Automating Intervention in Chinese Justice: Smart Courts and Supervision Reform

Abstract

This article examines how smart courts enhance the reform of judicial responsibility and the “trial supervision and management” mechanism. It holds that smart courts, while meant to provide better judicial services and improve access to justice, have the additional goal of enhancing the restructuring of accountability and power structures. It argues that automation and digitization help institutionalize and codify political supervision. Smart courts help resolve tension between the two opposing requirements of Chinese courts to maintain legal rationality and independent adjudication on the one hand, and the need for flexibility to allow intervention on the other. This article provides an account of the automation of “trial supervision and management” and explores the role of technology in enhancing political intervention in China’s legal system. This investigation draws on internal court reports and central and local judicial documents, supplemented with a review of Chinese empirical scholarship.

1. Introduction

Chinese courts are increasingly automating and digitizing their judicial process by integrating information technology (IT) in all aspects of their work. This “judicial informatization” (sifa xinxihua, 司法信息化) has been ongoing for two decades but has accelerated under the new judicial reform agenda, formulated in 2014 by the Central Committee (“Fourth Plenum Decision”). ¹ As part of this agenda, the judiciary of the People’s Republic of China (PRC) has accelerated the mass digitization of its procedures and archives of court judgments and introduced live-broadcasting and online video depositories of trial hearings. ²

This virtual judicial environment is part of the foundation for what is officially called the policy of “building smart courts” (jianshe zhihui fayuan, 建设智慧法院). It envisions a fully digital judicial process in which judicial officers use technological applications sustained by self-learning algorithms, where most tasks are automated, with minimal opportunities for human discretion or interference. Smart courts have also developed parallel with other reforms to strengthen hierarchical control over the judiciary and judicial formalism. ³ These reforms, such as the guiding cases system, ⁴ the centralization of budgetary authorities of local courts, ⁵ and the introduction of circuit courts, ⁶ have brought the local judiciary closer to the centre.

This article argues that digitization and automation of courts are meant to enhance the restructuring of accountability and control structures by guaranteeing stricter procedural compliance and facilitating vertical control over adjudication power. To this end, this article provides a first account of two smart court reforms that are meant to enhance two judicial reforms in particular, namely the judicial responsibility reform and the reform of “trial supervision and management” (shenpan jiandu guanli, 审判监督管理). ⁷ Through these case-studies, the article shows technology is envisioned as a pathway to resolve the tension between two opposing requirements of China’s judiciary. It shows how digitization and automation are perceived and operationalized in the unification of the normative and prerogative state.

This article conceptualizes the supervision mechanism as a channel through which the prerogative state can exercise its sovereign power in adjudication. China’s legal system has developed into a dual system in which a prerogative state exists that rules according to political priorities but leaves conventional matters to legal rules. ⁸ However, the Chinese party-state needs to be able to suspend legal rationality at any time to ensure that it can intervene in the legal system whenever its bottom line is in jeopardy by real or imagined threats. ⁹ “Trial supervision and management” is one such internal mechanism that allows this.

Smart courts help resolve the tension between these two opposed requirements of China’s judiciary. The digitization and automation of justice are envisioned as being able to enhance legal rationality and independent adjudication, while simultaneously leaving enough discretionary space for political intervention that the central party-state considers appropriate. The judiciary sees technology as a key pathway to solve this tension between the contradicting need for legal rationality and independent adjudication on the one hand and the need for flexibility to allow party intervention on the other.

The scholarship has not yet fully explored how smart courts are envisioned as a pathway towards resolving this contradictory duality. Earlier law and technology literature has already underscored the centralizing potential of technology as automation and digitization provide enhanced control mechanisms for those in power. ¹⁰ The literature has already discussed smart courts’ policy context, legality, and functional purposes. ¹¹ In line with this article’s premise, one academic publication has argued that smart courts illustrate how Chinese government agencies strengthen social control and how automation and digitization play a crucial role in centralizing judicial power. ¹² However, none of the above-cited literature examines these developments in parallel with specific judicial reforms.

In addition, when it comes to the automation of justice, other countries in the region are lagging behind. Many of these countries have yet to progress beyond the digitization of their judicial services and are far from the comprehensive systems that Chinese courts have established. For example, Indonesia has recently set up its e-court system, ¹³ whereas Japan only began experimenting with e-courts in 2019. ¹⁴ In Vietnam, e-courts and online litigation were virtually non-existent before the COVID-19 pandemic. ¹⁵ China is at the vanguard of algorithmic justice and governance within Asia and globally. Therefore, this article provides a valuable stepping stone for future academic inquiries into the effects of the increasing ubiquity of advanced IT, algorithm-based applications, and artificial intelligence (AI) in both administration and adjudication.

The rest of this article is organized as follows. First, the article discusses the data and methods. The two case-studies were taken from the 2020 and 2021 China Court Informatisation Development Report (“the Report”). ¹⁶ The Report contains multiple “model” cases of smart court initiatives. These model cases were chosen because the judiciary deemed them important and successful enough to include them. Model cases are meant to bring together disparate practices into a unified national approach to court work. Therefore, they provide excellent material for case-studies.

Second, the article discusses the dual character of Chinese courts. First, it examines how courts should deliver judicial services while simultaneously adhering to their political imperative of guaranteeing social stability. Furthermore, it describes the role of political supervision of justice in maintaining this dual character. Finally, through an analysis of new procedural rules, it examines how supervision has changed since the judicial responsibility reform.

Third, the article provides a brief overview of the general smart court reform. It covers the different policy documents that have guided judicial informatization since 2016 and discusses its principles and goals. It then moves on to the two case-studies. The first case-study is a trial e-management platform, designed and implemented by the Jiangxi Provincial High Court. The second case-study is the full process automated supervision platform, designed and implemented by the Yibin Intermediate Court. The article takes them as paradigmatic, representative cases of different levels of China’s court system to provide a more complete picture of smart court development in China.

