Tenzin Norgay

Long march to perfecting the state of surveillance in China

facial recognition

File photo of Hikvision’s facial recognition technology capable of recognizing “ethnic minorities”.

Three years ago, during a mock exercise Chinese police in Guiyang City challenged the BBC’s John Sudworth to go anywhere in the city without being found by them. Within seven minutes after the reporter left the surveillance control room, he was caught by security officers based on his location caught on camera. Through this mock exercise, China sent a loud message: the Chinese state is omniscient and omnipresent.

A lot has been written about the technological surveillance prowess of China and its trialing in Tibet prior to wider rollout. To a significant degree, surveillance aided by technology has deterred human rights activists in China for the fear of being caught by the state’s eyes all around them. Deterrence of freedom and rights activism in Tibet has also taken a hit due to stepped-up surveillance in Tibet in the wake of popular protests against Chinese rule in the spring of 2008. This raises the pertinent question of how powerful the surveillance technology is and whether China has perfected surveillance technology. Making the location of citizens scrutable and legible to surveillance data gathering is not the same as knowing what they think or say or intend. How useful is this knowledge to a party-state seeking complete control?

Sinologists Jessica Batke and Mareike Ohlberg set out to answer these questions. In their quest to understand the State of Surveillance in China, Jessica Batke and Mareike Ohlberg asked six questions: “To what degree is Xinjiang a model for the rest of the People’s Republic of China (PRC)? Who, exactly, are local governments elsewhere trying to track? Why do they think such surveillance is necessary? How much does the application of national surveillance plans vary from place to place? How costly is it to local governments? And how well do any of these systems actually work?”

Based upon analysis of 76,000 surveillance technology procurement documents on the government procurement network spanning 16 years between 2004-2020, the authors conclude that China has not perfected surveillance technology although the intent is clearly “to eliminate any public spaces where people might remain unwatched.” China is not yet an Orwellian state as Chinese leaders would want us to believe to project the infallibility of the Communist Party of China. But it may be well on course to become one in the long run by training both machines and humans.

The authors argue that the surveillance technology and conceptual framework is the same across China, although scale and purpose of deployment may vary from location to location. While dissidents and potential criminals may be the object of surveillance in parts of the Chinese heartland, all members of a particular “ethnic” or religious group are targeted for their “ethnic” or religious affiliation. This finding comports with what Tibetans and Uyghurs have experienced for several decades solely due to their distinct identity and the socio-political-historical context under which the two territories and peoples became “ethnic” minorities under Beijing’s rule.

Parsing through 76,000 documents surely is overwhelming, and explaining the nuanced findings is a huge challenge without boring your audience to death with statistical and technical jargon. The authors skillfully told the story by doing a comparative analysis of three case studies in terms of demand and deployment of surveillance technology. The case studies focused on Shawan County in far west Xinjiang, Xijiao in southeast coastal Guangdong province, and Harbin in northeast Heilongjiang province. The three locations are scattered on the map of China, thereby making the case studies representative of the research conclusion.

The authors found that the local authorities have wide latitude in deciding the type of surveillance technology, and the scale of deployment in the three places varies according to threat perception. In Xiqiao, surveillance is deployed to check daily activities of the people with a focus on “key persons,” a term for types of people the authorities view as dangerous. Officials in Harbin were working on surveillance capacity to predict where the city residents will go and what they are likely to do. In Xinjiang, unlike other parts of China, surveillance pervades daily life not only through technology but the massive presence of “convenience police stations” and security checkpoints throughout the region. Other parts of China pale in comparison to the intense scrutiny that exist in Xinjiang although the technology and conceptual frameworks are the same as in other parts of China.

The following are notable takeaways from the report:

  • The pervasive and invasive surveillance system in Xinjiang is designed to target the Uyghurs as a group. It is discriminatory by design to distinguish Uyghurs from other groups present in the region. For example, the 8 million Han Chinese in Xinjiang accounting for 40% of the Xinjiang population are not subject to the same level of surveillance meted out to the Uyghurs. The authorities target Uyghurs as a group instead of focusing on outsiders, dissidents and criminals like in other parts of China.
  • The state of surveillance technology, conceptual frameworks and programs in Xinjiang is not unique, although the scale of deployment and the intensity are when compared to other regions in China.
  • Purchase of surveillance technology has increased dramatically in the past two years.
  • 14 billion RMB ($2.1 billion) was spent between 2016 and 2020 for deploying the “Sharp Eyes” program alone, in addition to expenditures on other surveillance projects. Surveillance expenditure sometimes reached half of the annual total public security expenses.
  • Last year, at least 998 counties spread across China purchased surveillance equipment of some type.
  • Facial recognition cameras are not omnipresent, although they are gaining traction in deployment.
  • The Public Security Bureau accounts for 65% of purchase of surveillance technology.

