instituted early versions of a “grid management” system to help police keep tighter control over urban neighborhoods. The system, which divided neighborhoods along a grid, was built to encourage a smooth flow of information and enable quick action to head off problems. Each square on the grid was assigned a manager, whose job it was to report goings-on to the police, take photos and video where necessary, and intervene when trouble appeared to be brewing.

The most impactful of the projects, the Golden Shield project, focused on exploiting information to neutralize security threats. One critical component, built using cutting-edge firewall technology from the United States and Canada, was a fine-tuned system for filtering out unwanted internet content. The other key element was the construction of a computer network connecting the Ministry of Public Security with local police bureaus around the country, combined with a national online database containing the ID numbers and personal information of every adult in the country.

On that foundation the Party hoped to build a surveillance system that would incorporate tracking of internet use at the individual level, closed-circuit cameras, and smart ID cards. Plans also called for the eventual incorporation of speech and facial recognition—a full 15 years before they would be put to use in the crackdown on the Uyghurs.

After Qian’s death, the insights he and his students had shared with the Party helped it continue to maintain control, and thrive, in ways that have surprised the world. While most Chinese people have come to realize Mao Zedong was merely human, one Chinese scientist told Science magazine in 2018, “To a circle of scientists in China, Qian Xuesen is now, in their mind, the new god.” If Qian seemed prophetic after Tiananmen Square, however, putting his commandments into practice could be a challenge. In the Hu Jintao era, both grid management and the Golden Shield had short-circuited on several occasions, sometimes to a degree that made leaders in Beijing nervous.

The rise of Xi

The Party’s effort to reboot its approach to control wouldn’t begin in earnest until after November 4, 2012, the day that it anointed a powerful new general secretary.

In November 2012, as the audience gathered for Xi’s coronation, it was clear the next generation of Chinese leaders would have to take the country in a more deliberate direction, but the Party, ever secretive, had been tight-lipped about its plans. When he strode onto the stage in the Great Hall of People to an explosion of camera flashes to deliver his first speech as the Communist Party’s new leader, Xi Jinping was still a mystery to many in China. The paunchy 59-year-old with an impassive smile had been chosen by Party elders to rise to the top post and the country’s presidency five years earlier but gave few clues about what he planned to do once he got there.

The Party was too lumbering and corrupt to manage further change, the reasoning went, and could only solve the country’s current problems by loosening its grip. Xi began to lay out a nationalistic vision that placed the Party at the center of a final push to reclaim China’s lost glory.

Three years before Donald Trump started stitching “Make America Great Again” on red baseball caps, Xi had delivered the Chinese equivalent. Over the next few years, Xi would outline an ideology that wasn’t merely anti-Western but a throwback to a troubled earlier era of Chinese politics. He cast the China Dream as the latest evolutionary phase of a grand historical project that began with Mao Zedong’s revolution, one that gave the Communist Party a reason to reassert its primacy in all aspects of life in the country. He became the “Chairman of Everything,” sitting at the head of every important committee.

In one of his first major speeches as president in 2013, he declared it a priority to turn China into a tech power. “An important reason that Western countries were able to hold sway over the world in modern times was that they held the advanced technology,” he said.

Outwardly, becoming a leader in strategically important technologies like artificial intelligence would help China compete with the United States. Inwardly, it would help the Party build a new system of control that would ensure its own well-being.

The volume of personal data theoretically available to the Party strains comprehension. By the start of 2020, close to 350 million cameras recorded the comings and goings in Chinese streets, in public squares, in subway stations, and around commercial buildings. More than 840 million smartphones bounced around in the purses and pockets of Chinese pedestrians, sending a steady stream of location data to telecom operators. Mobile payment systems logged millions of transactions a day in databases that offered searchable, ever-evolving portraits of human activity rendered in breathtaking detail.

Government officials can scrutinize your private chat history, reading and viewing habits, internet purchases, and travel history and can crunch the data to judge how likely you are to help or harm public order; in which artificial intelligence companies work hand in glove with police to track down fugitives, find abducted children, and publicly shame jaywalkers; and in which public services, rewards for good deeds, and punishments for misbehavior are all delivered with mathematical precision and efficiency.

much of this vision remains aspirational. Some of it the Party may only partly realize, and some of it might not materialize at all.

Annual double-digit increases in GDP and rising standards of living guaranteed that Chinese people would fall in line. But over the past ten years, that growth has begun to slow. Exploding debt and demographic pressure threaten to weaken it further. Now, with the previous social contract starting to fray, the Party has turned to digital surveillance to help it write a new one.

Xi is a keen enough student of history to know that the utopian dreams of Mao and other twentieth-century totalitarians ended disastrously. He understands that the Cold War era’s high priests of surveillance, East Germany’s Stasi, failed to foresee their own demise in time despite compiling dossiers on suspicious citizens that would have stretched more than 111 miles laid end to end.

The difference for Xi lies in technology. What the Stasi did with close to 100,000 agents and more than 170,000 informants, China’s own domestic spy agency has begun to automate using algorithms and its suffocatingly expansive network of digital sensors

By mining insight from surveillance data, it believes it can predict what people want without having to give them a vote or a voice. By solving social problems before they occur and quashing dissent before it spills out onto the streets, it believes it can strangle opposition in the crib.

From Africa to Europe, countries around the globe have shown an interest in what Xi is selling, in many cases literally purchasing Chinese-made surveillance systems designed to hone and strengthen state control. The list includes autocracies like Saudi Arabia and hybrid regimes like Uganda but also city governments in advanced democracies like France and Germany.

All of this comes at a time of profound weakness for democracy. Countries where democracy is deteriorating have outnumbered those where it’s improving every year since 2006, In 2021, the discrepancy grew to record size, with 73 countries backsliding and only 28 recording improvements. Nearly three-quarters of the world’s population, including 330 million Americans,

To maintain control there, he established a vast paramilitary organization, the Xinjiang Production and Construction Corporation (more commonly known as the Bingtuan), made up of decommissioned Han Chinese soldiers. The Bingtuan’s mission was threefold: enforce security, promote economic development through modern agriculture, and dilute the influence of minority populations by encouraging waves of Han migration. It was remarkably successful in meeting the last goal. The Han population in Xinjiang rose from 5 percent in 1949 to more than 40 percent in 1978.

After Mao’s death in the late 1970s, the Party returned to the Soviet model, loosening religious controls and introducing affirmative-action-style policies, such as added points for Uyghurs on college entrance exams. But in the early 2010s, still unsatisfied with Uyghur loyalty, it started relying more on brute force to muscle them into submission

Despite being remote, Xinjiang was central to Xi’s most cherished foreign policy idea: a trillion-dollar global infrastructure plan called the Belt and Road Initiative that would resurrect old Silk Road–era trading systems and reorient international commerce around the Chinese economy. One aim of the plan was to funnel Chinese goods through Central Asia to the Middle East and Europe through railway networks that would travel, just like the Silk Road caravans, through the mountain passes of Xinjiang. An uprising among Uyghurs could scuttle those plans, eviscerating billions of dollars in investment. It would also make the Party look weak, both abroad and, worse, at home.

Over his first four months in charge, he built 4,900 convenience police stations—giving Xinjiang twenty times as many police stations per capita as Chicago—the vast majority in neighborhoods dominated by Uyghurs. He would order construction of at least 2,800 more over the following year and a half. Each police station was connected to a network of surveillance cameras and to a series of plastic alarm buttons installed in nearby businesses that owners could push to alert authorities of suspicious behavior.

History is littered with examples of how data gathered even under neutral or positive auspices can later be exploited for regrettable, and sometimes terrifying, ends. Censuses conducted in Europe in the early twentieth century that asked people to note their religious affiliation later helped the SS identify the largest pockets of Jews in each country they occupied during World War II.

