The False Promise of the Surveillance State
In May 2017, I walked through the main gate of the Id Kah Mosque complex on the main square of Kashgar, the Uighur-majority city on China’s far western frontier. The mosque has been the center of Kashgar for hundreds of years, a reminder of the city’s ancient Central Asian heritage. The six video cameras hanging from a metal railing above demonstrated the Chinese Communist Party’s determination to forge a techno-authoritarian modernity. At the time, in 2017, the Party’s escalated campaign of repression against their Uighur Muslim minority was in its early stages. The cameras above my head showed that Beijing was already placing surveillance at the center.
Three years later, China’s nationwide system of facial recognition is one of the Communist Party’s proudest accomplishments. The country hosts several young tech firms valued at over $1 billion that specialize in selling facial recognition software to China’s government and around the globe. Beijing’s worldwide exports of surveillance tech seek to spread its governance methods and bolster authoritarian allies. Foreign visitors entering China have their fingerprints scanned at passport check, after which a computer screen declares “fingerprints captured!” Then their photo is taken and, presumably, uploaded to a facial recognition database. “Face captured!” the computer screen declares, as an animated face smiles happily at you.
Most of the debate over the PRC’s vast surveillance apparatus has focused on whether it is a good thing to have your “face captured” by China’s Communist Party. Beijing’s tech-enabled surveillance state enables the security services to track more easily anyone who dares question the wisdom of Xi Jinping Thought. For China’s Uighurs, who are subjected to batteries of cameras at every turn, and forced to upload tracking apps to their phones, the rollout of surveillance tech has been accompanied by forced labor and re-education camps.
China’s construction of this vast surveillance apparatus is a tremendous achievement, we are frequently told, giving the party new tools, better information, and, as a result, increased longevity. The notion that intensified surveillance improves governance outcomes and stabilizes the regime is the justification for Beijing’s techno-authoritarianism at home. It is also the sales pitch for Chinese tech abroad.
But are these claims true?
Medieval China is said to have had Four Great Inventions—gunpowder, paper, printing, and the compass. Yet despite this technological prowess, the Industrial Revolution happened in Northern England rather than Northern China. Today, Beijing is betting on the Four Great Inventions of Xi Jinping’s China—facial recognition, AI, smartphone tracking apps, and 5G connectivity—to undergird the Party’s rule over the next generation. New technologies have funneled extraordinary volumes of information to the Party’s security apparatus. China’s government probably has more data on its people than any government has accumulated in history.
New surveillance technologies are justified on the grounds that they give governments more useable information, allowing authorities to resolve previously insurmountable dilemmas, and to address longstanding challenges with fewer resources. More data and better tracking ought to allow individualized responses to policy problems. We ought to see governments tackling threats before they arise, and using more targeted interventions in the place of blunt methods. For China, surveillance tech ought to replace Maoist mass mobilization—the political technology of the peasant society—with precision technocracy. If surveillance tech works, in other words, it ought to make China’s government less overt in its strategies of repression. It should also make the Communist Party more capable and responsive, thereby extending its hold on power.
There is no doubt that technology gives autocrats new tools. But does it improve their political skill? In China, the homeland of surveillance tech, the promised benefits have not panned out. There is little evidence that Chinese politics is being transformed by technology. In China, the rise of surveillance tech has coincided with a turn away from technocracy. Except for the smiley animations on its facial recognition software, Beijing has abandoned the pretense of moving toward Singaporean-style soft authoritarianism. During the days of Xi Jinping’s predecessor Hu Jintao, many analysts in China and abroad suggested that the Communist Party had embraced technocracy and looked to Singapore’s Lee Kuan Yew as an example of enlightened governance. Yet under the People’s Leader (as Xi is now officially known), Chinese politics have re-emphasized sticks over carrots. And the construction of the surveillance state has been accompanied by a reinvigoration of Marxist-Leninist ideology and resuscitation of Mao-style mass mobilization campaigns.
Beijing has forged a surveillance state without peer. But information alone provides no ironclad guarantee of the Communist Party’s future. There is some correlation between governments’ ability to procure information and their capacity to govern. The Romans’ road network was built with this aim in mind. The U.S. Constitution empowers Congress “to establish Post Offices and Post Roads” for a similar reason. But bringing information from the provinces to the center only accomplishes so much; the information must be analyzed and deployed, too. The infrastructure needed to understand and to interpret—including transparent institutions and well-aligned incentives—is very different from the equipment needed to collect data. This is why China’s surveillance apparatus has not created the ultra-efficient techno-dystopia predicted by sci-fi novelists: the modern surveillance cameras have simply been fitted into the standard authoritarian dystopia bequeathed by Marx and Lenin. China’s tech ecosystem has blossomed, but Chinese politics are little changed, which is why Chairman Xi is reinvigorating the personality cult of Chairman Mao. Artificial intelligence, it turns out, is no guarantee of intelligent government.Artificial intelligence, it turns out, is no guarantee of intelligent government.
