Behind the Golden Shield: China Reforms Public Security Intelligence

Publication: China Brief Volume: 17 Issue: 16

China's recently-retired public security chief Meng Jianzhu presided over a shift to AI-enabled policing.

In his last address before retiring as China’s security chief in October, Meng Jianzhu (孟建柱) briefed members of the Politburo Standing Committee on the use of artificial intelligence (AI) to predict threats to social order (Sina, September 21). AI-enabled policing is the latest evolution of the Golden Shield Project (金盾工程)—a nationwide network that is attempting to link surveillance assets nationwide with personal digitized information stored in public records. However, as China’s police employ more sophisticated technology, and collect ever-greater volumes of data, they are also adapting their organization to better process information (China Brief, June 3, 2011). Over the past two decades, the Ministry of Public Security (MPS) has built a “public security intelligence system” (公安情报信息体系) to improve how analysis reaches decision-makers. New intelligence structures are helping China’s security state cope with an increasingly complex, fluid, and networked society.

Intelligence Reform

In 2004—a year after the launch of the Golden Shield—the annual Meeting of National Public Security Bureau Chiefs (全国公安厅局长会议) produced a plan for a “public security intelligence system”. China’s top cops agreed that new ‘integrated intelligence structures’ would help “improve the level of intelligence collection, analysis, and assessments”. [1] The plan aimed to address longstanding problems with information sharing and a backward analytical culture at every level of public security.

For much of the 1980s and 1990s, public security intelligence work was a specialist function of the MPS’s “operational departments” (行动部门). Criminal Investigations (刑侦), Economic Crime (经侦), Counter-Terrorism (反恐), State Security (国保) and others contained ancillary intelligence offices (情报科) that supported their investigators. Operational departments set their own collection tasks and managed their own information resources—and intelligence was commonly seen as the exclusive property of whoever collected it. Information not of direct use to the collectors was discarded or withheld from other departments for fear of benefiting rivals.

Public security intelligence work was treated as a back-office job that served the narrow requirements of the department—not a guide for security policy. Information was rarely shared beyond departmental walls. There was no unified leadership over intelligence work, and no direction over how information was collected, analyzed and disseminated. In the early 2000s, scholars at the Public Security University in Beijing began to lament the lack of an “intelligence cycle” (情报流程) to synthesize information into a valuable product for decision-makers. While officers made use of tradecraft such as interrogations, interviews, covert surveillance, agents and informants, the lack of a process to synthesize information meant that traditional public security “intelligence” barely met the defining criteria of that term at all. What Chinese analysts derided as “small intelligence” (小情报) could not lead investigations, forecast threats, or guide strategy. [2]

By the mid-2000s it was clear that the MPS’ existing intelligence system was no longer capable of serving the needs of law enforcement. Local officials described their difficulties in maintaining the security state’s information dominance (制信息权) over an increasingly fluid and IT-literate society. They feared that China—like other globalized economies—was facing ever more complex and dynamic forms of criminality and dissidence (, July 29, 2014). Moreover, the erosion of the danwei (work unit) system and the uncontrolled flows that accompanied market reform deprived the party-state of its traditional eyes and ears. For one analyst, the MPS’ traditional intelligence system was “no longer suitable for the real requirements of the fight.” [3]

The Blueprint: Intelligence-Led Policing

In 2008, at a meeting of MPS police chiefs in Nanjing, former MPS Vice Minister, Zhang Xinfeng, heralded the achievements of the Golden Shield in laying a nationwide information infrastructure onto which new intelligence structures could then be grafted (Boxun, May 7, 2009). The blueprint for the new system is “Intelligence-Led Policing” (ILP; 情报指导警务), a British law enforcement model pioneered in the 1990s. It has since been enthusiastically adopted by law enforcement agencies across the globe, including in the United States. In China, it has become a mantra in public security policy speeches and academic analysis.

