Is Big Data Increasing Beijing’s Capacity for Control?

A ChinaFile Conversation

China’s authoritarian government is using big data to develop credit scoring systems, and is urging data-sharing between companies and governments, putting ordinary Chinese squarely in the digital spotlight. How should Chinese netizens and global citizens concerned about privacy react? —The Editors


China’s government is leading the country into an Information Technology-backed authoritarianism. It uses Big Data and cutting-edge IT systems to strengthen its central authority and control capacity. The basis for this will be an all-encompassing system of cyber surveillance. The scope of China’s new so-called Social Credit System will go far beyond financial-credit scoring systems in Western societies by linking different data sets.

A number of private technology companies, including the Internet giants Alibaba, Baidu, and Tencent, are deeply involved in developing this scoring system. They are willing to share the data they collect on their social media, shopping, or even dating platforms with governmental agencies, and they are working on solving the immense technological challenges of integrating those vastly different data sets into one national system.

This will bring the government into a position to significantly expand its knowledge about what people do and think at any given time. It will attempt to influence behavior by offering incentives for good social credit scores and by threatening the exclusion from economic and social activities to those with low scores. For Chinese citizens this should be extremely worrying. Many, though, do not care about or even support this development.

The consequences of China’s IT-backed authoritarianism, however, reach far beyond China.

First, China’s IT-backed authoritarianism has the potential to become an export hit. If China can prove the effectiveness of its social scoring system, it will become an attractive model and provider of cutting-edge technology for other authoritarian governments. This will have significant impact on the power and endurance of non-democratic regimes all over the world.

Second, China’s credit scoring system will not only monitor citizens, but also companies. International companies with China activities may soon need a good social credit score to enter the Chinese market or to continue doing business in China.

International technology companies should refrain from helping China speed up the implementation of this system. Tech companies with expertise on big data and credit scoring should carefully weigh the political risks before providing Chinese partners or investors with their technologies.

Western governments should step up their efforts to integrate China into international data-security and standardization agreements. In the end, concerns over data privacy will not prevent the Chinese government from pursuing its Orwellian plan. But at least we should not make China’s job easier.

In many ways, the social credit system isn’t very new. The Chinese Communist Party government has always sought to keep tabs on its citizens, for instance through the “personal file” (dang’an) system of a few decades ago. Academics and policymakers were discussing the possibility of a social credit scheme far before the advent of widespread Internet and mobile technology. What has changed is that the leadership now seems convinced that technologies such as big data, cloud computing, and the mobile Internet will provide the boost needed to turn the system into reality.

For the moment, I remain somewhat skeptical about the feasibility of the project, at least, as the all-encompassing structure to monitor behavior presented in policy documents. First and foremost, governments everywhere in the world have a history of bungling ambitious information and communications technology (ICT) projects, and the simple scale and complexity of the envisaged Chinese system provides a particularly daunting technical challenge. One element of this is information-gathering. The Shanghai municipal government, for instance, published a catalogue of information points to be entered into the credit system, which listed over 1,200 items. About 1,000 of these concern businesses, while the rest are information concerning individual citizens. Entering, maintaining, and updating these entries is a mammoth task, and every error reduces the efficacy of the system. Moreover, this list is to be supplemented by the vast amounts of data generated in private online business platforms, through different systems, with different modes of information collection, processing and storage, and so on. It will be a considerable feat merely to ensure interoperability. Even if all of this was to work, the next challenge would be to derive meaningful knowledge and insight from the resulting ocean of data. Again, this would be hugely costly in terms of resources, with possible gains uncertain and the potential for error considerable.

But perhaps the biggest challenge is political. At the most basic level, the purported idea of scoring individuals is to reward desirable behavior and punish undesirable behavior. That, in turn, may require some debate about how certain behaviors are to be classified and what weight should be given to each. This question is rather similar to the question how cadres should be evaluated—an infamously thorny problem that equally remains to be resolved.

Nevertheless, I think that the social credit system will generate a considerable number of partial successes. Even if it will not be possible to rate everyone all the time, it is likely that certain categories of professionals, such as doctors or teachers, will come under increasing scrutiny. Even if it is impossible to correlate all kinds of behavior, it is already the case that convicted debtors are barred from making luxury purchases. To me, the interesting question is which analytical framework is most apposite in understanding what happens: is this a classically authoritarian attempt at mind control, or a somewhat more banal Weberian expansion of the bureaucracy?

Could be missing something about this question —is there some place in the world where Big Data is not increasing state control?

