The value and application paradigm of the construction of a general model for legal supervision of big data for administrative prosecution
Cui Qinglin Yan Hui
Digital prosecution is an important support for the modernization of procuratorial work. In the context of the digital era, administrative procuratorial supervision and case handling should further promote the procuratorial big data strategy, promote the transformation of the case-handling mode from "case-based and quantity-driven" to "similar case-based and data-empowered", discover governance loopholes or supervision clues through data analysis, data collision, and data mining, and actively perform administrative procuratorial supervision duties in accordance with the law, and help promote the modernization of the national governance system and governance capacity.
The big data legal supervision model helps improve the quality and efficiency of traditional work
The first is to promote the efficiency of administrative procuratorial supervision. Administrative procuratorial supervision work involves multiple parties such as courts, administrative organs, and parties to the case, and the interests and demands are diverse, and accurately and efficiently verifying whether administrative entities have enforcement and punishment powers over relevant matters, and judging whether administrative organs' application to the court for compulsory enforcement is reasonable and lawful, will affect the quality and effectiveness of administrative procuratorial supervision. Big data technology should be fully used to integrate data resources from all links and departments of administrative law enforcement, to achieve traceability and queryability of processes, to ensure that oversight is objective, truthful, complete, and legal, and to provide new paths for improving the quality and efficiency of procuratorial case handling.
The second is to promote the standardization of administrative procuratorial supervision, unify the application of law, and improve the social guidance effect of administrative procuratorial supervision. In judicial practice, reasons such as uneven personnel quality and imperfect supervision and restraint mechanisms have affected the high-quality and efficient development of administrative procuratorial supervision. Relying on big data retrieval, analysis, and processing, procuratorial organs can more accurately discover problems such as inconsistent application of law in judicial and law enforcement activities, and the construction of a big data legal supervision model can realize the push of similar cases and accurate legal application suggestions through model comparison, data screening, and early warning of abnormal data, help procuratorial organs standardize the performance of administrative procuratorial supervision functions, and improve the quality and efficiency of procuratorial case handling.
The third is to promote the refinement of administrative procuratorial supervision and improve the methods of collecting evidence. Administrative procuratorial supervision often requires procuratorial organs to extract relevant evidence from different administrative departments, the sources of evidence are different, the data structure is not uniform, and the electronic evidence after evidence collection also needs to be reviewed and fixed by the procuratorial organs, affecting the efficiency of administrative procuratorial supervision. Big data technology can quickly process evidence from different data sources and different data structures to achieve accurate and rapid evidence collection. Big data electronic evidence collection can also leave traces throughout the process, forming a complete, closed, and traceable electronic evidence chain to ensure that administrative procuratorial supervision is legal and effective.
The big data legal supervision model helps the paradigm shift of administrative procuratorial supervision
First, the concept of procuratorial has been upgraded, extending from litigation supervision to social governance. The path of big data legal supervision is to summarize the rules and characteristics from individual cases, and then screen out similar cases in massive data according to the laws and characteristics, and find problems in legislation, law enforcement, judiciary, mechanisms and other aspects in batch cases, and finally solve social governance problems. In this process, procuratorial organs can take advantage of the big data of the litigation and investigation center to expand the channels of case leads for legal supervision, integrate into the overall pattern of party committees, governments and social governance, promote the docking of procuratorial investigation and joint mediation, resolve social contradictions through multiple channels and means, resolve contradictions and disputes at the grassroots level, and create harmony and stability at the grassroots level.
The second is the iteration of the work paradigm, from passive supervision to active supervision. In the past, administrative procuratorial supervision did not have sufficient grasp of the sources of information and illegal leads of administrative acts and trial activities, and it was difficult for procuratorial organs to fully grasp supervision information and accurately discover clues to problems, resulting in the timeliness of supervision work being weak. In contrast, big data technology can overcome the limitations of personnel experience and physical distance with the help of Internet means, and actively discover clues of violations and crimes through the collection, screening, and analysis of exponential and incomplete data volumes, and on this basis, realize active supervision and multiple coordination through methods such as investigation and evidence collection, active guidance and investigation, and complete the transformation from passive supervision to active supervision.
The third is to expand the volume of supervision, changing from case-by-case supervision to similar case supervision. With the support of big data technology, the procuratorate can process all the data related to a certain phenomenon without relying on random sampling or using random analysis methods. In other words, big data can see more clearly the details that the sample itself cannot reveal. Procuratorial organs use big data to expand supervision channels, and after collecting data and cleaning data, they must conduct statistical analysis of the data to visualize the results. In this process, the mode of legal supervision has changed, from the traditional case-specific supervision to the supervision of similar cases based on the full sample of big data.