The coronavirus coverup in Wuhan is but the latest evidence that China’s surveillance state is not delivering on its promises. A properly functioning surveillance apparatus would have immediately notified Beijing of the outbreak. Instead, local authorities had early knowledge of the virus but declined to share this news lest they be punished. When the hospitals were too full of people dying of pneumonia to cover up, Beijing was finally notified. Then Chairman Xi chose to deny and delay until the virus had spread across China and the world.
China’s state media have celebrated the country’s tech firms for developing apps to manage the crisis and enforce the quarantines. More significant, though, is the failure of the surveillance apparatus to prevent the pandemic in the first place. Terabytes of data were streaming toward Beijing, volumes unimaginable just years earlier, from surveillance cameras, from smartphones, and from the internet of things. None of this provided any benefit in the crucial early weeks of the COVID-19 outbreak when the virus was allowed to spread from Wuhan to the world.
Many governments have struggled with the new coronavirus, so Beijing’s inept policy response cannot be blamed solely on its surveillance state. But the surveillance apparatus deserves little credit, either. The causes of the COVID crisis—excessive secrecy, distorted incentives, an unaccountable elite, and suppression of criticism—are standard fare in Chinese politics. They explain why in addition to COVID, the Communist Party has also bungled the handling of a swine fever over the past year and a half. These factors were also behind the Party’s failure to promptly tackle SARS in 2002 and AIDS before that.
The Chinese Communist Party’s eventual response to the COVID crisis—a mixture of mass mobilization and denunciation—are standard policy responses in the People’s Republic, whether to epidemics or other threats. In the spheres most important to the Party’s hold on power, there is little evidence that technology is improving policymaking. Consider Beijing’s campaign ostensibly to crack down on violent Uighur separatism. Rather than target individuals linked to violence, Beijing has reportedly locked up hundreds of thousands of Uighurs in re-education camps. The pinpoint accuracy promised by surveillance state advocates ought to enable the government to arrest hundreds rather than hundreds of thousands. Yet instead of careful tracking, the CCP is sorting citizens by ethnicity—a type of “facial recognition” that requires neither smartphones nor software.
Digital facial recognition technology may well provide some benefits to humanity. Owners of the newest iPhones appear generally to like signing in with their faces rather than their thumbprint. When people begin returning to airports after the COVID quarantines subside, they will find it increasingly common to board planes by scanning their faces rather than their boarding passes. But it is one thing to sign onto an iPhone, and quite another to manage a political system. In the world’s epicenter of surveillance tech, the thesis that it will transform politics is being daily disproven.
The main accomplishment of China’s surveillance state has been to reproduce the indignities of authoritarianism in the internet age. Denunciation culture has gone digital. The surveillance apparatus did nothing to stop the COVID epidemic from spreading, yet apps proved capable of denouncing people “infected” with the virus. Singaporean-style soft authoritarianism this is not. Pinpoint accuracy remains a far-off dream.
Will the surveillance state at least provide the political stability that the Communist Party craves? History does not suggest cause for optimism. Authorities tasked with guaranteeing political stability always want more information. All security chiefs find frightening the volume of things they wished they knew but do not. Yet there is no obvious relationship between the size of a surveillance apparatus and the longevity of a regime. Nor is there a straightforward link between the quantity of data flowing into a government and the quality of analysis that comes out.
All governments want information, and regimes that fear for their hold on power tend to want more of it. For centuries, secret police forces have opened letters, sought informants, studied rumors, and employed spies. In the years before the French Revolution, King Louis XVI’s agents tracked his opponents not only across the country, but in London, too. From the late 19th century, the Russian czars’ fearsome Okhrana ran operations across Europe to follow dissidents and would-be assassins, even placing double agents inside revolutionary cells. Even democracies gather reams of information, both on external foes and on purported internal threats. At the height of Cold War-era insecurities about communist subversion, FBI Director J. Edgar Hoover ran a series of illegal information gathering efforts against domestic rivals such as Martin Luther King, Jr.
