China’s Secret to Controlling the Internet (2024)

When the internet first became accessible worldwide, optimists hoped that it would be able to break authoritarian regimes’ ability to control information. Dictatorships, the thinking went, would be powerless in fending off the information revolution enabled by the decentralized nature of the internet. In 2000, U.S. President Bill Clinton famously compared China’s attempts to control cyberspace to “trying to nail Jell-O to the wall.”

However, Chinese leaders did not think so. Although they did not have a comprehensive plan at first, they developed tactics over the decades that nailed the proverbial Jell-O to the Great Firewall of China. Beijing’s approach to taming the information revolution is unique: It has focused on controlling access to the internet, not just censoring its content. Chinese authorities prioritize knowing who is online—and this allows them to identify, track, intimidate, and punish those who are potential threats.

When the internet first became accessible worldwide, optimists hoped that it would be able to break authoritarian regimes’ ability to control information. Dictatorships, the thinking went, would be powerless in fending off the information revolution enabled by the decentralized nature of the internet. In 2000, U.S. President Bill Clinton famously compared China’s attempts to control cyberspace to “trying to nail Jell-O to the wall.”

This article is adapted from The Sentinel State: Surveillance and the Survival of Dictatorship in China (Harvard University Press, 336 pp., $35, February 2024).

However, Chinese leaders did not think so. Although they did not have a comprehensive plan at first, they developed tactics over the decades that nailed the proverbial Jell-O to the Great Firewall of China. Beijing’s approach to taming the information revolution is unique: It has focused on controlling access to the internet, not just censoring its content. Chinese authorities prioritize knowing who is online—and this allows them to identify, track, intimidate, and punish those who are potential threats.

This concept—which the Chinese state calls “battlefield control”—underpins one of Beijing’s most widely used surveillance tactics both online and offline. For the Chinese Communist Party (CCP), which sees suppressing any threats to its power as a life-and-death struggle, cyberspace is a new battlefield. And the CCP doesn’t just rely on technology to control it. It also uses technology to know where and when to deploy extra police manpower to the front lines of its war for public control, whether those are train stations, shopping malls, hotels, or universities.

To dominate the battlefield in cyberspace, the government relies on two bodies for surveillance: the Central Cyberspace Affairs Commission, which is part of the CCP, and police units. Party censors determine which content to block or filter, while police enforce censorship through tactics such as installing surveillance hardware, blocking and filtering suspects’ communications, and conducting investigations and arrests.

The CCP began to assert control over the internet in the mid-1990s, but it was not until 2014 that it established the Central Cybersecurity and Informatization Leading Group, a nationally integrated agency charged with both regulatory and censorship responsibilities. (In 2018, President Xi Jinping promoted this group to the status of a central commission.) Local jurisdictions, meanwhile, have set up equivalent offices attached to their CCP committees.

Local outlets of the Cyberspace Affairs Commission—which I’ll refer to as “cyber agencies”—lack the workforce and technological capabilities to conduct sophisticated surveillance. Instead, their main tasks are routine censorship and promulgating disinformation. For instance, the municipal cyber agency of the city of Longnan, with just under 3 million people, reported that, by the late 2010s, it used big data and cloud computing to monitor online public opinion; in 2019, the agency monitored 515,000 pieces of online information about Longnan, 8,000 of which were deemed to be negative. Local cyber agencies also recruit internet commentators to conduct online campaigns to manipulate public opinion and spread disinformation.

Cyber police units, on the other hand, take charge of enforcement and surveillance. Cyber police were first organized in public security bureaus, or PSBs, throughout China in the early 2000s. The city of Yanan’s cyber police unit has reported that its main missions include “monitoring and controlling harmful information; collecting, analyzing, and reporting developments on the internet; enforcing regulations on internet cafes; and investigating and dealing with cybercrimes.” The cyber police units in local PSBs are relatively small even though they perform such duties. A typical county cyber police unit has about five to six officers.

Both the local cyber agencies and cyber police units deploy high-tech solutions to patrol the internet 24/7. For example, the cyber agency of Santai county has used web-monitoring technology called Real Time eXchange. Cyber police officers are required to report important developments to the PSB leadership and also to the county party committee and government.

Although local records do not specify this, it appears that cyber agencies determine which content to delete and block and then instruct police to execute it. The cyber police unit of the city of Ergun’s municipal PSB, for example, has stated that it is responsible for “organizing and implementing the ‘routine work’ of Ergun’s cyber agency.” (“Routine work” almost certainly refers to censoring online content.) Reports of cyber police taking bribes from businesspeople to delete critical posts also indicate that cyber police are tasked with enforcement. And when cyber agencies discover online materials requiring investigation, they alert the cyber police. For instance, as soon as the cyber agency of Yunlian county became aware of what it called a serious “internet rumor” in 2017, it contacted the cyber unit of the county PSB to investigate.