The US census of 1940, the country’s most ambitious and wide-ranging to that point, would be used two years later to help Washington identify blocks of Japanese Americans and ship them off to World War II internment camps.

The more difficult and more urgent effort came in the countryside, where resistance to the Party was strongest. In Tibet, Chen had ordered groups of Party members from the cities to live temporarily in villages, where they delivered government dictates and conducted “gratitude education” to teach locals how to appreciate life under Party rule. Xinjiang had already begun copying the practice before Chen took over there. Now he modified it, tasking the visitors with collecting intelligence on their hosts. He sent more than a million people (most carrying gifts like electric kettles or cooking oil, and most, though not all, Han Chinese) to rural areas scattered around the region, where they invited themselves into Uyghur and Kazakh homes and portrayed themselves as “big brothers” or “big sisters.

The manual told them not to jump straight into interrogations but rather to find a natural place in the conversation to slip in pertinent questions and to expect to conduct two or more visits before a family would start to reveal what they really thought.

A smartphone app that police used to interact with it, reverse-engineered by Human Rights Watch, showed that the data came from a web of sources that extended beyond surveillance cameras and human intelligence gathering. Information from security checkpoints, visitor management centers at residential compounds, and Wi-Fi “sniffers” that detect the unique ID numbers of smartphones and computers all flowed into the platform, along with information about mail deliveries, electricity use, and gas station visits.

The Qing Dynasty’s inability in the early 19^th century to match the technological progress of the newly industrialized West, a consequence of imperial arrogance and institutional inertia, is widely blamed for ushering in the Century of Humiliation. The catastrophe of Mao’s pursuit of industrial self-sufficiency during the Great Leap Forward, and the famine that followed, only buttressed the sense of China as having lost, perhaps permanently, the technological moxie that underpinned its former greatness.

The notion of China as a technological has-been destined to cower in the face of Western know-how persisted into the early years of the twenty-first century. Western elites believed it at least as strongly as their Chinese counterparts, buoyed by confidence in the explosion of information technology.

Clinton argued that letting China in would open the country up further to the internet and, by extension, to the ideals of freedom and rule of law. The Communist Party might try to control cyberspace, he said with a wry smile, but doing so was “like trying to nail Jello to the wall.” The punch line quickly morphed into a catchphrase among foreign policy wonks in Washington for the conviction, almost ubiquitous at the time, that the internet was an unstoppable democratizing force.

The White House, like the developed world writ large, was convinced Beijing didn’t know what it was getting into. For a long time, it seemed they might be right. Over the next decade, China’s government poured billions of dollars and incalculable hours into building and upgrading its system of internet filters in an effort to nail the Jello in place.

Inside what has since come be known colloquially as the Great Firewall, officials nurtured the creation of what was in effect a parallel internet populated with Chinese versions of Google (Baidu), Amazon (Alibaba), and Facebook and Twitter (Weibo), each beholden and ultimately subservient to Beijing.

By the time Xi was ready to take over, the Party had built the most sophisticated censorship apparatus on earth—one that largely had succeeded in controlling access to information for hundreds of millions of internet users. But the furor over Wenzhou and explosions of social media angst showed it was still struggling to pin down the gelatinous blob of human activity online.

One of the country’s most important goals, China’s AI development plan revealed, was to build a matrix of smart cities, smart courts, smart schools, smart hospitals, and smart government agencies that would help solve “hot and difficult issues in social governance” and “make social operations more safe and efficient.

One of the first places the Party put AI to use was in taming dissent online. Under pressure from Beijing, the country’s internet companies built hybrid filtering systems that blended human “content managers” with machine censors. New algorithms fed with mountains of data were now able to scan messages by the millions and automatically flag any that might cross the Party’s red lines.

Crucially, they could also scan images, PDF files, and video—the sort of painstaking task that previously had to be done by hand. The machine censors could still miss sensitive content, especially if it involved a new topic they hadn’t been trained to spot. In those instances, the human censors would step in to mop up. The cat-and-mouse game continued, but the Party was getting much better at it.

Xi also expanded technological controls in the real world. Almost immediately after he came to power, the Party’s top security officials worked with police and the ministry in charge of information technology to accelerate the rollout of a vast surveillance program, called Project Skynet, that required the installation of batteries of high-definition video cameras for real-time monitoring in cities across the country. Over time, police departments in some cities connected the Skynet cameras to new facial recognition systems and used them to track down fugitives who had assumed fake identities.

In 2016, the Party’s Central Committee, evidently pleased with Skynet’s results, approved Sharp Eyes, a new program that expanded the installation of cameras into rural areas.

Cameras multiplied, not just in Xinjiang but outside far-flung Tibetan monasteries and in remote hamlets scattered around the country. State media reported that police in Sichuan, the populous province known to most of the outside world for its spicy cuisine and pandas, had installed Sharp Eyes systems in 4,800 villages and other residential areas in the space of a year. Reports from other provinces suggested similar explosions. Sifting through public tender documents, researchers at the Asia Society discovered that by 2020 the government bill for Sharp Eyes projects alone added up to around $2.1 billion.

The government’s aim was to install video surveillance in all of China’s “key public spaces” by the end of 2020—a goal it would largely achieve. But the new networks were not just larger. They were also far better organized and interconnected, feeding into command centers where local authorities could collect and analyze surveillance footage covering broad areas with relatively little effort.

The government’s goal, Meng said, was to build systems that bolstered trust in the Party by automatically detecting signs of negative shifts in social media conversation and using that information to inform decisions. For example, such a system could scan Weibo for certain combinations of keywords previously associated with local environmental protests, like “corruption” and “pollution” and “factory,” and alert provincial officials when they reached a certain threshold in posts from areas under their control.

The officials could then step in to assuage local concerns before residents took to the streets.

During her search for a new framework to understand what she was seeing, Hoffman stumbled onto a paper about self-regulating “autonomic” computer systems. The systems had been developed by American technology giant IBM in the early 2000s in response to constant increases in the complexity of modern computing. The company couldn’t find enough human IT managers capable of keeping the new systems running, so it began to develop computers that could optimize themselves The notion of autonomic error correction matched what Hoffman was seeing in China.

One example she cited was the Social Credit System. A project the Party began to promote widely in 2014, it sought to combine regularly updated measures of financial, social, and political behavior into a single mechanism that would encourage trustworthiness in businesses and individuals through a mix of rewards and punishments. The system was complex and flawed, but its conception encapsulated the Party’s approach toward regulating society. In theory, Hoffman argued, it could be automated to incentivize obedience and preempt discord.

So many of the problems the Party encountered—so many of the protests that erupted either online or in the streets—were a response to abuses of power, greed, corruption, and criminal negligence among local officials, encouraged by a regime that allows political manipulation of the legal system. Building the political equivalent of an autonomic nervous system might be something of a fantasy, but even a partly functioning version of it would allow the Party to manage the country with vastly greater efficiency than it had in the past.

In recent years, state media had highlighted the books that Xi, a purported bibliophile, had on his shelf to start the year. Next to Marxist texts Das Kapital and The Communist Manifesto sat two books on artificial intelligence: Augmented: Life in the Smart Lane by futurist Brett King, a forward look at what life will be like in the coming age of predictive technology, and The Master Algorithm, a layman’s introduction to machine learning by computer scientist Pedro Domingos that imagines a universal learning algorithm capable of deriving all past, present, and future knowledge from data. (Domingos later told the German magazine Der Spiegel it was “both exciting and scary” to learn his book was on Xi’s shelf.