Fourth, the article discusses the cases in comparison: the two case-studies are examples of how technology is envisioned and operationalized to create an iron cage around the judicial process. China’s legal system has previously been described as a “bird in a cage.” ¹⁷ With smart courts, the bars of the cage have been reinforced with technology, allowing an all-encompassing surveillance of judicial behaviour.

Finally, this article concludes that automating the “trial supervision and management” mechanism allows court officials to detect specific cases requiring intervention more systematically without compromising the normative system. Additionally, smart court reform is an example of the Chinese party-state’s firm belief in the power of algorithmic technology to resolve inherent tensions in the legal and governance system without fundamentally altering it. Using technology in governance and justice reform is considered a facilitator for improving governance structures while maintaining and enhancing party dominance. In this sense, technology enhances authoritarian legality by allowing the prerogative state to exercise its sovereign power better and more efficiently through the legal system without undermining legal rationality.

2. Data and methods

In this article, I examine how digitization and automation are perceived and operationalized in judicial reform through case-study research. I base my case-studies on court reports about court informatization initiatives that focus on trial supervision and management. These reports were published in the Report, which is a bundle of research reports that evaluate the status quo and summarize experiences and achievements of different court initiatives across the country. They are the most up-to-date official reports on the development of smart courts. These reports are insightful objects of analysis because they give us an accepted official reading and evaluation of the status quo and lay out the path ahead for future developments. The practice of “summarizing experiences” has had a long history in communist policy-legal rhetoric since the establishment of the PRC. It is meant to bring together disparate practices into a unified national approach to court work. ¹⁸ These reports were supplemented by news articles, official press releases by courts, and official implementation measures. The analysis has also benefitted from informal discussions with experts on Chinese courts and judicial reform.

The first case-study is the Jiangxi Provincial High People’s Court’s (hereinafter “Jiangxi High Court”) “Trial e-Management” Platform (“shenpan e-guanli” pingtai, “审判e-管理”平台) from the 2021 report. The Jiangxi High Court is one of two high courts in China that has been covered in the Report for five consecutive years. The trial management platform is part of its e-series—a series of initiatives related to automating and digitizing the judicial process. For example, the 2020 report covers its e-assistant judge platform—an artificial intelligent assisted case-handling system that helps judges prepare and adjudicate cases.

The second case-study is the Sichuan Province, Yibin City Intermediate People’s Court’s (hereinafter “Yibin Court”) “Entire Court, Entire Staff, Entire Process” Supervision Platform (“quanyuan quanyuan quancheng” jianguan pingtai, “全员全院全程”监管平台) from the 2020 report. In addition, multiple other provincial courts have developed similar platforms, which were featured on a list of “example cases of judicial reform in people’s courts,” published occasionally by the Supreme People’s Court’s (SPC) Judicial Reform Leading Small Group to share experiences. Therefore, while the procedures and technology described are from the Yibin Court, they also apply to several other provinces’ courts.

I chose these cases because of their explicit supervisory intentions and because the reports directly refer to the above-mentioned opinions on the judicial responsibility system as part of their regulatory context. I decided to choose two cases from different levels of the court system to be able to give a more complete picture of how smart court reform is perceived and designed as part of judicial reform. The point is not necessarily to theoretically compare them, but rather to achieve a more in-depth and complete understanding of new developments.

The choice of only two case-studies might limit the representatives of the research outcomes. However, the case-studies presented in the reports are a selection of that year’s most promising smart court initiatives, meant to contribute their experience to future smart court reform. Therefore, the case-studies are “model cases.” The judiciary deems them important and successful cases that are meant to set an example and guide future reforms in other courts. Moreover, based on these examples, multiple other provincial courts have also introduced similar digital platforms. Therefore, these case-studies should be considered paradigmatic cases that highlight characteristics of digitization and automation in Chinese courts. ¹⁹ The article does not examine how judicial officers are using these systems and how this influences the operation of courts “in action.” Rather, the case-studies are representative of how the judiciary perceives digitization and automation, and how the judiciary wants digitization and automation to operate as part of judicial reform.

Lastly, these reports are written by judicial officials for officials, favouring the reliability and accuracy of the assessments and descriptions made in the report. Therefore, they are significant, not only because they are rich in detailed information on local Chinese policy developments, but also in their role as a reference for future reform. In this sense, these documents not only serve as informants, but can also be considered as actors in their own right because they shape and influence future policy and reform. Therefore, their official character makes them interesting in their own right. ²⁰

The next section discusses the dual character of Chinese courts, how this has influenced their operation, how political supervision remains one of the primary mechanism to maintain courts’ dual character, and how this changed after the judicial responsibility reforms.

3. Chinese courts: between a rock and a hard place

3.1 The dual character of Chinese courts

Chinese legal courts (fayuan, 法院) serve dual functions. On the one hand, they provide normal judicial services, “offering rules-based solutions to a wide range of social [and economic] conflicts.” ²¹ On the other hand, they are tasked with stability maintenance and ensuring the implementation of central policies, making them an integral component of state governance. ²² However, this does not mean that they are passive political actors who do not enjoy autonomy. ²³ To effectively fulfil their dual tasks, courts require substantial discretion and flexibility. It means that the judiciary, as a government institution, must primarily consider the consequences of its decisions for the broader political, economic, administrative, and social context. ²⁴ This duality explains Chinese courts’ preference for discretion and informality. ²⁵ It also means that courts were long vulnerable to local political and economic interests, which led to many corruption scandals. ²⁶