Some amount of surveillance is needed to investigate run-of-the-mill crime. Information gathering can also protect against violent opponents of the constitutional order. But surveillance is only ever a tool, not a panacea. The failures of secret police agencies are in many ways more spectacular than their successes. They were not enough to help the Bourbons or the Romanovs, who remained blind to the threat even as the revolutionaries were pounding down their palace gates.
And it was not only the old monarchies who wrongly trusted in their secret police. The most technologically advanced surveillance state before Xi’s China was found in East Germany. The Stasi had recruited one in ten Germans to work as informants, and, in addition to old-fashioned letter-opening, employed an advanced system of surveillance technology. Yet this did little for the East German Communists. Their regime’s collapse in 1989 came almost as a complete surprise. They had almost unlimited information, but little understanding of the revolutionary surge that ripped through East German politics in mid-1989, at the speed with which the coronavirus tore through Wuhan in January.
It is far easier to collect data than to analyze it, and easier to recognize faces than to understand the minds behind them. Some parts of human life, like traffic and shopping, follow patterns that are easy to model. But though prophets from Karl Marx to Nate Silver claim to have discovered a science of society, politics rarely follows predictable patterns. Most of the world’s great revolutions came as an absolute surprise. And many of the contemporary world’s most frequently predicted revolutions, from the downfall of Venezuela’s Nicholas Maduro to Iran’s religious leaders, stubbornly refuse to happen.
Because analysis is hard, and because predictions are vulnerable to falsification, surveillance chiefs prefer to devote resources to collecting rather than predicting. A spy chief is rarely criticized for finding new ways to tap phones or hack emails, but they can easily be attacked for forecasts that go wrong. Yet surveillance apparatuses face a problem like the one that economist Friedrich Hayek diagnosed in centrally planned economies: they can collect many types of information, but not the ones that they really need. “Knowledge of the circumstances of which we must make use,” Hayek wrote, “never exists in concentrated or integrated form but solely as the dispersed bits of incomplete and frequently contradictory knowledge which all the separate individuals possess.” The rough political equivalent is the adage that “all politics is local”—something that centralized databases are unlikely to manage well.
Soviet economist Leonid Kantorovich won the Nobel Prize in Economics in 1975 for his work on the algorithms to optimize a planned economy. Yet even when he ran their numbers through the USSR’s most advanced computers, his best efforts never approached the efficiency of the decentralized price mechanism.
The KGB tried to manage Soviet politics with a similar emphasis on centralized information planning. But it faced similar dilemmas as did Gosplan. “Garbage in, garbage out”—the constant problem of the data collection machine. The Soviet security services reported on all sorts of capitalist subversion that didn’t exist, but that they assumed bosses would like to hear about. They neglected to report on the far more important fact that the security services and party elite were themselves losing faith in socialist ideology and the Soviet system. If planning the Soviet Union’s economy was difficult and inefficient, planning its politics proved impossible, as its completely unexpected collapse proved.
The reality is that politics cannot be planned or predicted, regardless of the quantity of data streaming toward the center. Every U.S. President comes into office with a plan, written by the experts that staff their transition teams. Then presidents are forced to toss these plans out as they spend their days in office dealing with unexpected shocks: terrorist attacks, hurricanes, recessions, and pandemics. Authoritarian systems have more need for planning and predicting. Presidents need only to survive four or eight years in power, while dictators must plan for indefinite rule.
Surveillance and information gathering can help solve some problems, but only in a supporting role. South Korea, Taiwan, and Singapore have led the world in tackling COVID thanks in part to technology like tracking apps. The West, too, is likely to adopt new surveillance tech in the aftermath of the epidemic, making use of instant virus testing, smartphone GPS data, and perhaps also facial recognition to manage quarantine and social distancing regulations. But these types of health surveillance will only be as effective as the institutions in which they are embedded. Today, some Republican governors are as uninterested in taking COVID-19 seriously as were Communist Party leaders in Wuhan in January. No amount of virus surveillance data can change this. Those societies that were always on top of the virus—using old-world techniques like face masks and contact tracing—will find that new tools graft straightforwardly onto existing institutions. South Korea has used apps to track the virus, for example, but no one thinks that its politics have been in any way transformed. Technology complements state capacity but it cannot substitute for it.
Advanced surveillance tech will, therefore, prove a false hope for politicians worldwide, as it already has for China’s leaders. They hope that it will provide the information needed to predict politics and shape the future. And yet the most advanced technologies seem rarely to address the question most need answered: what will come next? Nobody knows—and least of all someone convinced that the answer will be found by accumulating ever more information about the present rather than thinking about the future.