China’s leaders wanted to redefine government using the same tools that Google, Facebook, and Amazon had used to remake capitalism. In an ideal future, with enough data and the right algorithms, they wouldn’t have to impose their will with violence the way Stalin, Mussolini, and other twentieth-century tyrants did. Instead, like a Google advertiser, the Party would be able to divine when and how to nudge its subjects to behave in ways it prefers. They could engineer away dissent. Instead of freedom or oppression, China would have optimization

It was unlikely the Party would ever achieve such a flawless idea of control, but as Xi Jinping entered the middle of his second term he seemed determined to travel as far down the path as possible.

For as powerful as the Communist Party is, its ambitions for social control depend heavily on companies in the one sector of the Chinese economy not dominated by the state. Nurtured inside a walled garden, China’s tech giants are unique. They often boast deeper insight into the lives of their users than Facebook, Google, and Amazon, those pioneers of what author Shoshana Zuboff memorably dubbed “surveillance capitalism.” They also have a much closer relationship to the state than their Silicon Valley counterparts, depending on government support for development and, ultimately, survival.

At the same time, like all creatures of global capitalism, they are driven by a powerful thirst for profit—one that doesn’t always align with the state’s interests. To the Party, that makes them both an invaluable asset and a source of profound anxiety.

Inside the Great Firewall, the Party granted tech companies surprising leeway. As long as they obeyed censorship orders, internet companies were left pretty much to do as they pleased

The new and improved WeChat allowed users to make purchases and transfer money by scanning QR codes. That meant vendors could accept mobile payments simply by printing a code on a piece of paper and taping it up somewhere customers could scan it. In a cash-bound country that had never embraced credit cards, it was a transformational development. Nearly everyone in nearly every city, township, or village in China now conducts their business through Alipay or WeChat, whether it’s buying street food, booking flights, or investing in money-market funds. In 2020 alone, Chinese people used mobile payment systems to conduct more than 123 billion transactions worth $67.5 trillion, nearly eleven times the total global credit and debit payment volumes recorded by MasterCard that year.

The way people spend money is the most reliable and revealing window into their desires and beliefs.

Amazon can only see how money is spent on its own platforms. Tencent and Alibaba, meanwhile, can see how users spend money across China. Through their other services, they can also see where users spend time, who their relatives and friends are, what movies they like, how much electricity they use, and what they like to do on vacation—a degree of behavioral insight that is breathtaking in its breadth and clarity.

Feng used it to pay for nearly everything: his morning coffee, Ralph Lauren shirts at the nearby outlet mall, snacks from a vending machine near the entrance to his office. On days when he didn’t feel like driving, he used it to take the bus or subway.

Before a mass purging of records in 2021, China’s public court database contained thousands of files that showed prosecutors using mobile payment data and other information from Alibaba and Tencent platforms to convict people of crimes. In some cases, police had pulled the data directly from the suspects’ phones. But others were mysteriously vague about where it had come from.

Chinese laws dictate that internet content firms must monitor and control the information posted on their websites, or else they risk losing their license to conduct business. Those rules exert significant pressure on internet companies to develop systems capable of fine-grained monitoring and filtering content. “If I say something illegal or bad using Gmail, then I am responsible for it, not Google,” said Lokman Tsui, a former journalism professor at the Chinese University of Hong Kong who previously served as Google’s Head of Free Expression in Asia and the Pacific. “In China, it’s the opposite.

Bored during one confinement ahead of a Communist Party meeting in 2017, Hu took a friend’s advice and bought a slingshot online for target practice in the yard of his apartment complex. He used WeChat Pay. A day later, he recalled in a phone conversation, a state security agent asked why he’d bought a slingshot. “They thought I was going to use it to shoot out their cameras,” he said. “They understand what you spend money on, where you spend it.” The example was mundane, but that was the point. Even something as harmless as buying a slingshot was enough to get flagged. Everyone, Hu said, now had a spy in their pocket.

Just like Xinjiang, Hangzhou is rife with cameras—including more than a hundred over a two-mile stretch of road along the West Lake near the city center. But this dense network of sensors is meant to improve residents’ lives as much as control them. They feed data into algorithms that alleviate traffic congestion, monitor food safety, and help escort first responders to a scene more quickly.

Alibaba’s other products and platforms make it easier for city residents to pay their utility bills, take the bus, get a loan, even sue local companies in online court.

these collaborations have turned Hangzhou into the “smartest” of Chinese cities and a model that others around the country are rushing to emulate.

Hangzhou’s camera surveillance network has been credited with finding abducted children. Over the decades, China’s birth limits had spawned a black market in stolen children that gave parents around the country fits of anxiety. Offering AI-powered cameras that could track down a lost son or daughter earned the government immense goodwill, even if only a scattering of families would ever need to rely on them. Other data helped the city manage the flow of tourists at crowded attractions, optimize parking spaces, and design new road networks. More than any other place in China, Hangzhou was the picture-postcard version of the Party’s vision for a new, digital authoritarianism

While researching smart cities, we had come across news reports of an initiative in Little River Street called City Eye. The reports described it as an AI-enabled, modern method of city management by the local branch of the chengguan, an urban management force akin to junior varsity police. We sent in a request to the local chengguan branch to visit and learn more about how the system worked. To our surprise, they said yes.

Known formally as the Urban Administrative and Law Enforcement Bureau, the chengguan are responsible for enforcing rules that govern the everyday running of cities. In practice, they spend most of their time taking care of tasks that police don’t want to bother with: chasing away street peddlers, punishing unauthorized trash dumps, tracking down vandals, and handing out parking tickets. Over the years, the chengguan have become the object of close-to-universal loathing across China, and not without reason. Chinese social media is crawling with videos that show vicious attacks by chengguan on street vendors, frequently among the poorest and weakest members of Chinese society.

In neighborhoods, where people are either on the verge of moving up into the middle class or trying hard not to fall out of it—where well-being feels widespread but fleeting. It’s these in-between places, where laundry dangling between apartment windows obscures peeling paint behind billowing rainbows of T-shirts and underwear, that worry the Party the most. The rich don’t have an incentive to make trouble, and the destitute don’t have the power. The people in the middle have just enough of both. And the pressures they face trying to make their way in modern China—merciless work hours, bad health care, constantly rising prices, pollution and food-safety scares, a capricious stock market—make them more likely to lash out.

As a result, local officials often have to tread a fine line, exerting enough force to keep the streets in order and the economy growing, but not so much that they trigger an unpleasant reaction.

Like many Chinese across the country, the people in Hangzhou shared the same disdain for chengguan as everyone else, Qiu said. “Residents only see the street management part of the job,” he complained. Even in the wealthiest Chinese cities, sidewalks are vulnerable to appropriation by the messy, insistent, and occasionally desperate forces of developing-world economics. Farmers selling fruit from the backs of beat-up mini- vans. Migrant vendors hocking socks, cheap knockoff handbags, or the region’s famous stinky tofu out of unlicensed carts. Beggars blocking foot traffic with stories of personal tragedy scrawled on scraps of scavenged cardboard. Qiu and his team were responsible for keeping Gongshu’s streets free of those phenomena. It was true, they sometimes had to get physical, he said, but that was only in extreme cases. No one saw the effort they put in behind the scenes trying to reason with violators.

Surveillance cameras with AI technology kept a 24-hour watch on the streets and sent an automatic alert with a screenshot any time it saw something out of order. Among the things it was trained to notice were piles of garbage on the street and mobile street vendors selling their wares on unauthorized corners. Officers would then decide whether it merited a response. The system also relied on public shaming. Once a week, the chengguan gathered data on violations in each of the district’s nine residential areas and published them through a public account on WeChat that anyone could see. Below the data table, Qiu’s team inserted a list of the most common violations in each area—invariably one form of illegal parking or another

The most common was “mis-parked bicycles” (149), followed by “mis-parked vehicles” and “business conducted outside of shops” (29 each), “drying laundry along the street” (10), “illegal advertising” (6), and “rubbish piles” (2). Notably, not a single residential area had recorded an instance of the last item on the list: “abnormal gathering of people,” a category that included protest.

