SHARON LI does a deep dive into China’s surveillance development and its domestic and global implications.

As China races towards establishing itself as the world’s leading technology superpower, new developments in advanced technology have seen large-scale integration on the state level, particularly within the expanding presence of the country’s surveillance state. In previous decades, surveillance has largely been limited to vast sets of video footage and live-feeds, all of which require human labor to monitor and comb through; however, algorithmic surveillance is steadily becoming the norm with the help of machine learning and other forms of artificial intelligence.

In the past half-decade, global surveillance equipment sales increased by 55%. According to HIS Markit, the Chinese market accounted for 44% of all global revenue in video surveillance equipment, much of which can be attributed to the government’s recent expansion of local and national-level security programs. Just two of the many programs in the country include Xue Liang, or Sharp Eyes, and the government’s “safe city” or “smart city” project. The former is a surveillance and facial-recognition security program currently being piloted in fifty rural towns, while the latter targets urban areas and currently has 500 pilots across the country. Unsurprisingly, the companies that develop the equipment and technology for such programs have enjoyed major financial windfalls.

Hikvision, a Hangzhou-based company founded in 2001, currently stands as the world’s largest manufacturer of surveillance equipment. Their controlling shareholder is the Chinese government, which operates through the China Electronics Technology Group, a 100%  state-owned defense and military electronics manufacturer. Contracts from the local and national government have, naturally, significantly bolstered Hikvision’s finances. In 2017, the company took in a revenue of USD 6.07 billion—in the same year, they won five security contracts in Xinjiang, totaling USD 270 million. One such project is in Moyu, a county which reportedly saw the disappearance of half of their Uyghur population to re-education camps. According to the original tender obtained by Agence France-Presse, the contract calls for a network of 35,000 cameras to monitor everything from streets to schools and office buildings to mosques—specifically, 967 facial recognition cameras for each of the 967 mosques in the county. These new cameras, along with 840 dome/bullet cameras equipped with full-color nighttime vision,  will be integrated into the already existing mosque surveillance system in the region, which currently consist of five basic surveillance cameras for each mosque. Aside from mosque surveillance in Moyu, Hikvision will also develop and provide six video monitoring systems in the form of panoramic cameras for the re-education centers, which allow for complete surveillance with no blind spots. In return, Hikvision is guaranteed an 8% annual return which will likely be generated through the increased fines that will result from the surveillance system.  Other contracts with Hikvision include a USD $53 million facial recognition deal in Pishan—which, according to the original tender for the competitive consultation announcement, will include mosque video lecture systems and networking systems—and a USD $73 million contract in Xinjiang’s capital city of Urumqi, which will require 30,000 additional security cameras. And these are just three of the dozens of surveillance projects in Xinjiang—three of the USD $1.2 billion-valued surveillance projects that Hikvision and its competitor Dahua have won in Xinjiang in the past few years. Already, Xinjiang’s security costs far surpass any other part of China’s, and in the past year alone, the region’s surveillance and security spending has almost doubled, with the government citing the threat of extremism and terrorism as the cause.

The government’s method of spurring development within the surveillance market is also notable. Rather than directly pour funds into the market and risk immediate market distortion, the state instead provides local governments with funds so municipalities can independently put out competitions that guarantee certain companies massive contracts and long funding-cycles. Such million and billion dollar contracts come in the form of public-private partnerships (PPP), in which winning companies finance the majority of the spending for “guaranteed investment returns.” These contracts are incentivizing surveillance companies such as Hikvision to update their equipment and develop the more advanced technology required for biometric surveilling and algorithmic facial recognition through machine or deep learning. Deep learning, in particular, has gained attention as the new frontier in AI and is able to go beyond machine learning in its capability to “self-learn”—in this case, to classify certain features or patterns— based on an artificial neural network. Already, Hikvision has launched multiple products with deep learning algorithms, which consist of capabilities from auto-tracking to license-plate identification, to immediate classification of blacklisted individuals. And the company has made other strides towards algorithmic surveilling. In 2014, they showcased thermal monitors that could record under ultralow light conditions, license-based recognition, and explosion-proof cameras that could “alert authorities to large gatherings.” This last type of camera, the ‘dome-shaped bullet’ camera happens to be the same ones installed in Moyu County of Xinjiang.