Surveillance in Tibet

The authors mention Tibet only one time in the 17-page report. But this does not mean that Tibet is not a concern in terms of deployment of surveillance technology. In fact, both Tibet and Xinjiang are at the same extreme end of the surveillance spectrum in China. Both are outliers in terms of scale compared to surveillance practice in the Chinese heartland. Both have a massive presence of so-called “convenience police stations” and security checkpoints installed across their homeland. Surveillance technology is being used not only to monitor “criminals” but for “social management” of Tibetans and “social stability” in Tibet.

A simple keyword query for Tibet in the Chinese government procurement network reveals that the public security bureau in Tibet has no intent of scaling down the level of surveillance already in place. Besides the usual oppressive tools of choice, it is most likely that drones will soon be part of the surveillance mix to monitor the Tibetans.

A wide range of documents are available on the Chinese government procurement network (www.ccgp.gov.cn), but three items of interest in terms of recent procurement notices will be highlighted here.

  • Genome surveillance   Based on anecdotes, it has been known for long that DNA of Tibetans is being profiled by the state. The project does not seem to be complete yet, although some observers had earlier reported the project to have been completed. Select recent procurement notices for DNA database construction (July 8, 2019), DNA reagents and Consumables (July 21, 2020), Ultra-micro magnetic bead method DNA extraction kit (June 27, 2020) show that it is an ongoing project. Since DNA profiling is highly controversial, the strategy so far appears to be to keep the project out of view to avoid condemnation for mass profiling. But recent procurement notices reveal that the strategy has changed, as the public security bureau is actively bidding for rigorous DNA profiling; one such procurement notice comes with a price tag of 1 billion yuan. DNA data of Tibetans is scrutinized intensely by the Chinese authorities. Geneticist Yves Moreau, an engineer and professor at the Catholic University in Leuven, Belgium, said in an interview with NPR that the DNA profiles of “Tibetans are studied 40 times more intensely than the Hans, and the Uighurs are studied 30 times more intensely than the Hans.” In Crackdown on Genomic Surveillance, Moreau wrote that half of the genome studies of Tibetans and Uyghurs are authored by the police force, military or judiciary.

    The official justification to profile DNA is to catch criminals. However, such a justification is problematic, as a wide range of Tibetan activism for language, environment, culture, freedom of opinion and expression, etc., is criminal by the official Chinese definition. In recent years, even discussing the Dalai Lama’s Middle Way Approach for conflict resolution has been criminalized through law.

    DNA-profiling technology has made great headway in solving crimes worldwide, and it is legitimate for law-enforcement agencies to use the technology with stringent safeguards and oversight. However, use of such technology is problematic in Tibet and in China, where the Communist Party of China is the final authority free of any oversight for human rights abuses.

    It is not outlandish to imagine Chinese leaders ordering targeted extradition of Tibetans in exile by furnishing DNA of family members back home as proof of their Chinese citizenship to foreign governments. For example, last month Indonesia extradited three Uyghurs to China instead of Turkey on the merit of China producing DNA of their family members in Xinjiang to prove their Chinese citizenship. China’s claim of universal jurisdiction of its national security law also throws open a range of imaginable situations under which DNA profiles can be used for extradition of dissidents (read: criminals in Chinese) including Tibetans from foreign countries.
     
  • Drone surveillance   Use of drones for domestic surveillance is another controversial issue worldwide. Drones to surveil civilian Tibetans has not been observed so far—at least publicly—but that looks to change soon. In light of a procurement notice (June 22, 2020) on the Chinese government procurement network, the Lhasa Public Security Bureau will soon deploy drones to surveil the Tibetans in Lhasa.