Residents likewise seemed happy. The pedestrian experience in particular had improved, they said. Streets, previously cluttered, were now clean and free from trash. Electric bikes were parked neatly within allocated white lines on the pavement instead of randomly on the sidewalk.

The task of optimizing life in Hangzhou as a whole fell to Alibaba. The company’s tree-lined headquarters, Taobao City, is a full-sized, twenty-first-century futurama located in Yuhang District, covering nearly 3 million square feet on the northern edge of Hangzhou. Autonomous robots zoom around the grounds delivering packages and groceries, and cameras at the cafeteria payment kiosks automatically recognize what dishes employees have taken from the buffet, charging them accordingly.

At a “hotel of the future,” guests can check in, access the elevators, and unlock their rooms with just their faces. Down the street, a mall for employees features virtual reality mirrors that allow shoppers to try on outfits without having to bother with dressing rooms.

Alibaba trained AI algorithms that could spot traffic accidents and violations and push alerts to apps on the phones of traffic police within twenty seconds. It developed another algorithm to automatically control traffic lights to optimize the flow of cars. Two years later, Alibaba’s data showed that notoriously car-clogged Hangzhou had fallen from fifth place on the country’s list of most-congested cities to 57^th.

One feature of the system is an AI-powered navigation tool provided to ambulances that manipulates traffic lights to clear a path through traffic.

The ambition of Chinese cities to make their residents’ lives easier is real. Local governments and companies spent $24 billion on smart city technology in China in 2020, a figure that was likely to rise to about $40 billion by the end of 2024. China’s state planners are pushing for smart city systems to soak up data from a more diverse web of sensors: not just cameras and smartphones but also QR code readers, point-of-sale machines, air quality monitors, and radio frequency identification chips used to store biometric information in advanced ID cards.

As next-generation 5G mobile internet networks become more widespread, the list of potential data sources is likely to expand exponentially. Connected devices ranging from surveillance drones to smart toasters are set to produce new streams of data for city governments to tap. If everything goes as planned, officials will be able to crunch the information from those censors to revolutionize everything from sewage treatment and health care to crowd control.

China, for all of its associations with authoritarianism, has done more than any other country to democratize state surveillance around the globe. It has accomplished this by helping its companies sell useful and affordable digital tracking systems to governments and by offering to train local police in how to get the most out of the technology. As a result, almost any regime can now get their hands on advanced surveillance tools to do with as they see fit.

Chinese surveillance systems have expanded across the globe with breathtaking speed. As of 2019, Huawei said it had installed Safe City systems in 700 cities across more than 100 countries and regions. ZTE claimed to have built similar systems in 160 cities spread across 45 countries

In the same way that the United States launched the Marshall Plan to rebuild war-ravaged Western Europe as a bulwark against Soviet expansion in the 1940s, Xi wants to use the Belt and Road Initiative to rebuild the global trading system in ways that protect China’s interests. One part of the trillion-dollar plan involves creating what Beijing describes as a “Digital Silk Road.” By seeding other countries with Chinese technology in three areas—communications infrastructure, cross-border e-commerce, and smart (or safe) cities—the Communist Party hopes to rewire cyberspace according to its own vision.

Since the launch of the Belt and Road Initiative, Chinese diplomats have morphed into full-blown salespeople for built-by-China information technology projects abroad, brokering deals with the promise of easy financing that governments find difficult to reject. As of 2019, such projects accounted for an estimated $79 billion in investments around the world. Most of that money has gone into infrastructure like internet cables and 4G base stations, but surveillance has made up an ever-larger slice of the pie.

Like many people in the developing world, most Ugandans couldn’t afford the luxury of worrying about abstractions like privacy. “They say, ‘You know, I’m not a criminal.’ Why should they care?” Mukasa said. But for opposition politicians who only used to have to worry about being physically tailed, the cameras meant they now also had to be careful not to linger too long in public. Ugandan security forces fanned out across the country, using data from license plate readers and facial recognition—all connected through Huawei’s networks—to identify and detain protesters caught on the Chinese company’s cameras. A week after the protests kicked off, Ugandan police delivered a final tally of 45 killed and 836 arrested. Even before Museveni deployed his new system to crush the protests, his partnership with Huawei had already helped focus Western attention on the rapid spread of Chinese surveillance exports.

Neither China’s economic nor its political model is well suited for export. China’s miraculous growth in the post-Mao era was the product of a rare blend of factors that few countries could replicate: a massive and compliant population, strong leadership, and a generous dose of good timing. China’s success with surveillance is likewise the result of a concurrence of advantages—a large and disciplined bureaucracy, massive stores of data, and deep financial resources—that few other countries can match.

Chinese citizens have been issued with detailed identification cards since the mid-1980s. In Africa and other parts of the developing world, public records are poor, and in nations like Sierra Leone, complete records may not even exist. To a country with limited resources, successfully copying the China model was no less an improbability than replicating the American one.

In promoting the China Solution and exporting surveillance, the Party was trying to undermine the notion that democracy is the only legitimate form of government to which a country can aspire. its ultimate aim was “to make the world safe for autocracy” and, by extension, itself.

While Chinese surveillance systems generally failed to reduce crime, the scholars discovered, they were associated with a significant increase in human rights abuses.

By the time Huawei was pushing the expansion of the Chinese surveillance state abroad, American tech companies, investors, and academics had already played vital roles in nurturing it inside China’s borders. Many of them profited from it, some handsomely.

China was then already buying close to $20 billion a year in telecommunications equipment from foreign firms, according to Walton’s calculations. Sun Microsystems, one of Silicon Valley’s original computer and software success stories, helped the Ministry of Public Security build a national fingerprint database. Nortel, one of North America’s dominant makers of networking gear at the time, provided state-owned Shanghai Telecom with state-of-the-art equipment that allowed the Chinese company to filter out unwanted URLs at the point where individual subscribers accessed the internet, allowing for fine-tuned censorship.

The globalization of finance and the growth of venture capital allows American corporations to earn revenue from foreign markets without setting foot on foreign soil. Both developments have made it possible for American businesses and investors to reap rewards from assisting oppressive regimes while keeping their hands relatively clean. The result has been a massive expansion in efforts to do just that.

The neural networks required for deep learning needed immense amounts of processing power. Chipmakers, particularly manufacturers of graphics processing units or GPUs, originally designed for gaming, proved an ideal fit for deep-learning applications like facial recognition. With no formidable Chinese competitors, Santa Clara–based Nvidia was selling vast quantities of chips to surveillance companies all over the country

The database might also be connected to experiments Chinese forensic scientists have been doing since the late 2010s using a process known as DNA phenotyping, which can extrapolate biographical traits such as hair and eye color, age, and skin pigmentation from a genetic sample. The technology, which has also been deployed in some instances in the United States, theoretically allows police to reconstruct the image of a person’s face based on a blood sample.

To help them keep out new arrivals, border agents built a registry and ordered all Chinese residents to carry identification papers with their name, age, occupation, last place of residence, and “physical marks and peculiarities. In turn, the Chinese community launched a decades-long effort to undermine the system using fake paperwork that became, in Parenti’s description, “the largest informal anti-surveillance movement in United States history.