“Data” can mean very specific things to people in the tech world, but in society generally people assume that governments always will be able to collect information massively and instantly. People are as intimidated by inaccurate information collection as they are by accurate information collection. It is “data” as an absolute that makes them cautious and suspicious. While in the United States we associate government data collection with passive surveillance and regard the voluntary surrender of huge amounts of personal information to commercial entities as some other kind of thing, in China there generally is no illusion that such a distinction exists. As for social credit, it is already established in the United States. Your FICO score influences how quickly you will get on an airplane (or whether you will at all), and your social media ratings from Facebook or commercial “reputation” sites influence whether you get a job or a promotion. The snares will only increase from here on. The difference is, the Chinese public is already more prepared for the integration of commercial, military, and security information collection. As a consequence, Chinese are better controlled through intimidation and self-censorship than their counterparts in the U.S. will be for a couple of decades.

Though social control through big data has strong parallels in both China and the U.S., there are differences. One is the unwillingness of the Chinese government to outsource completely its surveillance and censoring technologies to the commercial sector—not least because American companies are still critical in China’s big data effort. It is a structural difference with the U.S., where big data companies are full partners with government and occasionally can advance their own priorities at the expense of state control. In addition, China uses its commercial law (and particularly investment thresholds) to exclude tiny tech start-ups from competing with the larger, fully-tamed corporations; in this way “disruptive,” rogue enterprises that are the normal breeding ground for Virtual Private Network (VPN) and proxy schemes that might circumvent the censoring structures are unviable by definition. The institutional strangling of a wide variety of innovation is obvious. More important, big data depends upon scope and speed —you need more and more of both. The Chinese strategy at present is to increase the speed (of intervention and accumulation, not of data exchange) by narrowing the scope, making China a digital island. It can’t work, because the focus on controlling routers and logarithms and even physical signals ultimately will divert attention from understanding data content to controlling the perimeters of data transfer, inhibiting both speed and scope. We already see some of the effects of this in the exit from China of major international data corporations. The state must hope that by the time the data system breaks, the social and political fears of “big data” will be sufficient to accomplish the goals on their own.

Judging the merits of China’s efforts to employ big data to use credit scoring systems as a form of internal surveillance can only be done within the broader context of Beijing’s effort to modernize the internal security apparatus. We are much more accustomed to hearing about “informatization” in the context of military modernization, but, ever since 10,000 Falungong practitioners surprised the leadership by demonstrating in front of Zhongnanhai in April 1999, the internal security services have been researching how to better employ intelligence and surveillance technology to police, shape, and control Chinese society. At the heart of this kind of surveillance is the desire to preempt problems before they begin.

As Rogier Creemers has pointed out, China’s social credit system, is not new—even within the context of technology-enabled surveillance and policing. The Social Credit System has clear antecedents in the broader official Chinese discussion of “public security informatization” in the late 2000s and early 2010s, and the disgraced former security chief Zhou Yongkang arguably is the father. The ideas behind the social credit system were part of a series of articles (since scrubbed from the Internet) in which Zhou is credited as the author describing a “social management system” to monitor happiness, encourage compliance, and shape decisions that could affect social stability. In a social management sense, the concept of monitoring, compliance and shaping stretches back decades, but the technological advances to see these aspirations realized is more recent.

The “grid management system” that caught Western attention in the last two years is one of the best examples of the marriage between traditional C.C.P. governing strategies and modern surveillance technology. The current version of the “grid(-ized)” (网格化) policing concept first emerged, publicly, between 2001 and 2002, particularly in Shanghai, and was based on earlier concepts that designated administrators to monitor a “grid” based on a specified number of people. Pilot “grid” policing programs were implemented in multiple cities throughout the early 2000s. These focused on harnessing modern technology to engage in the prevention and control of social order problems—applying the concept of pre-empting threats to modern technology. This developed by 2005 into “grid management” (网格化管理), a slightly broader social-management oriented concept, which was first openly implemented in Dongcheng (Beijing), and several other cities from 2005 onward.

Pre-empting is about more than simply identifying and eliminating threats; it is also about a process of shaping social demands so that the Party-state’s security apparatus has better control and faces smaller challenges. Social management is the process by which the C.C.P. leadership attempts to manage its relationship with both the party cadres and society, to ensure that it remains the ultimate authority in power. Part of the social management innovation process is about not only using technology to censor thought but also to actively shape it. The political security the Party is trying to preserve is best served when people choose not to challenge Beijing’s authority.

The question is not whether the Social Credit System will increase Beijing’s capacity for control, but how it will affect the behavior of Chinese citizenry on top of the broader, more action-oriented surveillance policy and capabilities. The evolution of state surveillance since the turn of the millennium has followed a clear path:

  1. deployment of more sophisticated surveillance systems and rebuilding of informant networks;
  2. database integration between local, provincial, and national authorities;
  3. automation of data sharing and task assignment; and now
  4. deploying a system-wide surveillance system that allows individual tracking with more scrutiny than currently allowed by true-name registration for telecommunications services, travel and logistics, as well as a variety of other services in daily life.

The effectiveness of the Social Credit System ultimately will be whether it succeeds in shaping people’s behavior in ways predictable to the authorities, because the system will undoubtedly strengthen surveillance and investigative capabilities of the Chinese state.