The duality of their institutional character has influenced and guided judicial practice since the establishment of Chinese courts. ²⁷ In the 1980s, courts mainly fulfilled their political task through the administration of criminal justice. ²⁸ Later, as China’s society and its economy became increasingly complex, they became important public spaces in which Chinese citizens could legitimately air their grievances and demands against the state, as well as each other. ²⁹ Therefore, courts are also supposed to use the law to resolve disputes. However, as part of a political system that greatly emphasizes social stability, courts are sensitive to the threat of unrest. This sensitivity has made protesting, petitioning, or simply the threat to do so “a successful means for litigants to pressure courts to rule in their favour or to alter decided cases.” ³⁰

In this sense, there exist extra-legal factors that influence judicial decision-making. Depending on the case, legal technicalities might be of secondary importance. ³¹ Therefore, courts find themselves between a rock and a hard place: ironically, the tension between, on the one hand, fulfilling their political tasks and, on the other, using the law undermines their own governance capacity and popular confidence in and support of courts. ³² The dynamics of this precarious balance are especially visible in administrative litigation, ³³ in which courts are caught between stronger administrative agencies and the law, and criminal justice, in which principles of due process are sacrificed for swift and severe justice. ³⁴ Similar dynamics exist in civil and commercial law but are less pronounced. ³⁵

In sum, Chinese courts exist to serve both the prerogative and normative state. The Chinese party-state has created a system of “authoritarian legality” in which resolving conflicts and disputes that have no substantial impact on the Chinese party-state’s bottom line is outsourced to Chinese courts, while the party-state maintains the prerogative to intervene when its interests are at stake. ³⁶ In other words, legality only exists in clearly demarcated zones and is “intentionally subordinate to the exercise of power by the party-state.” ³⁷ Therefore, China’s legal system contains “zones of lawlessness” to enable the prerogative state. These zones exist entirely outside the reach of legal and procedural guarantees to protect sovereign power from any alleged or actual threats. Furthermore, judicial mechanisms and procedures allow the suspension of legal and procedural rights to create these zones, making them a systemic feature of China’s legal system and court practice. ³⁸

One such mechanism designed for this purpose is the “trial supervision and management mechanism” (shenpan jiandu guanli jizhi, 审判监督管理机制). The mechanism functions as a procedural entry point that allows the prerogative state to penetrate the legal system and exercise its power without undermining it. It is important to note that this is an internal judicial mechanism and does not necessarily involve any interaction between political and judicial actors. The next sections will first detail the ideology behind political supervision of the judiciary, and then explain the supervision mechanism and how it has changed as part of the judicial responsibility reform since 2014.

3.2 Supervision of justice in China

The notion of supervision as crucial to the administration of justice is born out of China’s Marxist-Leninist-Maoist approach to justice, which advocates the “mass line” approach. This approach is characterized by informality and particularism, favouring mediation and settlement over litigation. ³⁹ The party-state must have the ability to intervene and “correct errors” in judicial decisions when citizens petition to challenge final judgments. According to this approach, a judgment’s political correctness is more important than preserving its finality. ⁴⁰ Especially in the early stages of China’s legal development, it was a “pragmatic response to the problem of how to overcome shortcomings in the judiciary at an early stage of institutional development.” ⁴¹

From this ideological perspective, supervision is a so-called “last line of defence” to ensure social fairness and justice, guarantee trial efficacy, standardize judicial behaviour, prevent judicial corruption, and overcome judicial unfairness. ⁴² Additionally, it serves as an institutional check on individual judges and offers the required flexibility to bring decisions in line with external policies of the central government. The idea is that court leaders (i.e. the court (vice) president, division (vice) chief), being senior court members with political responsibilities, have a better idea of how to decide sensitive cases. However, this approach to justice also means that political officials within and outside courts could involve themselves in cases. This phenomenon was especially prevalent under Hu Jintao, where party-state officials would pressure courts to change decisions to appease the public. ⁴³ Nonetheless, the reforms under the Xi Jinping administration have reduced this phenomenon. ⁴⁴

Many systems and procedures exist to supervise the judiciary and control judicial decisions. Appeal and retrial procedures, a fundamental aspect of any judicial system, are also considered a way to “supervise” trial work in China. ⁴⁵ Prior to the judicial responsibility reforms, court leaders primarily supervised through the vertical hierarchy of the case-approval system. Chinese courts operate highly similar to a bureaucracy, with a clear hierarchy of command through a three-tiered vertical hierarchy: the collegiate panel at the lowest level, which functions as the basic adjudication unit. ⁴⁶ A front-line judge, often a junior in rank, is expected to adjudicate routine cases independently but report more complex issues to the presiding judge. The presiding judge will report the case to the next tier for potentially more serious cases, namely the divisional meeting. At the meeting, both the division chief and deputy chief and more experienced judges may give their input on how to decide the case. These meetings also play an important role in determining whether cases will be reported up the chain to the higher decision-making body within a court, namely the adjudication committee. ⁴⁷ Especially when senior staff anticipate a social controversy surrounding the case, the decision might become more political than purely legal. ⁴⁸ This explanation is not exhaustive but gives a basic understanding of how supervision worked in courts under the case-approval system. The above-described hierarchy also exists between a lower-ranked court and a higher-ranked court, which can guide and co-ordinate with the lower-ranked court to decide a case. ⁴⁹

Therefore, supervision in China’s judicial system refers to both internal, administrative mechanisms in a legal court that allow court leaders to monitor and guide trial activities when necessary and supervision by higher-ranked courts. It does not necessarily involve interaction between party officials and judges. However, court leaders are also party members and take political considerations into account when supervising sensitive cases because they bear political responsibilities. ⁵⁰ In this sense, China’s judiciary is firmly embedded in its political system. ⁵¹

It bears reminding that the type of influence exercised through this mechanism is “allowed, preferred, and even urged by the regime.” ⁵² It implies that an individual judge should decide cases without interference from government agencies, social groups, and other individuals, but not necessarily from their collegiate panel or adjudication committee. ⁵³ However, it also bears repeating that “a great majority of the cases handled by the Chinese courts are routine and straightforward.” ⁵⁴ A single judge generally tries these cases, and the decision is final.