 

Conversations with the people he met persuaded him that concern with privacy as a right in China was the province of the wealthy and well-educated. Residents in other parts of the country were too concerned with making ends meet to afford that luxury. Moreover, the proliferation of cameras, not just in public but in people’s homes and pockets, had made the act of watching and being watched a regular part of life. “For people on the lower rungs, surveillance, streaming video, it’s just a part of how they interact with the world,” he said.

 

The Truman Show, the 1998 science fiction film starring Jim Carrey, about a man who is adopted and raised by an entertainment company inside a simulated world crawling with cameras, was an underground hit in China in the early 2000s, disseminated around the country on black-market video compact discs. Two decades later, it owned a rating of 9.3 (out of 10) on Douban, China’s equivalent of Rotten Tomatoes, with close to a million 5-star reviews. Tech companies had created a version of The Truman Show on a society-wide scale, Xu said.

 

“The world has become a gigantic film studio. Surveillance cameras everywhere, live-streaming twenty-four hours a day,” he said. “The performances of the people in the feeds are incomparable—more convincing than you could get from any actor.

 

By 2020, more than 60% of Chinese people believed facial recognition technology in China was being abused, according to a survey by a newspaper based in the southern city of Guangzhou. Roughly a third of the 20,000 people surveyed reported privacy and property losses because of the leak and abuse of their facial data.

In late 2020, Hangzhou drafted rules to ban residential compounds from requiring face scans to access shared areas. The following summer, China’s highest court issued a judicial interpretation that made clear that facial information was considered personal information. Hotels, shopping malls, and other businesses needed to get consent before using facial recognition, the court said, only in limited ways.

What was going on? The Party, for all its power, still had to respond to changes in public opinion. It was also growing concerned about its relationship to the country’s tech giants. The biggest was a sprawling law on the protection of personal information, which passed in the fall of 2021. It was modeled on the world’s strictest digital privacy legislation, the European Union’s General Data Protection Regulation. The GDPR was vehemently opposed by Silicon Valley when it was approved in 2016 because of the threat it posed to mass data collection. It required internet companies to get explicit consent to collect data, collect only as much as needed, limit the amount of time personal information is stored, and allow users to access, correct, and delete that information at will.

 

China’s version contained all those same provisions, but it differed from Europe’s legal regime in one vital respect: it placed few limits on data the state could access.  The law allowed citizens to sue government agencies for mishandling their personal data, but in a country where the ruling party controlled the courts, that was not a meaningful restraint. Besides improving security, throwing its support behind privacy also gave the Party an important weapon in constraining the power of the country’s internet giants.

 

Alibaba and Tencent, but tensions had grown. The companies often acted with impunity, shrugging off regulators to pursue new lines of business, such as lending, that officials considered risky. More importantly, the companies’ swelling stores of data gave them far better insight into what was happening in many areas of society and the economy than the Party had—a power imbalance that didn’t sit well inside the leadership compound in Beijing.

 

Chinese police and state security agents could request data on individuals and small groups, but the companies pushed back against the mass access needed for the type of sweeping analysis and social modeling the Party wanted to do.

 

Conflict spilled out into the open in November 2020, as Ant Financial was preparing to list in the United States in what was expected to be the largest initial public offering in history. Investors had already committed to buying $34 billion of the company’s shares. But with days to go, Xi Jinping shocked everyone by ordering Ant to cancel the IPO. The cancellation was in part a response to a recent speech Jack Ma had given in which he had been dismissive of regulators, but there was another motivation: not long after, Chinese regulators began pressuring the company to share its troves of consumer credit data. Later, in a further message to Jack Ma to fall in line, they slapped Alibaba with a record $2.8 billion fine for “abusing its market dominance.

 

The assault on Ant and Alibaba was just the first move in a sweeping crackdown on China’s tech sector. Using a combination of antitrust and privacy rules, the Party asserted its dominance over the country’s most powerful companies. A year after the Covid-19 pandemic slowed global commerce to a crawl, China’s economy was in a fragile state. That didn’t stop Xi Jinping from making his move. Ride-hailing, delivery, and social media giants—all big drivers of consumption and employment—came under burning scrutiny over their business practices, including how they handled user data. The rectification shaved hundreds of billions of dollars off the value of Chinese tech stocks as investors panicked, but leaders in Beijing didn’t blink. By the end of 2021, there was zero doubt left about who ultimately controlled the country’s data.

 

Part of it was practical. “If you’re not willing to have your data collected, you basically can’t live,” he said. He still believed that privacy was a concern mostly among the well-educated. “There’s a difference in how higher-tier and lower-tier cities see privacy. It wasn’t that lower-tier China didn’t understand or value privacy but that economic pressures focused its attention in other places. Even if privacy consciousness did spread nationwide, Xu said, he believed it would look different from anything in the West. The continuing influence of collectivist ideas in China would produce a different notion of privacy,

 

Public group activities like dancing and singing are rare in the U.S. but a daily feature of virtually every park in every city in China.

 

The state looms over almost everything in China, but Chinese discussions of privacy seldom target the government directly. Whether out of resignation, acquiescence, or agreement, most Chinese don’t deny the government’s claim to their personal data, even as some rage against privacy violations committed by others. Given the malleability of privacy, that could always change. For the moment, however, the Party has found a way to make China’s version of privacy compatible with surveillance.

 

China’s network of digital sensors is itself often described as “all-seeing.” Without a doubt it sees more than any state surveillance network that has come before. And yet, it’s riddled with blind spots. The overwhelming majority of Chinese people can still find refuge from the Party’s prying eyes inside their own homes, Broad stretches of public space likewise fall outside of the government’s field of vision, whether because of neglect, indifference, or broken cameras. Gaps exist in Chinese cyberspace as well, a result of missing or corrupt data and jumbled, incompatible databases.

 

Personal cars, remote stretches of desert, and crowded markets all offered one degree or another of space for snatches of candid conversation and exchanges of information that the authorities would rather not happen.

 

In the Mao years, the Chinese leaders had funneled precious resources into literal Potemkin villages, like the community of Dazhai in the dusty inland province of Shanxi, meant to embody the intent, if not the actual reality, of policies dreamed up in Beijing.

 

In the reform era, the Communist Party embraced more bottom-up policymaking but never fully relinquished its fondness for central planning. Accordingly, the safest path to a promotion for any official was to respond (or appear to respond) with extreme zeal to dictates from Beijing. When the order of the day was GDP growth at all costs, local governments fell over themselves to erect hastily planned bridges and residential towers—some of which themselves later fell over or apart.

 

It was inevitable that lower-level officials would jump into digital surveillance with the same fake-it-till-you-make-it enthusiasm.

 

Where Potemkin erected elaborate facades, Sadowski writes, Silicon Valley marketers deploy technological buzzwords “as though they were a magician’s incantations: Smart! Intelligent! Automated! Cognitive computing! Deep learning! Abracadabra! Alakazam!” Enchanted audiences were none the wiser.

 

The prisoners only had to believe that they were being watched.

 

The United States took 150 years to build up its credit system. In capitalist countries, owners of capital have a strong incentive to find ways to sort would-be borrowers according to their ability to repay debts. In the 1950s, specialized credit-rating agencies in the United States trawled newspapers for “lifestyle” information to include in individual credit files, including drinking habits and sexual orientation, and helped revive what had been a long-dormant privacy movement.

 

After decades of unbridled growth, Chinese society was dripping with cynicism and dismissive of the government. A few years earlier, batches of baby formula tainted with the industrial chemical melamine killed six infants and sickened as many as 300,000 more. Memories were also fresh of the catastrophic high-speed train crash that summer in Wenzhou that led to accusations of a government cover-up and became a metaphor for the country’s headlong pursuit of development no matter the human cost.