But what does this mean for surveillance in China? Undoubtedly, with the help of Hikvision’s development, advanced surveillance equipment will only become more ubiquitous in already heavily policed areas such as Xinjiang. Even outside of autonomous regions, this type of advanced surveillance is already being embraced in the form of expanding “safe city” and “smart city” programs, particularly in large urban areas. Predictive policing, “voice printing,” social credit scores—all of these methods have been used and will continue to develop with the help of vast swaths of available data, loose data laws, increasing power of the state, and an influx of funding. And as China continues to push for the incorporation of more advanced technology in the surveillance state, companies such as Hikvision will continue to lead the way. 

However, the political and macroeconomic implications of such surveillance extend far beyond just China. Democratic countries such as the United States have begun to very publicly voice concerns that China’s technology is intentionally bolstering an authoritarian world order. Such fear is not completely without reason. Chinese surveillance technology has proven to have a far and penetrating reach, and is undoubtedly increasing the ease at which governments can monitor and potentially control their people. Furthermore, Chinese companies known for their role in developing advanced surveillance techniques for the state such as Sensetime, CloudWalk, Hikvision, and Yitu not only lead in countries that are launching their own smart city programs or advanced surveillance systems such as Malaysia, Singapore, and Zimbabwe, but also in advanced democracies like the United States.

By providing more effective means for governments to seamlessly integrate these programs, China—though not exactly imposing an authoritarian world order—is advancing it, and simultaneously making it more enticing. And it is bolstering not just the traditional form of authoritarianism, but an unprecedented digital authoritarianism. However, it is crucial to recognize that China is not single-handedly creating this new system. As mentioned, other countries are not only financially supporting such endeavors but creating parallel systems as well. The United States has dramatically escalated mass surveillance programs in the past decade, many of which were headed by the National Security Agency and Department of Homeland Security. And such surveillance databases and policing techniques—which typically operate to curb terrorism, espionage, money-laundering, and other large-scale crimes—have been accused of disproportionally targeting people of color. Furthermore, through surveillance programs such as PRISM and USA Patriot Act, the United States government works and partners with commercial companies such as Facebook, Apple, Google, AT&T, and T-Mobile to surveil users. This may sound familiar, as China is taking similar—if not identical—measures to surveil their own people. Unlike in China, however, the United States governments lags far behind in terms of integrating artificial intelligence and more advanced technology into their surveillance methods, which can largely be attributed to the bureaucracy and red tape that restricts the tech sector from state-level adoption. Furthermore, as shown earlier last year with the protests that resulted from Google and its involvement in using “AI to process drone video feeds” for the Pentagon’s Project Maven, civilians have more of a voice in objecting to what they see as technological misuse—or at least, when such information is public. Thus, while the United States is creating domestic surveillance programs that echo the very ones they criticize China for, there is too large a gap between the private sector and the state to create the advanced surveillance system that China has put in place.

For now, it is difficult to say exactly how much machine learning technology and other forms of surveillance are altering the democratic world order, given such technology has only recently begun to be developed. Furthermore, we cannot claim that, without China, such technology would not exist or would not eventually develop. On the other hand, without the rapid and unfettered development of surveillance technology that the unique political and economic conditions of China allows, it is highly unlikely that the international community would presently be facing authoritarianism in this form and at this gravity. Ultimately, one thing is clear. A new potential of digital authoritarianism is more real than ever. And while China may be leading it, countries like the United States are also contributing to its establishment. But for this potential to actually become a threat, it would take acceptance, adoption, and integration on an international-scale. At least in some parts of the world, this seems to be a likely reality.

Sharon Li can be contacted at sharon.li@yale.edu.