Key principles of the big data legal surveillance model
Although the legal supervision model of administrative procuratorial big data is not artificial intelligence in the traditional sense, its construction principle is closely related to the above three elements, and the cornerstone of which is data, so special attention should be paid to the cornerstone utility of procuratorial big data. To build a legal supervision model for administrative procuratorial big data, three key issues should be solved, one is the source channel of data, the other is what kind of data is collected, and the third is how to use data. Returning to the specific scenario of administrative procuratorial supervision, the following three steps can be summarized:
First, build a big data supervision platform for procuratorial data and form a data pool. On the one hand, expand and extend the source channels of data, such as the national procuratorial business application system, the provincial and municipal procuratorial data application platform, the China Judgment Document Network, the administrative law enforcement information disclosure platform, the enforcement information disclosure network, the trial information network, the Chinese government website, domestic banks, citizen hotlines, and the message area of media news websites; On the other hand, the first layer of data collection of judgment documents is formed by setting the filter conditions of region and time, and then the second layer of data collection is formed after adding subject information or other keywords to filter.
Second, data is classified according to the business scenario to form packets. The same platform, different supervision models need to collect different information, such as the same administrative service platform, can collect the transportation company's vehicle registration information, "Road Transport Certificate", "Road Transport Business License" and other information, can also mine illegal oversized transport company, vehicle and driver information, and correlate the above data related to business scenarios, through the association data collection to obtain the cleansed data package.
Finally, effectively carry out big data collision and find the direction of accurate supervision. Taking the digital supervision model of "administrative organs illegally exercising their functions and powers of administrative fines" in a certain place as an example, the model analyzes the collision data of administrative organ fines, the collection of fines and forfeitures by the Finance Bureau, and the enforcement of administrative non-litigation by courts, so as to dig out illegal clues such as the fine decision has been made but not paid or not paid in full, the fine has not been paid nor applied for enforcement, and the court has approved enforcement but has not been transferred for enforcement. Specifically, the type of administrative procuratorial supervision is first selected, the rules for research and judgment are designed, and the relevant procedural supervision rules and substantive supervision rules are taken into account, and then data collisions are carried out, data packets are compared with rules, and abnormal clues are obtained through data analysis.
Construction of a general model for legal supervision of big data for administrative prosecution
Identify supervised model kinds and scenarios. The legal supervision model of big data for administrative inspection is mainly divided into risk data supervision model, demand data supervision model and effect data supervision model. The data supervision model originates from the need for procuratorial organs to carry out legal supervision activities in practice, so the supervision model must be designed according to specific application scenarios and practical needs. The application scenarios of the big data legal supervision model for administrative prosecution include administrative enforcement supervision (including non-litigation enforcement), administrative judgment supervision, administrative illegal conduct supervision, and administrative adjudication personnel illegal conduct supervision.
Develop anomaly case detection models. The core key of big data strategy is the big data legal supervision model. Some scholars have identified the path of big data empowerment of legal supervision clue discovery as "analysis of typical cases→ sorting out the law of case occurrence→ data sharing and aggregation→ data collision comparison, → research and judgment of clues in similar cases→ transfer lead verification→ carrying out precise supervision→ tracking and supervising implementation→ and promoting social governance." Among them, the sorting out of the law of case occurrence is the key node of the big data legal supervision model, which must extract the characteristics of abnormal cases and realize them through data.
Enhance the use of human-machine coupling methods. Specifically, by setting thresholds, system early warnings, and pushing case-handling guidelines based on algorithms, procurators are guided to improve verification and confirm abnormal leads through file verification, exchange feedback, and on-site visits. After all, the analysis results of big data do not reflect the uniqueness of the case, and prosecutors need to participate in the intervention, cooperate with the machine, analyze and compare the data in real time, and timely identify and consider the abnormal and small probability situations in batch cases, so as to better improve the quality of supervision data, optimize the self-learning of machines, and ensure the accuracy and efficiency of the model.
(The authors are the chief procurator of the People's Procuratorate of Dali Bai Autonomous Prefecture in Yunnan Province and the researcher of the Intelligent Judicial Research Center of Southwest University of Political Science and Law)
Source: Procuratorial Daily