China, An AI Surveillance State

Security cameras in front of the portrait of Mao Zedong in Tiananmen Square

In 2000, U.S. President Bill Clinton dismissed China’s attempts at censoring the internet as “like trying to nail Jello to the wall.” Six years later, China completed the Great Firewall of China which banned almost all western news outlets including the CBC, all western social media platforms that would subsequently play a role in the Arab Spring, and even all pornographic content including livestreams of banana-eating! This among other forms of social control echoes those depicted in George Orwell’s dystopian novel, 1984, in which a totalitarian “Big Brother” oversees all citizens by security cameras. Seventy years after the novel, big data is bringing “Big Brother” to China.

In January 2017, China announced a pilot social credit system that rates Chinese citizens in the four areas of administrative affairs, commercial activities, social behaviour, and the judicial system. The system will roll out in Shanghai Municipality, Jiangsu, Anhui, and Zhejiang provinces, regions with a combined population six times that of Canada. The rating is reported to determine access to privileges that include travelling abroad and sending children to the best schools.

44 financial firms have so far signed up to submit data to the social credit system and 300 commercial banks have been connected for monitoring transactions. In addition, China’s biggest technology companies are reported to share data with the system. China’s “App for everything”, WeChat, confirmed in September 2017 in its privacy policy that it discloses data of its near 1 billion users upon “request by a government authority, law enforcement agency or similar body.” Rumours that this occurs in all large Chinese technology companies circulate as Baidu, Alibaba, and Tencent have internally instituted party committees that assess daily operations.

This social credit system is lenient compared to an experiment conducted in 2010 in Suining County of Jiangsu Province in which citizens were rated from A to D. Failing to care for elderly relatives or causing disturbances that block governmental buildings lowered the grade, whereas being classified as a model citizen improved it. Those graded with Ds were excluded from official support or employment.

China became a complete surveillance state when it installed 200 million street cameras with artificial intelligence technology that is able to determine pedestrians’ age, gender, and attire. The system may soon be able to recognize citizens’ faces as China is leading in facial recognition technology. Face++, backed by Chinese and Russian state-owned funds, has stated, “we want to build the eyes and brain of the city, to help police analyze vehicles and people to an extent beyond what is humanly possible.” Though China has not yet established surveillance on its 1.3 billion citizens, this future is within reach. In September 2017, 25 wanted individuals were arrested at a beer festival surveilled by a facial recognition system with access to a police database of suspects and a 98.1% accuracy rate. Moreover, the database of citizens’ faces already exists, as all citizens must have a government-issued photo ID by age 16.

Armed with artificial intelligence-powered surveillance and propaganda apparatus, China would be equipped to manage the growing discontent against a litany of social, economic, and environmental problems. Social unrest grew tenfold from 8,700 in 1993 to 90,000 in 2010, all suppressed with China’s public security budget, which exceeds China’s defence budget! The coming social credit system and mass surveillance will further instil fear in civil society, and reduce this cost of suppression.

The data explosion from the mass surveillance of economic activity may also play a role in restoring stability in China’s troubled economy. China has announced to use big data to systematically identify credit risk as part of a larger effort to curb bad debt. Using machine learning, the commercial data from mass surveillance has even been suggested by Jack Ma to predict market demands, and thus reduce the inefficiency of China’s planned economy. The same inefficiency inherent to an economic system of state-ownership plagued the Soviet Union during the Cold War. Yuval Noah Harari wrote in Homo Deus that “Capitalism won the Cold War because distributed data-processing works better than centralized data-processing.” With machine learning, China may continue its “central data processing,” sustain economic stability, and yield no influence to the free market.

In fairness, China is not alone in surveilling its citizens, nor is mass surveillance not without merit in the age of terrorism. Other governments with covert mass surveillance programs include the United States, the European Union, the United Kingdom, Germany, Australia, Switzerland, Sweden, France, India, and Russia. Canada also collects metadata of phone calls and emails, but not the content of them. However, open debates about the ethics of artificial intelligence and institutional checks and balances in many of the countries above are lacking in China.

Ray Kurzweil’s law of accelerating returns states that technological advances beget technological advances, and thus technology charges ahead at an ever accelerating pace; this is a generalization of Moore’s Law. Similarly, if surveillance begets data explosions, improvements in algorithms, and increased suppression, which in turn clears the way for further surveillance, then where is China relative to the point of no return on the path towards a dystopian future?