This division of labor makes administrative sense. It is unnecessary to duplicate technological capacities across agencies, and housing surveillance capacities in police buildings poses less of a security risk, since they are better guarded than government buildings. The police also operate China’s Public Information Network Security Surveillance and Control System—the so-called Great Firewall.

The work of surveillance and control is labor-intensive—cyber police must conduct in-person investigations and visit individuals suspected of violations such as posting censored materials. In 2016, the cyber police in the city of Guiyang’s Baiyun district investigated 85 individuals in person. The cyber police in a neighboring district, Yunyan, were even more aggressive, reporting 200 such investigations that year. Penalties for harmful online activities include detention, fines, and “criticism and education.”

Cyber police use several tactics to detect the identities of violators and potential violators. One approach is to trace IP addresses, the unique identification associated with a local network that connects to the broader internet. This is easy to do because online access is provided by state-owned telecommunication companies. But more sophisticated users can avoid this sort of identification by routing their online traffic onto a virtual private network. And additional measures are needed in order to identify the owners of social media accounts, which allow anonymous posting. In February 2015, new rules required that all internet users provide their real names when registering accounts on chatrooms, the popular messaging app WeChat, the extremely active microblogging service Weibo, and all other social media outlets.

Cyber police also surveil access points such as internet cafes and public Wi-Fi networks in hotels, malls, airports, and other venues to identify users beyond their home networks. (Internet cafes are not as common today as they were in the early 2000s, but they’re still widely used, especially by gamers.) Regulations help them do this. In 2001, the government mandated that private internet cafes, which were supposed to require ID from every user, retain customer information, including identities and online activities, for 60 days. Internet cafe operators must also obtain a license from the local PSB and Cultural Bureau, a government agency that regulates the entertainment sector. More recent rules require that internet cafes install ID card readers so that customers gain online access only after scanning IDs that store identifying information, including a color headshot.

Similar regulations were in place even before the introduction of cyber police but were loosely enforced. Today, cyber police strictly enforce these rules and also train “security attendants” at internet cafes, who presumably ensure compliance with regulations and also perhaps spy on customers. Some jurisdictions even require that internet cafes install cameras to monitor customers.

As for monitoring public Wi-Fi networks, in the late 2000s local cyber police began requiring operators to install unspecified “security technical measures.” A concerted national effort on this front likely began in 2014. That year, Wuhan police, for example, initiated a three-year program to install “security management systems” in all public Wi-Fi networks. In 2016, the Yunyan PSB installed 560 public Wi-Fi-monitoring systems. Thousands of similar systems were installed in two counties in Sichuan province in 2017-18. On average, a Wi-Fi surveillance device costs about $306, indicating that this surveillance program—including post-installation maintenance and operation costs—requires significant resources.

Cyber police pay special attention to users designated as “key individuals,” people placed on a blacklist by local authorities for engaging in activities viewed as threats to regime security and public safety. These likely overlap with political dissidents, liberal scholars, human rights activists, members of illegal religious organizations, and practitioners of Falun Gong and other movements deemed by the state as “evil cults.” Key individuals also reportedly include some well-known pro-government personalities, indicating the party’s paranoia about individuals with a significant public following regardless of their political loyalty.

The number of key individuals online varies in different parts of China, demonstrating that jurisdictions have wide latitude in making these designations. In 2018, Hengyang county had 100 key individuals under surveillance. The cyber police of Oroqen in Inner Mongolia claimed to have had 25 under surveillance in 2015. The city of Changzhou’s Yunhe district reported that its police had in-person contact with 62 key individuals in 2016. Several jurisdictions, however, have identified many more. Jishan county had 1,141 such targets, roughly 0.3 percent of its population, under surveillance in 2018. Between 2011 and 2014, cyber police in Tancheng county “registered and controlled” 3,475 key individuals, around 0.4 percent of the population.

Although there are few public details on how China surveils key individuals, this is one of the most important surveillance tactics in China because it allows the authorities to keep track of individuals most likely to cause trouble. At a minimum, it seems that cyber police have special files on their targets, as reported in several jurisdictions. For instance, according to a report issued by the city of Neijiang’s PSB in February 2011, the cyber police unit was instructed to collect basic information on all types of key individuals, designate officers to employ “various technical means” to scrutinize them and use an unidentified special police database to ascertain their online identities. The report also mentions real-time surveillance of targets through information obtained from internet service providers and internet cafes.

Such a system of distributed surveillance has allowed the Chinese surveillance state to control the battlefield in cyberspace and neutralize threats to CCP rule. As with its other forms of preventive repression, these tactics allow the state to monitor regime opponents’ activities and restrict their freedom of action.

This surveillance state relies on the CCP’s strengths in organization and mobilization. New specialized bureaucracies—in this case, cyber agencies and cyber police—form quickly to carry out the central state’s agenda. Although technology is important, informants and police investigations and intimidation are essential. China’s success in controlling the battlefield in cyberspace reveals the CCP’s tactical adaptability in the face of a novel threat to its hold on power. If anything, it shows that we must never underestimate the party’s determination to perpetuate its rule—or its capabilities for doing so.