How Smart Are Smart Cameras? Surveillance and Protest Dynamics in China

• In Event: State Repression and Human Rights II (Virtual Session)

Abstract

Dictatorships around the world are deploying advanced digital surveillance tools to monitor, track, and surveil citizens. China’s rapid and extensive expansion of surveillance camera networks has drawn global attention. China’s camera systems are often depicted as high-tech instruments of state repression that use facial recognition and artificial intelligence to identify and track 1.4 billion people. Yet, little is known about the relationship between surveillance cameras and protest in China. In this paper, we explore the mechanisms through which surveillance cameras affect protest. We created an original dataset of local camera density for approximately 3,000 Chinese counties from 2013 to 2019 using machine learning classification of Baidu Maps street view images. We combine this data with comprehensive records of over 215,000 geographically located protests drawn from social media data. We use a border regression discontinuity design to estimate the causal effects of camera expansions on protest. We find that the expansion of digital cameras in Chinese counties significantly reduces the number of protests regardless of their size. In addition, surveillance cameras lead to a decrease in protests involving violence and police deployment. The evidence suggests that cameras have a strong deterrence effect rather than an effect of smart policing. It also implies that Chinese protesters are strategic rather than spontaneous.

Authors

• Xu Xu, Princeton University

• Han Zhang, The Hong Kong University of Science and Technology