Nonetheless, due to its highly opaque and discretionary character, the case-approval system was extremely susceptible to abuse, which led to widespread corruption, unaccountability, and unlawful interference in adjudication. ⁵⁵ As a result, the judiciary launched a series of reforms to address this abuse and change how responsibility was assigned (“judicial responsibility reforms”). ⁵⁶ These reforms included (1) the deprivation of power from court leaders to review and sign off on decisions drafted by front-line judges (encapsulated in the slogan: “letting those who adjudicate judge, and making those who adjudicate responsible,” rang shenglizhe panjue, rang panjuezhe zeren, 让审理者裁判, 让裁判者负责), effectively ending the case-approval system; (2) reducing the power of adjudication committees; (3) establishing “professional judges meetings” ⁵⁷ to provide the adjudicating judge or collegiate panel non-binding advice for a case; (4) making the recording of intervention mandatory; (5) keeping judges accountable for life for the quality of the cases they adjudicated. ⁵⁸

Empirical research has found that this reform has significantly strengthened individual judges’ autonomy and accountability, substantially eliminating opportunities for corruption and judges’ means to shirk responsibility. The main reason for this is that individual judges are now responsible for life, so they are more resistant to court leaders’ intervention. Additionally, the requirement to record interventions has made court leaders think twice about intervening because it is difficult to know how this will reflect on their performance evaluation. Leaving a record of intervention could potentially destroy their career prospects. ⁵⁹ Another reason for this reform’s success is the increased consolidation of political power under Xi Jinping, incentivizing local institutions to follow central directives better and implement reforms more diligently. Without it, these reforms would probably not have been as successful. ⁶⁰

3.3 Changes to supervision since the judicial responsibility reform

Despite its success, the judicial responsibility reform also generated several issues regarding supervision. It is unsurprising, as the aims of the judicial responsibility reform and supervision fundamentally contradict each other. The initial document launching this reform captures this contradiction. On the one hand, it gives adjudicating judges the authority to issue judgments independently, without needing the approval of their superior, and prohibits court leaders from reviewing and issuing judgments on cases in which they did not participate. ⁶¹ On the other hand, it asks court leaders to diligently fulfil their “trial supervision and management” duties. It gives court leaders “the right to request the individual judge or collegiate panel to report the progress of a case and the results of the review” in the following circumstances: (1) cases that involve group disputes that may affect social stability; (2) cases that are difficult and complicated, and have a significant impact on society; (3) cases that may conflict with a decision of the court or a higher-ranked court; and (4) cases that involve reports of violations by the adjudicating judge. These cases are called the “Four Types of Cases” (“Four Types”) (silei anjian, 四类案件). Where a case gets identified as either of the four types above, court leaders may ask the adjudicating judge to report on this case. Where court leaders disagree on the trial process with the adjudicating judge, they must submit the case to a professional judges’ meeting or adjudication committee for discussion. Any advice is non-binding and must be recorded and included in the case file. ⁶²

In other words, the previous case-approval system gave court leaders substantial discretion in deciding the outcome of a case because the judgment was not valid until they signed off on it. With the responsibility reforms, court leaders lost this power. However, where a case is deemed sensitive (based on the criteria stipulated above), court leaders now have to get involved in the process and keep up to date with the case’s progress. Additionally, they can no longer give oral instructions individually, but rather via a professional judges’ meeting or the adjudication committee. Moreover, this advice or guidance is non-binding and must be recorded. This mechanism is what “trial supervision and management” refers to, namely the responsibility to supervise the full process of sensitive cases. ⁶³

This new situation created tensions: “A contradiction emerged between the independent operation of judicial power and the supervision and management of trials.” ⁶⁴ In short, the contradiction lies in the fact that the responsibility reforms gave front-line judges more autonomy in performing their trial responsibilities while one of the primary responsibilities of Chinese court leaders is ensuring the “political correctness” of their subordinates’ work. ⁶⁵ Additionally, the most important goal of the entire judicial reform in the past decade has been to strengthen party leadership and put the party’s political construction first. ⁶⁶ These goals naturally clash with what the new autonomy judges are supposed to enjoy. It has led to inertia among court leaders, as they became unwilling or even afraid to fulfil their supervision duties and responsibilities. According to courts’ research reports and Chinese empirical scholarship, court leaders thought exercising their supervision powers might be considered undue interference. ⁶⁷ As a result, court leaders were concerned that adjudicating judges enjoyed too much power, that this was exacerbating inconsistencies in decisions, and that they were losing control over sensitive cases. ⁶⁸

The primary cause of this contradiction lay with the ambiguity and vagueness of court leaders’ new supervision responsibilities and duties. ⁶⁹ The 2015 Opinion remained vague about the exact content and scope of the sensitive cases mentioned above. It also did not specify the exact measures that court leaders could take as part of their supervision duties and responsibilities. Therefore, the challenge of further reforming judicial responsibility and the “trial supervision and management” mechanism boiled down to “correctly dealing with the relationship between the delegation of [adjudication] power and supervision.” ⁷⁰

To address this, the SPC issued a series of subsequent documents to attempt to refine and clarify the different responsibilities and duties for all court staff. ⁷¹ They also provided lists of powers and responsibilities for court leaders, specifically stating that performing these duties when following procedures and within the scope of the list would not constitute undue interference. ⁷² It was supposed to reassure court leaders that their intervention was not illegal and, therefore, punishable if they followed correct procedures. These central documents also tasked local courts with formulating detailed rules to further demarcate court leaders’ exact powers and responsibilities.