ILP is sometimes misunderstood as simply an increased use of intelligence operations by police. In fact, it is a management philosophy that places assessments derived from intelligence and analysis at the center of all strategic and operational decision-making. Seamless information sharing is crucial in ILP, as intelligence serves as the command and control link between strategic decision-makers, operational officers, and frontline cops. A municipal-level public security officer in Anhui accurately describes the logic of ILP as a “virtuous cycle”: intelligence is gathered from a criminal environment, collated, analyzed, and presented to decision-makers for consideration. Decision-makers allocate resources and formulate strategies on the basis of this intelligence. Frontline officers apply resources and execute the strategy to shape or disrupt the criminal environment.

There is a clear alignment in the benefits of ILP and the requirements of law enforcement in China. ILP offers resource efficiency when China’s police continue to struggle with personnel numbers (China Brief, April 12, 2012). In ‘Tier One’ cities such as Shanghai, Beijing, Guangzhou, and Shenzhen, population size would probably be overwhelming for a more traditional, investigation-led police force. This is possibly also true in the countryside, where budgets are small and jurisdictions are large (China Brief, September 4, 2015). Moreover, intelligence-led, preemptive policing frees up resources as police spend less time trailing leads. And, by consolidating intelligence and analysis from multiple law enforcement fronts, analysts can uncover “hidden” (内幕性) connections between criminal trends and incidents. [4] In other words, ILP lets the MPS do more—and better—with less.

Integrated Intelligence Departments

At the center of public security intelligence today are “Integrated Intelligence Departments” (IID) (综合情报部门), which sit at central, provincial, municipal, and county level public security bureaus. Staffed by a new cadre of “highly educated” analysts (one Chinese analyst reported 28,000 nationwide in 2015), they draw on the intelligence produced by operational departments as well as the statistical analysis offered by big data crunchers in Command Centers.  [5] IID’s analytical product is an integrated assessment of statistical predictions and specific threats. It not only reads the conditions in a community or region that support the generation of crime, but also the criminal entity, their intent, their target, their plan, and the resources they have available. The clarity of threat offered by an ILP-enabled system supposedly helps police manage the most persistent and complex forms of crime and disorder.

Perhaps the most important function of IIDs is how they have helped introduce an intelligence cycle into public security work. IIDs hold regular “intelligence consultation” (情报会商) meetings, where participants receive tasking from above and issue targets to collectors. These meetings can serve tactical purposes such as planning local law enforcement activities, or strategic ones such as resource allocation. When necessary, meetings can bring in government agencies or industry representatives. It would be logical for the MPS to use IIDs to share assessments and coordinate strategies with China’s other security agencies, such as the Ministry of State Security (MSS) or the quasi-military Peoples’ Armed Police (PAP). Cross-agency sharing of this type would serve the broader objective of improving cooperation within China’s intelligence community, which is an official requirement of the 2017 National Intelligence Law (, 27 June).


The result of China’s new public security intelligence system is that the MPS is more tactically and strategically aware of existing threats to the party-state, as well as those over the horizon. Centralizing authority over intelligence work allows more effective oversight and ensures that every department and jurisdiction has access to the same information and quality of analysis. Finally, the emergence of public security intelligence as a standalone discipline will allow the MPS to stand tall in China’s intelligence community. After a checkered history of intelligence work, it appears as though the pressure of informatization on policing structures has reshaped the MPS into a competent intelligence organization. In the future, an improved intelligence capability may strengthen the MPS’ policy clout among China’s leaders, and possibly increase its involvement in national security  work, which has traditionally been monopolized by the MSS and PLA.


  1. Ma, Dehui. 2015. “中国公安情报学的兴起与发展”, 情报杂志, 34 (11), p. 8.
  2. Peng, Zhihui. 2016. “论大数据下公安情报流程的优化”, 情报杂志, 35(4), p. 16.; Lü, Xuemei. 2015. “公安综合情报部门的发展困境与战略转向”, 情报杂志, 34 (6), p. 18.
  3. Wang, Yichen, 2010. “公安情报共享的障碍与解决对策”, 公安研究, (12), p.81, p. 77.
  4. Ma, Wenhai. 2009. “以信息化改革和创新警务机制探析”,公安研究, (10), p. 50.
  5. Peng, p. 16.; Lü, p. 16; Sun, Xiaowei. 2010. “综合情报信息机构设置问题探讨”, 公安研究, (190) 8, p. 76.