Authoritarian regimes increasingly use digital surveillance of citizens to deter mobilization, but information alone does not prevent protest. To address this, dictators have historically relied on human “street-level agents”—such as local police and grid workers—to act on the data that technologies like facial recognition and social media monitoring provide. However, governments historically had very limited ability to observe whether these agents were truly patrolling and reporting honestly. Moreover, these agents often lack the motivation to perform risky interventions and/or lack the capacity to complete tedious tasks associated with protest prevention. This results in a “principal-agent” problem where they shirk their duties.
Han Zhang, the Young Family Assistant Professor of Sociology and International and Public Affairs at Brown and a PSTC affiliate, argues that the Chinese government has solved this shirking problem by creating two distinct layers of surveillance. The first is the conventional understanding of digital surveillance, targeting the public, to detect and deter mobilization. The second is a digital workplace surveillance program that monitors the government’s own frontline workers themselves. While these two layers operate independently, each capable of reducing protest on its own, the potential that they may also reinforce one another exists.
Zhang and his team tested this “Two-Layer Panopticon” framework in China using an original dataset of 51,611 government procurement contracts that captures digital workplace surveillance of agents alongside mass surveillance of citizens. They demonstrate how Chinese cities use tools like GPS tracking, body cameras, and AI-powered "smart platforms" to watch their employees. Grid workers, who are tasked with monitoring households and mediating conflicts, must now carry smartphones that track their real-time movements and require them to upload live photos to verify they are actually on patrol.
“Our findings show that cities implementing digital surveillance to monitor grid workers experienced a statistically significant decrease in protest occurrences,” Zhang explains. “Through tools like real-time GPS tracking and automated reporting systems, rulers can detect whether agents are patrolling assigned areas, filing accurate reports, or responding to risks, and then reward or punish them accordingly.”
Digital surveillance on human agents reduces shirking problems and enhances protest prevention. Furthermore, this strategy is especially attractive to dictators, not only because it’s easier to implement and costs less, but also because it’s more publicly acceptable than population-wide mass surveillance.
By bridging the gap between digital authoritarianism and bureaucratic theory, Zhang’s research offers a sobering look at how modern regimes use AI not just to spy on the public, but to tighten the grip on their own administrative machinery. This “Two-Layer Panopticon” creates a more resilient form of control where the state's eyes are turned both outward toward society and inward toward itself.