In response, many different courts issued various implementation measures, regulations, and checklists to flesh out court leaders’ supervision duties and responsibilities. For example, in 2019, the Shenzhen Intermediate People’s Court issued a series of “trial supervision and management checklists” that detail the exact boundaries and scope of court leaders’ supervision responsibilities to ensure that they “accurately grasp the boundaries and content of supervision and management, and eliminate the worry of court leaders that supervision is exercised without legal authority.” ⁷³

In 2018, the Hunan Provincial High Court issued the Measures to Further Implement Court Leaders’ Trial Supervision and Management Duties (for Trial Implementation), refining the exact content and scope of the “Four Types of Cases,” as well as the measures a court leader may take. Similarly, in 2019, the Jilin Provincial High Court issued the Measures to Standardise Court Leaders’ Trial Supervision and Management Duties (for Trial Implementation), stipulating supervision measures and prohibiting other actions for court leaders. ⁷⁴ Also, in 2018, the Ningxia Autonomous Region High Court issued the Measures to Further Implement Court Leaders’ Trial Supervision and Management Duties (for Trial Implementation) to do the same. ⁷⁵

Finally, in November 2021, the SPC issued national clarifications, taking steps towards institutionalizing and codifying the practice. ⁷⁶ Article 10 of the Guiding Opinions on the Supervision and Management of “Four Types of Cases” shows the options available. Among others, a court leader may: request the judge or collegiate panel to report on the progress of the case; review the case material; attend and observe court hearings; request the individual judge or collegiate panel to reconsider their judgment and report on it; decide to submit the case to a professional judges’ meeting or adjudication committee for discussion; decide to report a higher-ranked court, and any other necessary supervision and management measure appropriate to the court leader’s position. The court leaders are required to exercise the above measures within the scope of their authority, “which does not constitute interference or prying into cases in violation of provisions.” ⁷⁷

It is the first time a central document has stipulated the exact measures that constitute “trial supervision and management” under the new responsibility system. Additionally, it changed the meaning and scope of the “Four Types,” elaborating each type of case. ⁷⁸ Other courts now have to amend their implementation measures to comply with this central Guiding Opinion but do get a certain margin to adjust to local conditions. ⁷⁹ The Guiding Opinion stays silent on what cases trigger what specific measures so, in reality, it is likely a court leader remains free to choose whichever measure he prefers to fulfil their supervision duties. The last measure also leaves some discretion to court leaders in deciding how to exercise their supervision powers exactly.

Therefore, although the Guiding Opinion makes substantial progress in improving the “trial supervision and management” mechanism, some questions remain unanswered. For example, the Guiding Opinion did not incorporate a list of prohibited actions present in the Jilin High Court measures. In addition, it does not explicitly prohibit court leaders from giving oral instructions to adjudicating judges, which was the primary exercise of influence over cases. Neither does the Guiding Opinion specify the procedures for recording the supervision. Article 7 also allows court leaders to apply “trial supervision and management” measures to specific cases outside the scope of the “Four Types,” even including the phrase “or where it is otherwise necessary to apply [these measures].” In short, the Guiding Opinion leaves some crucial aspects unaddressed, and court leaders maintain a certain degree of discretion. As one observer commented, as long as there is no complete set of procedures for the entire system of supervision, guidance, and review of cases, it remains unstable. ⁸⁰

Nonetheless, albeit incomplete, these developments point toward institutionalizing and codifying the “trial supervision and management” mechanism. Chinese scholars claim that these reforms have turned the supervision mechanism, previously highly opaque and discretionary, into a transparent and rule-bound process with clear delineation between adjudication powers on the one hand and supervision powers on the other. ⁸¹ Other researchers remain sceptical. For example, the rules for “trial supervision and management” obligate court leaders to record their intervention and include it in the case file when performing their supervision duties. ⁸² However, they find that, lawful or not, court leaders in their surveyed courts avoid registering the intervention. The reason for this is that court leaders and political leaders remain powerful and have large discretion over the career trajectory of judges, so they have the leverage to dissuade an adjudicating judge from recording the intervention. Moreover, the recording of intervention may trigger a formal investigation and jeopardize the careers of everyone involved. ⁸³

Despite this, the mere existence of this obligation alone has made a difference, compelling court leaders to think twice about intervening. ⁸⁴ Additionally, the SPC gave a list of reasons they issued a Guiding Opinion. They claim that some courts had neglected supervision duties, that some courts inappropriately expanded the “Four Types” scope, and others had reinstated the case-approval system in all but name. They also mention the above-described concerns that court leaders had, such as fear of being seen as illegally interfering with a case. ⁸⁵

From a broader perspective, it is safe to argue that institutionalizing and codifying the “trial supervision and management” mechanism was necessary to maintain the dual character of courts, given the rampant abuse in the pre-2014 reform era. Courts needed to improve independent adjudication while at the same time guaranteeing they were still able to fulfil their political tasks. The key to doing this is to identify cases that require intervention, which is very difficult. Therefore, courts needed to carefully insulate individual judges from outside interference, namely illegitimate and illegal, while also keeping open a window, albeit tightly circumscribed and controlled, to allow the Chinese party-state, represented by court leaders, to intervene whenever they deem their bottom line is affected.

However, as discussed above, Chinese courts still have difficulties in distinguishing this “red line,” and the reforms have not yet fully erased all opportunities for abuse. In this sense, the reforms are still incomplete and might not achieve the desired results. It is why the Chinese judiciary has vested its hopes on IT, algorithm-based applications, and AI, allowing it to strengthen further judicial services and supervision and control. Therefore, the documents analyzed in this section also call upon courts to leverage the power of digital technologies to help enhance and enforce the new reforms. ⁸⁶

The next part analyses two cases in which courts have attempted to do this. The judiciary wants to automate the “trial supervision and management” mechanism through technology to create a tightly guarded “safety valve” that both protects the autonomy of adjudication and facilitates intervention. The automated and digital environment creates a sealed-off corridor through which the prerogative state can exercise its sovereign power whenever it finds this expedient.