 

The Party still wanted a financial credit system that would unlock economic growth, according to Lin, but to that it had added a much grander ambition: to rebuild a moral structure that had collapsed in the post-Mao era.

 

“China doesn’t have churches to enforce morality. You can’t rely just on education. If all you do is lecture people, saying ‘This is how you should do it,’ or ‘This way is immoral,’ you’re not going to get good results,” he said. Using a mechanism like social credit that offers real-world rewards and real-world punishments “makes the moral construction process a little more efficient.” It was a solution taken from the business world to manage a society in which making money had become the last unifying pursuit.

 

Under Mao, the Communist Party had compelled people in urban areas to behave correctly by compiling secret dossiers on every individual. Known as dang’an, the files contained records of each person’s family background, education level, political activities, job history, and personal achievements and failings. In the pre-reform era, when the state doled out housing and jobs, the dang’an had a terrifying power to determine a person’s path through life.

 

The Social Credit System would need to touch on every area of business. It offered a dizzying list: exports, exhibitions, health care, culture, education, sports, scientific research and stock trading, criminal prosecution and legal defense, pollution control, public security, intellectual property protection, tourism, media, insurance sales, even veterinary medicine. The aim, according to the plan, was to “allow the trustworthy to roam everywhere while making it hard for the discredited to take a single step. In order to do this, the government would need to collect, organize, and analyze unheard-of quantities of data.

 

The blacklist system was established in 2013 by China’s highest court as a way to help lower courts enforce rulings. In addition to having their names listed on the court website, people who ignored court orders could have their assets frozen and be prevented from taking out loans. Over time, the list of penalties expanded to include bans on air and high-speed train travel as well as stays in luxury hotels.

 

The lists were linked together in a system of “unified rewards and punishments” that copied names from one list to the others. This meant someone who stirred things up online wouldn’t just have their internet access limited by the cyberspace authorities but might also be banned from taking out a mortgage by the banking regulator. Shame was a feature of the system as well. In some cities, mobile operators fixed it so that anyone calling a blacklisted person would hear a warning that they were about to talk to someone untrustworthy. The threat of blacklisting turned out to be effective in scaring businesses into respecting rules they had previously ignored, or not even bothered to read in the first place.

 

Official newspapers and TV stations for the most part continued to produce a steady stream of social credit propaganda, running story after story on new bad behaviors that had supposedly come under the social credit umbrella. By 2021, the list had come to include, among other things, running red lights, evading subway fares, exceeding birth limits, defrauding social welfare systems, cheating on tests, plagiarizing academic work, failing to send children to school, backing out of joining the military, involvement in sports doping, refusing to leave the hospital when released, taking someone’s seat on the high-speed train, and being insufficiently filial.

 

They also revealed a system committed to delivering the impression, if not the reality, of omnipotence.

Whether or not the Social Credit System eventually cuts its way through the knot of technological, bureaucratic, and legal obstacles to achieve the immense powers Lin ascribes to it ultimately may not matter. More important to the Party is the ability to persuade people that it already has, putting the idea of surveillance in the back of people’s minds no matter where they go, compelling them to behave as if they were all the while being watched and judged.

 

HOW CHINA HANDLED COVID WITH SURVEILLANCE TECHNOLOGY

 

News of the lockdown exploded across the country. Never before had any government attempted to quarantine a city of that size in an effort to stop a disease from spreading. Convinced that the situation was worse than local authorities let on, hordes of people rushed to leave the city before police cut off the exits at 10:00 A.M. the following morning. Once out, many of them mingled with the hundreds of millions of other Chinese people traveling to their hometowns for the holiday—the perfect scenario for a public health catastrophe.

 

Unfettered access to SIM card location data through state-run carriers and rules requiring every SIM card to be linked to the user’s government ID meant the government could uncover the location of any mobile user in the country at any time. Quickly, China Mobile and the other carriers started to compile reports on users who had been in Wuhan, tracked where they went, and sent names to local governments. Matching that information with data sifted from airlines, the national rail operator, and tourist agencies, they were able to further narrow their target lists.

 

Michel Foucault prefaced his introduction of Jeremy Bentham’s circular prison with a grisly traipse through the landscape of seventeenth-century European plague quarantines. He described how each street in a stricken town was overseen by an official, who locked each family inside its house from the outside and periodically ordered them to appear before a window to be counted. Militia guarded the gates and sentinels prowled the neighborhoods. Anyone who flouted the rules, whether resident or official, did so on penalty of death.

 

Lists were made of the name, sex, and age of each inhabitant, and notes were kept of every death, illness, or irregularity, he wrote. The incessant observation, recording, and supervision represented the height of society’s power to impose discipline:

 

With the virus spreading around the country, the Party had reached back into its Mao-era playbook and deployed thousands of neighborhood watchdogs. While in Mao’s day they had been deployed to carry out Communist Party mandates, resolve disputes, and maintain grassroots social order, now they would go door-to-door checking on residents, taking temperatures, and enforcing lockdowns.

 

Cities began shutting off small side streets and side entrances of apartment complexes, funneling residents through main roads and gates. At some villages around Wencheng, village leaders rolled logs across road entrances to prevent anyone from leaving or entering. Volunteers and village officials—contactless thermometers in hand—would staff the new checkpoints, questioning each person who came by and verifying their movement permits. At its height, the quarantine covered at least 20 provinces and regions. The Wall Street Journal conservatively estimated that China’s government had at one time or another confined more than 500 million mostly healthy people to their homes in January and February 2020.

 

The virus was so infectious it moved from group to group before data could be gathered and analyzed. Vivian’s experience was a good example. It had taken the authorities almost two weeks to find her in Wencheng. Had she been infected, she could have passed the virus on to dozens or even hundreds of others in that time.

 

In Xi Jinping’s hard-line China, local officials afraid of Beijing’s wrath often try to bury negative news in the hope that it will stay unnoticed long enough for them to move on to another job. In that crucial early period, when Wuhan could have used the country’s new digital tools to corral and eradicate a major threat to the country as the Party envisioned, the realities of Party politics incentivized officials to sit on their hands.

 

A drone company in southern China, Shenzhen Smart Drone UAV Co., started offering the use of its high-end temperature-detecting drones to local police. The company’s deputy general manager, Kellen Tse, told us that the drones were typically used to spot forest fires. Now, police were flying them over crowds to check for people running fevers.

 

As the infectious disease began to ravage Western nations and China steamrolled ahead with its recovery, Liu found, a second wave of popular support rose around the belief that democracy was fundamentally flawed and couldn’t cope with public health crises.

 

SURVEILLANCE IS EXPANDING ALL OVER THE WORLD

 

Growing numbers of governments are following the Party’s lead in using digital surveillance to solve social and political challenges. In 2020, India, the world’s largest democracy, deployed facial recognition systems developed by a local start-up to identify protesters marching against a new citizenship law in its capital city, New Delhi. Singapore, one of Asia’s wealthiest countries and a regional partner of the United States, announced that it aimed to install more than 200,000 police cameras as soon as 2030 despite low crime rates and a total land area smaller than New York City.

 

The Party is brimming with confidence in its ability to disrupt and rearrange the global order. That certainty spilled across the front page of The People’s Daily one day in late September in an editorial that exulted in the spectacle of the Communist Party’s one-hundredth-anniversary celebrations in Tiananmen Square months earlier. “The world today is undergoing profound changes unseen in a century,” it proclaimed. “One of the most striking features of these changes is the irreversible trend that the East is rising and the West is declining.” China’s leaders believe they have arrived at a new form of human civilization—one based on a strong state with surveillance at its core—that is more efficient, stable, and responsive than democracy. They also sense that the time is ripe to sell its virtues.