    In the ongoing global Covid-19 pandemic, journalists and observers have reported authorities in China stepping up collection of citizen’s personal data through a variety of health surveillance apps. With the deployment of talking drones, authorities warned ordinary Chinese citizens to confine themselves in their homes at the height of Covid outbreak in China.

    But the deployment of drones to surveil Tibetans would be altogether at a different level given the political context under which the technology will be put to use. The Chinese authorities already operate with a combat mindset in Tibet. It is anticipated that sending drones to surveil Tibetans won’t be as mundane as ordering someone to lock themselves up in the house to escape from Covid.
     
  • Big data analytics   Big data analytics feature prominently in the government of China’s plan to surveil and control everyone under the rule of Beijing. For Tibetans and Uyghurs as the two minority groups most distrusted for “stability maintenance,” Chinese security authorities deploy big data analytics and policing techniques to surveil and control Tibetans. Although it is not unique to Tibet, a skewed dataset is a major concern when compared to big data analytics in other parts of China.

    Analysts, including Batke and Ohlberg, point to the ease of generating vast amounts of data, which are then analyzed according to algorithms that predetermine propensity for criminalized behaviors according to rules, categories written into the algorithms. The naïve faith in high tech creates the illusion that algorithms are objective and capable of discovering criminal intent well in advance. In reality the old GIGO maxim holds: garbage in, garbage out.

    Besides the TAR CCP’s procurement notice for a classical video surveillance network (July 10, 2019), the public security bureau of the Tibet Autonomous Region is actively building cloud computing for data sets of Tibetans in the TAR in view of select procurement notices for cloud investigation (October 22, 2020) and Nagchu City cloud video surveillance network (September 10, 2019). These clouds are expected to be plugged into the national level police cloud maintained by China’s Ministry of Public Security. According to the Rand Corporation’s Chinese Views of Big Data Analytics, “The MPS [Ministry of Public Security Bureau] is exploiting new data sets that it plans to centralize in a ‘police cloud’. Eventually accessible to all provincial and municipal police authorities, the police cloud will increase the ease with which police can make connections across disparate databases—including non-crime-related systems, such as housing and employment records—to rapidly identify people, places, and businesses of interest.”

    The goal for building a police cloud is to preempt any demonstration or protest by Tibetans by proactively tracking activities of all Tibetans. Those deemed to harbor “ill thoughts” against the government or who have expressed their dissent in the past form the focus of people in big data analytics.

    Human Rights Watch warned that the police cloud “scoops up information from people’s medical history, to their supermarket membership, to delivery records … the Police Cloud system track where the individuals have been, who they are with, and what they have been doing, as well as make predictions about their future activities.” In other words, privacy is not a right, but a luxury Tibetans in Tibet can only dream of.

    Predictive policing based on big data analytics compounds the issue of human rights abuses committed by Chinese law enforcement agencies. Preempting the “culprits” before they have even carried out any activism is highly worrisome, given Chinese authorities’ track record of decades-old repression in Tibet and racial biases against Tibetans. Predictive policing is already a reality in Xinjiang that looks to be on the verge of replication in Tibet.

The Communist Party of China’s reliance on surveillance technology to govern everyone under its rule is widely known. The party wants the people it rules to believe that it is infallible by internalizing fear of its surveillance prowess. Jessica Batke and Mareike Ohlberg’s research gives the crucial analytical conclusion that the surveillance technology in China is not perfect. Despite its deep-seated intent, the party still has a long march ahead of it in perfecting the panopticon system. The implication of this research work is that the Chinese surveillance state is navigable, and there is still time to reverse or at least stop the state of surveillance at its current stage. The party’s mastery of surveillance at the expense of the people it rules is no longer confined within the domestic borders of China. Over 80 countries worldwide in awe of the government of China’s ability to control its citizens import surveillance technology from China. The state of surveillance technology in China has direct implications for liberties and freedoms of everyone on the face of the planet. The liberties and freedoms of people across the globe, irrespective of the political systems under which they live, are tied to whether the surveillance technology in China is perfected or not.

Tibetans have reason to fear China’s delusions of omniscience and omnipotence. Yet in the long run, more and more Tibetans are criminalized, classified as security threats and punished, without having done anything criminal. This only alienates Tibetans further and undermines the credibility of the state. China will eventually discover this has been counterproductive.