4. Smart courts: digital and automated supervision

4.1 Smart courts building

Starting in 2016, the SPC issued a series of documents outlining judicial informatization. The Five-Year Development Plan for the Informatisation of People’s Courts (“Development Plan”), ⁸⁷ the 2016 Opinion on Comprehensively Promoting the Synchronous Generation and In-depth Application of Electronic Archives, and the 2017 Opinion on Accelerating the Building of Smart Courts form the basis for what is now known as smart court reform. ⁸⁸ These documents call for full digitization of court work and judicial processes, going beyond the initial digitization initiatives from 2013. According to these documents, informatization needs to support reform, such as the circuit courts, improving judicial services, judicial responsibility, and supervision.

Courts must conduct the judicial process (i.e. case-docketing, trial preparation, trial hearing, issuing of judgment) digitally to achieve complete recording and real-time dynamic monitoring. Additionally, back-end processes such as trial management, file archiving, serving judicial documents, and so forth must be digitized and automated where possible. It should improve access to justice (people can now file claims and submit cases via the Internet or automated dockets in court halls) and promote uniform adjudication (automated software pushes the applicable law and similar cases as a reference to the case-handling judge when they are preparing a case). It should also strengthen process management and supervision (court leaders can now consult the progress of each subordinate’s workload on a digital interface, the software system sends automated warnings to front-line judges when a case’s deadline is approaching, and so forth). ⁸⁹

The purpose of this two-pronged approach of simultaneous judicial reform and informatization is to achieve effective supervision by building a “systemic iron cage” or “digital iron cage” around the judicial process. ⁹⁰ It reveals how an important goal of smart courts is to create a completely digital environment in which the judicial process takes place, to enable full process traceability and real-time monitoring of the judicial process and judicial behaviour (i.e. a court official can see the progress of a case and see when and who took what procedural step at any given time). As a result, it should make supervision and holding judges accountable easier and more efficient, the judicial process fairer and more transparent, and ultimately improve the judicial system’s integrity.

The SPC’s Opinion on Accelerating the Building of Smart Courts clarifies how technology is envisioned to facilitate better vertical control and formalize supervision. The Opinion asks courts to develop programmes that record all procedural steps of the judicial process, to enable both live and post-facto supervision. ⁹¹ In addition, it commands courts to develop digital trial management platforms that cover the entire process (e.g. digital case-docketing, review, and acceptance, trial preparation such as exchange of evidence via an online platform, issuing digital summons, online trial hearing, drafting and issuing a digital judgment, appeal via an online system, digital archiving), with full traceability, and active supervision. Most Chinese courts have digitized a significant part of their judicial process by now. ⁹²

The next section analyses two “model cases” of such platforms. These cases are used as exemplary models to unify future court reform. Consequently, a substantial number of courts have introduced similar platforms and procedures to automate “trial supervision and management.” ⁹³ Therefore, it is highly likely that these model cases provided guidance for other courts. Moreover, the Jiangxi Provincial High Court introduced its new platform in all courts across the province and the Yibin Intermediate Court did so in all its district courts. Finally, other courts developed similar procedural regulations to codify supervision responsibilities. The fact that the SPC issued a national Opinion on these procedures means that they have now been consolidated and unified: lower-ranked courts have to amend these procedures to fit with the national directives.

4.2 The Jiangxi Provincial High Court Trial e-Management Platform

The “Trial e-Management” Platform (the “Jiangxi Platform”) enables court leaders to supervise remanded and revised cases (fagaizai anjian, 发改再案件), long-term unresolved cases, and the “Four Types of Cases” (silei anjian, 四类案件) on the digital trial management platform. Remanded (fahui, 发回), revised (gaipan, 改判), or retried (zaishen, 再审) cases are legal cases in which a first-instance court has rendered a decision and one of the parties has appealed to the second-instance court. Where the second-instance court agrees with the first-instance court, it may reject the appeal and affirm the original decision. However, where it disagrees, the second-instance court has a few options: it may remand the case to the first-instance court for a new trial or decide the case itself. ⁹⁴ It is the normal appeal mechanism, part of every judicial system, to guarantee uniformity and quality of justice. The Jiangxi courts moved this mechanism to a new platform. However, it seems the “trial supervision and management” mechanism also applies to these cases, as remanded and revised cases can also fall under the “Four Types,” so there is some overlap. ⁹⁵

Previously, the Jiangxi Provincial High Court had already moved the judicial process to its central digital platform. As a result, court members can now scan and upload legal documents onto the platform. In addition, judges and court leaders can draft and issue documents as well as monitor and manage the trial process via the platform. ⁹⁶ It means a digital trial management platform already existed, and the court built the new e-Management Platform on this previous work. The report claims the new platform adheres to the requirements of the judicial responsibility reforms.

The central dilemma that the Jiangxi Platform tries to tackle is standardizing and monitoring supervision and communication between different hierarchical levels within the court and between higher- and lower-ranked courts. The authors state that the platform addresses several issues. First, they claim the consistency of rulings was not up to standard, implying that individual judges were not applying the law uniformly, leading to similar cases getting different outcomes. They argue that the lack of consistency in judgments put the court’s credibility in danger. Second, they were having difficulties in handling the large number of cases coming in, leading to significant delays in the closing of cases, increasing conflicts between different courts, and increasing mistakes in trial procedures.