 

In 2019, the White House began adding Chinese companies involved in state surveillance to the Commerce Department’s Entity List, which allowed the United States to cut off supplies of advanced American-designed chips, like the GPUs needed to run high-end facial recognition systems, that China wasn’t capable of replicating. The State Department followed by slapping sanctions on senior officials in Xinjiang, including Party boss Chen Quanguo, that banned them from entering the United States or using the American financial system. Washington also sanctioned the Bingtuan, the paramilitary organization that runs much of Xinjiang’s economy.

 

Under the Biden administration, the United States has stepped up the pressure further by leading a return to something resembling the post-Tiananmen moment. In 2021, the United States organized a coalition that included Canada, the United Kingdom, and the European Union in imposing additional, coordinated sanctions on Bingtuan and Xinjiang officials—the first time that Britain or Europe had imposed sanctions on China in twenty-two years.

 

In her 2020 book Twilight of Democracy, historian Anne Applebaum cites research that suggests a significant number of people in every society harbor an “authoritarian predisposition” that is repelled by diversity and disorder. “Authoritarianism appeals, simply, to people who cannot tolerate complexity”.  Wealthy members of the racial majority will almost always experience state surveillance as a source of security and convenience, having no incentive to imagine its side effects on marginal members of society until, whether by choice or circumstance, they find themselves living on the margins.

 

 

China’s one-child policy

 

the Club of Rome, a group of scientists alarmed by the growing human population, which they feared would deplete the earth’s resources to the point of undermining economic progress. The group had developed an elaborate mathematical model that produced scenarios for managing the population and distributing resources in what it believed was a scientifically optimal manner. The approach was criticized in the West for treating populations of human beings as if they were cohorts of fruit flies in a lab. Song found it an easy sell back in Beijing, where leaders had already begun to worry that the country’s population was growing out of control.

 

While accounts differ over precisely how much of China’s one-child policy is attributable to Song, Qian’s protégé indisputably helped drive its design. Song and his team submitted projections to Party leaders showing that without intervention, China’s population would swell to more than 4 billion—a size they predicted would result in the evisceration of resources and the environment—within the next century. They calculated the country’s ideal population at around 700 million and found that getting there by 2080 would require bringing the fertility rate for women down close to a single child as quickly as possible.

Xinjiang, Uyghur oppression

If one corner of China comes close to justifying the “all-seeing” label, it’s Xinjiang, where security forces have cultivated a tried-and-true ability to peer into every aspect of Uyghurs’ lives. Perpetually tracking 15 million members of a minority group spread over a vast area turned out to be difficult, even with the best surveillance systems China could muster.

In Xinjiang on the country’s remote western edge, at the doorstep of Central Asia, the police were hard and well armed. Any interaction with them unfolded in an aura of peril and potential violence.

They took Tahir Hamut’s blood first. Next came his voice- and fingerprints. They saved his face for last.

Tahir had assumed their time had come to be “educated.” It was hard to believe they had been able to simply walk out. Others were less fortunate. Over the next few weeks, Tahir and Marhaba watched as their neighborhood slowly emptied out. One day, Tahir wandered outside and realized the smell of the tandoor naan flatbread, a warm swirl of yeast and sesame that conjured a thousand memories, had suddenly disappeared. The young men who ran the ovens were nowhere to be found. Soon other young men—butchers, fruit vendors, drivers—began to vanish. Then middle-aged men and some women.

It didn’t take long for Uyghurs in Urumqi to draw connections between the spreading surveillance, the biometric data collection, and the disappearances. A proliferation of new security gates made it impossible for residents to move around the city without having their ID cards and faces scanned. Some Uyghurs’ cards or faces set off alarms, which led to them being shuffled off to police stations.

If this new system decided he was a threat to the social order, he wouldn’t have anywhere to hide.

Over the Communist Party’s seven decades in power, Xinjiang has been China’s most fractious region, riven by ethnic tensions between Uyghurs and Han Chinese migrants that have periodically exploded into deadly violence. Against the odds, the Party has brought the territory under total control using a combination of internment camps, brainwashing, and mass surveillance.

All of this was linked to biometric markers like those the Urumqi police collected from Tahir and Marhaba during their visit to the basement.

How the platform processed this ocean of data was a black box, but one of its most important results was a ranking of people according to the level of threat they posed to social order. In 2019, the International Consortium of Investigative Journalists obtained an internal Xinjiang police notice that showed that the platform generated a roster of more than 24,000 “suspicious persons” that ranged from unfair to capricious. Among those: collecting money for mosques “with enthusiasm,” moving into a locale, moving out of a locale, returning home after being away a long time, returning from abroad, being connected to someone abroad, and “for no apparent reason, being unwilling to enjoy policies that benefit the people or failing to participate in activities organized by the government or the Party.

People on the higher rungs, such as those who had been jailed for participating in the 2009 Urumqi riots, elicited a red alert and were to be immediately taken into custody.

Even in regions where extremist ideas are popular, only a tiny proportion of the population—a fraction of 1 percent—cross the line into outright terrorism.

Officials in Xinjiang began building dedicated “transformation through education” centers to reform Muslims they saw as especially fervent.

Among the 24,000 people that the ICIJ police notice said were flagged as suspicious during that one week in the summer of 2017, two-thirds were detained by police. A few hundred were formally arrested and sent to prison. The other 15,000 were shipped off to one of Chen’s new facilities

A cache of procurement documents collected by the French newswire Agence France-Press offered a hint to what they were like inside. The documents, copied by an AFP reporter just before government censors began mass-deleting them, requested haunting lists of equipment. One camp in Hotan, in southern Xinjiang, asked for 2,768 police batons, 550 electric cattle prods, 1,367 pairs of handcuffs, and 2,792 cans of pepper spray. Others sought suppliers for infrared monitoring systems, razor wire, phone surveillance systems, riot shields, helmets, tear gas, net guns, stun guns, spears, billy clubs, spiked clubs called “wolf’s teeth,” and tiger chairs

One former detainee described being bound to a chair for up to nine hours at a time and interrogated about links to religious groups abroad. Afterward, he and other inmates would be roused at 5:00 A.M. and forced to go on forty-five-minute runs, shouting, “The Communist Party is good!” Breakfast was bread and barely flavored soup. The rest of the day was taken up by political study, which involved reading Communist Party documents, watching videos of Xi Jinping, and singing patriotic songs. They were told to not pray or fast during Ramadan.

 

“People didn’t dare to speak even a single word out loud. Everyone was silent, endlessly mute, because we were all afraid of accidentally saying something wrong.

 

satellite images showed that the facility continued to grow in size and severity. The original plans had called for “teaching buildings” totaling 250,000 square feet, the equivalent of five American football fields. Over the next few months, construction crews would add an immense parking lot, dorms, more walls and guard towers, and 135,000 square feet of detention and custody buildings

 

Across Xinjiang, Uyghur towns and Uyghur-dominated neighborhoods in cities took on a ghostly aspect. With each passing week, more men disappeared into the camps. Stores and restaurants began to close, and bazaars that once thrummed with the clamor of haggling shoppers quieted to a murmur.

 

Adrian Zenz, a German researcher with experience digging through Chinese government websites, set about calculating an independent estimate. Based partly on a document leaked to a Japanese news organization, he offered a “speculative” range of between several hundred thousand and just over a million—close to a tenth of Xinjiang’s Muslim population. As more information seeped out of the region in the following years, he would revise it higher.

 

China’s government was savvy enough to know it couldn’t get away with mass murder of a religious minority in the twenty-first century. The stain of such a crime would destroy Beijing’s ambitions to build China into a respected global power. Nor would killing off the Uyghurs fit with the Party’s decades of propaganda touting its concern for the welfare of ethnic minorities.

