In the latest round of judicial reform of Chinese courts (from 2014 to 2017), China’s Supreme People’s Court (SPC) has been promoting the system of “similar judgments for similar cases,” in order to ensure the effective supervision of trial activities. The system of similar judgments for similar cases mentioned by the SPC means that judgment criteria should be consistent between a case that a judge is trying now and previous cases that have been concluded by the court concerned and the court at a higher level or other similar cases with guiding significance. The SPC hopes to achieve the similar judgments for similar cases goal through artificial intelligence (AI) technology.
An Alternative Mechanism for Supervising Judges in Chinese Courts
The system of similar judgments for similar cases is one piece of the Judicial Accountability System, the core of the latest judicial reform of Chinese courts. In the relevant reform of the Judicial Accountability System, the SPC is exploring how to supervise judges in the process of hearing cases.
Prior to this, Chinese courts had long implemented the Review and Approval of Judgment System, under which judges had to undergo review and approval by their superiors before rendering a judgment. However, Chinese courts understand that this practice will render judges open to inappropriately interference by others when making judgments. Therefore, in the reform of the Judicial Accountability System, the SPC canceled the Review and Approval of Judgment System and emphasized that “he who hears the case shall make the judgment and be liable therefore” (让审理者裁判、由裁判者负责).
However, the SPC is also worried that if judges make judgments independently and are not subject to any supervision, there will be inconsistencies in judgment criteria among various judges. This situation will not be conducive to a uniform application of the law. In fact, some local courts that have taken the lead in canceling the Review and Approval of Judgment System have already encountered such problems.
The SPC believes that while canceling the Review and Approval of Judgment System, it is also necessary to find an alternative mechanism for supervising judges. The SPC’s solution is the similar judgments for similar cases system. In the “Several Opinions of the Supreme People’s Court on Improving the Judicial Accountability System,” the SPC requires that each court shall unify the judge’s judgment criteria by referencing similar cases. If a judge does render a judgment conflicting with the judgment of a similar case already rendered by the court concerned or the court at a higher level, then the president of the court and the director of the division where the judge concerned works should initiate a supervision mechanism.
Therefore, the implementation of similar judgments for similar cases has become an important part of the reform of the Judicial Accountability System.
The Role of Artificial Intelligence
In the “Opinions of the Supreme People’s Court on Implementing the Judicial Accountability System and Improving the Trial Supervision and Management Mechanism (For Trial Implementation),” the SPC requires that judges must conduct a search for similar cases and related cases before making a judgment to ensure that the judgment criteria of similar cases are consistent.
This practice is called the “Compulsory Similar Cases Search and Reporting Mechanism” (类案强制检索与报告机制). The SPC is considering adopting this mechanism because of its confidence in the big data and artificial intelligence (AI) technologies of Chinese courts.
First, Chinese courts have made great progress in big data technology. Chinese courts have incorporated all the judgments throughout the country into a database, and are now digitizing the case files, which will facilitate the judge’s search.
Second, Chinese courts have tried to use AI techniques to assist and supervise judges. On the one hand, many local courts in China are developing a “similar cases pushing” system based on this database, which can push the judgments of similar cases to judges for reference. On the other hand, some courts have tried to develop an “abnormal judgment warning” function based on this database — that is, if a judgment significantly differs from the judgments of similar cases, the system will automatically send a warning to the judge’s superiors, prompting them to initiate a supervision mechanism on the judge concerned. At present, this function is mainly used in criminal cases to monitor whether the judge’s sentencing is reasonable.
Feedback From Chinese Judges
At present, some companies and some local courts in China are developing such “similar judgments for similar cases” AI systems. They all hope that their systems will be procured by more local courts in the future, or even be purchased by the SPC and then provided to all courts in the country. Therefore, there has been fierce competition among them.
Based on my communication with some judges who use these systems, their feedback is mainly as follows:
First, the “similar cases” pushed by these systems are not precise enough, and the similarities with the cases that the judge is trying are not enough. That means there is not significant reference value for the judges.
Second, for experienced judges, such systems are not necessary. However, inexperienced judges are willing to learn how to hear cases through the AI system. Therefore, this system is suitable for the training of new judges.
Third, the AI technology that judges need most is the “automatic generation of judgments”: the AI system reads the case materials, extracts key information from them, and then automatically generates judgments based on the judgment criteria of similar cases. Given the fact that Chinese judges are trapped in the midst of a litigation explosion – a direct cause of the surge in the workload of judges in recent years – the courts urgently need to use this technology to reduce the workload of judges. At present, some courts in Fujian province have tried to use this technology to deal with certain disputes with similar circumstances in batch, such as credit card debts.
Although the AI technology of Chinese courts is still very immature, the current needs of China’s judicial reform, such as “similar judgments for similar cases” and the litigation explosion, force the court to consider resorting to AI technology. Therefore, Chinese courts will continue to increase their attempts in this regard in the future, and keep exploring the future of AI technology in Chinese courts.
This article is written by Meng Yu and Guodong Du, co-founders of the China Justice Observer (CJO). The CJO is committed to present the real Chinese judicial system. Meng Yu is a researcher from China University of Political Science and Law and Guodong Du is an attorney of Dentons Beijing Office.