Other issues arose after the judicial responsibility reform: there were uncertainties about properly fulfilling “trial supervision and management” duties of sensitive cases, given that it lacked clear procedures. Additionally, the court was unsure about how to protect the autonomy of adjudicating judges while at the same time strengthening the supervision of specific cases. Most significantly, there were issues with detecting such sensitive cases, not only because of the vagueness in the 2015 Opinion, but also because there were no procedures for detecting and reporting these cases. Ideally, the Case Docketing Department (li’an bumen, 立案部门, the department in a Chinese court responsible for registering cases and explaining procedures to litigants) identifies sensitive cases and immediately reports them to court leaders according to clear procedures. However, the report admits that the responsible staff did this largely ad hoc, implying it relied too much on personal discretion to report a case. As a result, some sensitive cases were not reported to the court leadership and were detected too late. In sum, the court lacked proper screening procedures for incoming cases and warning mechanisms when a sensitive case was detected. In addition, there was no unified approach in how to deal with these cases and no proper communication channels for co-ordination between court leaders and front-line judges. ⁹⁷

4.2.1 The “Trial e-Management Platform”

The Jiangxi High Court addressed these problems by designing a new trial management system called the “Trial e-Management Platform.” Previous digitization work allowed the court to integrate cases with the new platform easily and thus completely digitize their supervision and management. According to the report, the Jiangxi Platform automatically screens the digital files of a docketed case to capture and categorize all the relevant data and compare these to data of other cases in the database. Based on AI, the platform “intelligently” identifies the type of case and produces a first analysis, as well as a risk assessment. In addition, it uses AI to realize “intelligent supervision” of the entire trial management process. It also enables intra-court communication, data integrations between different departments and services, and personnel management. ⁹⁸ With this platform, the Jiangxi High Court designed a central venue through which court clerks, front-line judges, and court leaders handle a case and interact during the judicial process.

The Jiangxi Platform integrates five functions related to “trial supervision and management.” The first is the supervision and management of remanded and revised cases. The Jiangxi Platform enables “real-time monitoring” of these cases—that is, a supervisor may consult the case’s progress at any time via the platform to ensure the “quality of case-handling.” It also standardizes, facilitates, and monitors the supervision process itself, meaning that a court leader cannot unduly interfere, and must perform their supervision duties through the platform so that it can be recorded. Additionally, they have limited options in what kind of supervisory measure they can take. It automatically sorts and indexes the differences between first- and second-instance rulings for remanded cases. It also searches and summarizes typical cases to provide a reference for the first judge when they follow up on the remanded case. ⁹⁹

In addition, the Jiangxi Platform allows the second-instance court to communicate with the first-instance court, functioning as an online, multiparty discussion forum. It automatically notifies the first-instance judge when one of their rulings is reversed or remanded and prompts them to reply to the second-instance judge as soon as possible. The first-instance judge may agree or object to the second-instance ruling. The platform automatically notifies the second-instance judge, who can then justify their decision in another reply. Where there remains significant disagreement, even after online evaluation, the case will be reviewed by the high court’s case review committee. The Jiangxi Platform can automatically assign eligible personnel from a centralized database to form a case evaluation committee and create an online discussion group. It permanently retains the opinions expressed by the evaluation committee. ¹⁰⁰

Second, the Jiangxi Platform can supervise and manage long-term unresolved cases. It automatically indexes and categorizes cases that have exceeded a certain time limit ¹⁰¹ into a database, automatically triggering “trial supervision and management” procedures. It also uses “AI, data mining and other technologies” to conduct an in-depth analysis of the case’s progress and identify causes for the delay. ¹⁰² The report mentions that the database includes all overdue cases of the entire Jiangxi Province, implying that intermediate and provincial-level courts are the primary users of this function. Cases that remain unresolved for longer than a year are, without a doubt, complex and sensitive cases. These kinds of cases always get sent to a higher-ranked court. The report claims that the Jiangxi Platform makes inter-court communication and co-ordination more efficient.

The third function is the supervision and management of “Four Types of Cases.” As stated earlier, the Jiangxi Platform screens and indexes the docketed cases. Again, based on AI, the platform detects and flags cases that fall under the “Four Types” and pushes them immediately to a court leader. It allows courts to detect and review sensitive cases more quickly and efficiently. Court leaders must choose a measure via the platform to commence “trial supervision and management” procedures. While the report does not mention what kind of measures the Jiangxi Platform allows, it does mention different implementation measures that stipulate these procedures. Therefore, it is safe to assume that the measures are similar to what has been consolidated in the 2021 Guiding Opinion, such as reviewing case material, asking the adjudicating judge or collegiate panel to reconsider their decision, or submitting the case to a professional judges’ meeting or the adjudication committee. ¹⁰³ It is especially important to note that the Jiangxi Platform records any measure a court leader undertakes. This function seems to respond to the general reluctance to record interventions directly. ¹⁰⁴

The fourth function is called “case information data quality monitoring,” such as general data analytics, cleaning, and indexing. The Jiangxi Platform detects and automatically “repairs” routine data errors and omissions in cases. Where it needs manual input of case information, it automatically contacts the judge who adjudicated the case and sends them reminders every day.