 

By July, the surveillance state Chen had built now hovered over Uyghurs like an apparition. At any moment it could materialize out of the dark and drag people away into the unknown. Police had set up desks on sidewalks where they would order pedestrians to hand over their smartphones, then plug them into scanning devices called “Anti-Terrorism Swords” that searched for more than 53,000 identifiers of Islamic or political activity. Anyone whose phone had a copy of the Koran, encrypted chat apps like WhatsApp, or a virtual private network for getting around the country’s internet filters would be taken in for interrogation. Even photos of Turkish film characters were enough to get someone dragged into a police station.

 

Homes weren’t much safer. Police had started making daily visits to some families to scan their QR codes and search for unapproved guests. Sometimes they ran through the rooms, examining bookshelves and rummaging through cupboards in a search for religious material. People gave away or hid their Korans to avoid trouble. Marhaba grew paranoid that the QR code had a hidden listening device. Tahir couldn’t persuade her otherwise, even though it was printed on paper. And even if he was right, he couldn’t say for sure there wasn’t a microphone lurking somewhere else.

 

Based on testimonies gathered by a nonprofit group in Kazakhstan, media reports of emptied streets and bazaars in Muslim districts, and the continued expansion of the camps, Zenz said he believed as many as 1.5 million people—roughly one out of every six adult Turkic Muslims—had been detained in one of the facilities at one point or another.

 

The camps were being kept full by the continued expansion of Chen Quanguo’s web of surveillance. Security gates, most equipped with facial recognition, had been installed in virtually every public venue: at the entrances to bazaars and banks, mosques, hotels, tourist sites, bus stations, train stations, fancy malls, and cut-rate department stores. The police purchased drones, also loaded with facial recognition technology, to track people in remote areas outside the reach of regular cameras.

 

Toward the end of 2019, a few months after the United States imposed its first round of sanctions, Chen Quanguo began releasing camp detainees. Life on the streets of Xinjiang’s cities began to regain some semblance of normalcy, with fewer police patrols and less razor wire.

 

But fundamentally, little has changed. Mosques remain closed or closely watched, and calls to prayer stay silenced. Uyghurs and other Turkic minorities still live inside a web of inescapable surveillance, afraid they could be dragged away at any time. Though some camps have been closed, many of the remaining facilities have been expanded or converted into prisons. Calculating the floor space of roughly 350 prisons and internment camps in Xinjiang based on satellite images and Chinese prison construction standards, Buzzfeed News concluded in 2021 that the region maintained the capacity to detain at least a million people at any given time—seven times higher than the criminal detention capacity of the United States.

 

Researchers uncovered a policy document posted on the website of the Ministry of Education showing that nearly half a million children—40% of the region’s pre-college students—were enrolled in boarding schools. It called for further expansion of boarding schools so that education officials could “to the greatest extent possible break the influence on children of the religious atmosphere at home.” The current generation of Uyghurs might be able to resist the brainwashing, but children were impressionable. Suddenly the Party’s assimilation plans began to seem more plausible.

 

In an echo of Nazi Germany in the early days of the Holocaust, the Chinese government had also invited media from friendly countries to visit select camps on tightly controlled tours. Uyghur detainees, under the watch of authorities, gave interviews praising the education they had received.

 

Uyghur women came out of the camps with stories of sexual assault and gang rape. Several reported being given injections or forced to take pills that stopped their menstrual cycles. The women also revealed that family planning officials in Xinjiang had begun strictly imposing birth limits in Uyghur communities—even as they desperately encouraged Han Chinese women to have more children to counter the aging of the country’s population elsewhere. Having three or more children was a common reason for Uyghur women to be sent to the camps, where many were forced to have intrauterine devices inserted to prevent them getting pregnant.

 

Combined with the policy of shipping Uyghur children off to boarding schools, the implication was clear: the Party was hanging its assimilation strategy on the next generation of Uyghurs, who would not only be educated to be Chinese but also be fewer in number.

 

Meanwhile, the surveillance technologies and some of the techniques the Party had deployed in Xinjiang were spreading to other minority-dominated regions in remote parts of the country. Systems for monitoring mosques migrated to Gansu and Ningxia, northwestern regions home to large numbers of Hui, another Muslim minority culturally and politically closer to the Han Chinese. In heavily Tibetan areas of western Sichuan province, the government installed facial recognition systems and a total of more than 180 surveillance cameras inside seven Buddhist monasteries. Government procurement documents showed that they also used drones to map some monasteries and built databases to categorize monastery residents.

 

U.S. surveillance

 

Based on dozens of public-records requests, it found that 16 states had granted the FBI access to their databases of driver’s license photos for use in facial recognition scans. They also found more than two dozen states that allowed police to conduct facial recognition searches on driver’s license databases. By the middle of 2016, they calculated, at least 117 million Americans over the age of 18—more than 50% of the adult population—had been placed in a virtual, never-ending police lineup, often without their knowledge.

 

Unlike their Chinese counterparts, who are happy to trumpet their new surveillance tools, American police tend to shroud their arsenals in secrecy. The more powerful the tool is, the more care they often take to shield it from view. The transparency built into the American system in some instances makes it harder to see how the police are watching.  The algorithmic tools now at the disposal of the NYPD and other police departments had reduced the time it takes officers to identify someone from days or weeks to a matter of hours or even minutes.

A combination of cost reductions and expanding budgets has given police departments across the United States access to a long list of what used to be considered military-grade spy tech tools, all of which reduce friction in one way or another: automatic gunshot detection systems that pinpoint the location and type of gunfire; “Stingray” cell-site simulators that can siphon data off mobile phones and intercept, or even alter, communications; social media profiling services that sift through the online activities of a population for early warnings about potential incidents.

The New York City Police Department, America’s largest municipal law enforcement agency and its best-funded at $5 billion a year. In a sense, New York City’s entire surveillance apparatus was an example of mission creep. Five years after 9/11, the NYPD unveiled plans for a $100 million Lower Manhattan Security Initiative, which called for the installation of 3,000 security cameras and 100 automatic license plate readers, along with mobile roadblocks and radiation detectors. A few years later, the NYPD and Department of Homeland Security announced Operation Sentinel, a program that would photograph every vehicle entering and leaving the city along bridges and through tunnels. Sold to the public as counterterror measures, the projects soon became integral to the NYPD’s regular crime-fighting efforts.

One of the system’s more popular features among police is Patternizr, software that sifts through historical data on hundreds of thousands of burglaries, robberies, assaults, and other offenses to help detectives spot patterns in criminal activity across precincts. The platform also includes a secretive predictive policing system, which uses similar data to attempt to identify areas where future crimes are likely to occur.

One of the government’s most active surveillance agencies, Immigration and Customs Enforcement (ICE), has made extensive use of two powerful tools: license plate readers and data fusion products. One of ICE’s partners, Vigilant Solutions, gives immigration officers access to 5 billion license plate detections collected by private businesses, like insurance companies and parking lots, plus another 1.5 billion records gathered by eighty law enforcement agencies in more than a dozen states. Another partner, Thomson Reuters, provides ICE with a massive cross-referenced collection of public and private data on individuals. Combined, the two companies allow ICE to run a “continuous monitoring and alert system” that tracks the life events and movements of 500,000 individuals per month, the overwhelming majority of them members of the Hispanic community.

Clearview AI, the company paired facial recognition with a database of more than 3 billion images scraped from Facebook, YouTube, Twitter, Google, Venmo, and countless other public websites spread across the internet—a facial image database seven times larger than the FBI’s. The New York Times analyzed the company’s app and found code that would allow it to be paired with augmented-reality glasses, potentially allowing someone to pull up the social media history of anyone they saw on the street.