The fifth function is “judicial risk dynamic prevention and control.” It has 41 “integrity risk nodes” in the judicial process and will send warnings in case of procedural errors at each key node. Lastly, a higher-ranked court may use the generated data from all these activities to evaluate judges’ and court leaders’ performance. ¹⁰⁵

4.2.2 Discussion

In sum, the Jiangxi Platform functions as a smart tool for judges. However, it also acts as a supervisor or line manager in the sense that it oversees and regulates the behaviour of judges. On a more abstract level, when judges can only work exclusively through a digital platform, it essentially becomes a closed-choice architecture. This architecture limits the choices of users when dealing with a case. Additionally, it records their actions. Because the users each have their accounts, these actions are also automatically tied to the correct staff member. As a result, a judge’s actions can easily be monitored, measured, and compared to their peers. Therefore, the Jiangxi Platform is arguably a self-monitoring tool and a supervising tool. ¹⁰⁶

This report does not allow us to make any claims about the effect of this automated system on how judges and court leaders operate and perform their “trial supervision and management” duties, or how this is influencing the way courts operate. Yet, this report presents a system that makes it seemingly incredibly harder for court leaders to abuse these mechanisms for improper interference. The system is presented as an “iron cage” that draws judges and their activities into a digitally closed environment. This allows automatic recording. It also allows courts to post-facto review a case’s handling and evaluate the “trial supervision and management” measures for their appropriateness. Court leaders’ performance evaluation includes how well they “supervised and managed” their courts’ work. Therefore, it is crucial to record this so that the higher-ranked court can review this. As noted earlier, empirical research has shown that recording intervention, even legitimate, has put court leaders on edge and has been avoided as much as possible. ¹⁰⁷ Both the normative documents analyzed in the previous section, the reviewed scholarship, and the research report have identified digital trial management platforms as a key pathway to strengthen “trial supervision and management” procedures. It indicates that the refusal to record intervention is a persistent issue, and that technology is envisioned as a way to finally overcome this issue. The Jiangxi High Court presents their system as such: drawing the entire process in a digitally regulated and closed environment is the only solution to enforce procedures.

The report does mention two issues regarding its application. First, it states that some local courts have not been diligent enough in using the platform and organizing their practice around it, resulting in a low awareness and use of the trial management platform. Second, first-instance judges are reluctant to object to revisions of the second-instance judges. Conversely, the report also states that second-instance judges were unhappy with first-instance judges “talking back.” ¹⁰⁸ On the one hand, it raises some scepticism about the success of these platforms in improving supervision. On the other hand, it shows how resistance and reluctance are two real obstacles to digitization and automation. The fact that the Jiangxi High Court decided to mention these issues, and not others, might indicate that resistance and reluctance to change are pertinent issues.

The supervision of lower courts’ work through revising and remanding decisions is frequently a source of tension in the relations between higher- and lower-ranked courts. This tension might explain the reluctance and dissatisfaction with the system: “Grassroot courts hate having their decisions reversed or remanded for retrial.” ¹⁰⁹ Reversal and retrial rates reflect badly on grass-roots courts’ legitimacy and competence as institutions because these rates negatively affect their performance evaluation and ranking relative to other courts. Therefore, it is understandable that a system that allows higher-ranked courts to have easier insight and control over grass-roots courts’ work leads to dissatisfaction among grass-roots judges. Again, the report shows how technology is perceived and operationalized as a way to resolve these issues.

In conclusion, the most important contribution of this digital trial management platform is that it allows ubiquitous supervision of every key node in the judicial process. Where courts have become “completely paperless,” it is almost impossible to avoid using a digital trial management platform. The Jiangxi High Court’s report tries to convey this feeling of unavoidability: the system creates an iron cage that draws its users and procedures into a closed-choice digital environment. It limits discretionary freedom and monitors every action undertaken. Nonetheless, we cannot make any claims about the system’s success, its effect on supervision, and its effect on courts’ operation. In addition, we do not know how this has shaped judges’ behaviour, such as the strategies that they have developed to avoid or resist surveillance. These are questions that remain unanswered as of now.

Similar to Jiangxi’s Trial e-Management Platform, the next digital platform was also developed to improve the oversight of judicial cases. The Yibin Intermediate’s Court model case is also a direct response to rising tensions and issues resulting from the new judicial responsibility reform.

4.3 The Yibin Intermediate Court’s full process automated supervision platform

The Yibin Court’s “Entire Court, Entire Staff, Entire Process” Supervision Platform (the “Yibin Platform”) combines automated and manual functions to enhance “trial supervision and management.” The report mentions that the judicial responsibility reforms have reinforced tensions between the court leadership and front-line judges. It argues that many judges were resisting supervision on the grounds of wanting an independent trial, while court leaders were worried their supervision would be misinterpreted as illegitimate interference. The report faults the “abstract and delegating” manner in which the reform documents formulated new responsibilities and duties, causing uncertainty about the “trial supervision and management” measures, which were unspecified and not standardized. ¹¹⁰

4.3.1 Reforming supervision procedures

In response, the Yibin City Intermediate People’s Court took several measures to “correctly handle the relationship between delegating [adjudication] power and effective supervision.” In 2017, the Yibin Court issued the Measures for the Supervision and Management of the Whole Court and the Whole Trial (for Trial Implementation), implementing the judicial responsibility reform discussed earlier. It stipulated new rules to clarify the duties and responsibilities of “trial supervision and management,” standardize the practice, and strengthen internal supervision, creating a procedural framework that covers the entirety of the judicial process. ¹¹¹

The 2017 Measure immediately underscores the importance of the “full supervision” principle: courts should supervise cases from docketing to archiving. ¹¹² Articles 18 to 24 detail court leaders’ measures when fulfilling their supervision and management duties. It gives court leaders the authority to review and approve procedural matters, but any substantive issues must be discussed and handled via professional judges’ meetings and adjudication committees. It lists the measures they may take via the online trial management platform, such as reviewing a case’s progress. Additionally, it explicitly prohibits court leaders from intervening in cases in which they did not directly participate in their trial hearing, nor are they allowed to give oral instructions, which the document calls a hidden form of the old “case-approval system.” ¹¹³ Finally, Article 23 gives court leaders the authority to check, operate, and monitor the progress of a case (within the scope of their duties), to control and review key nodes in the judicial process, “correct improper behaviour, and coordinate rectification measures.” It creates a “silent process supervision mechanism” to achieve a “full recording” of all case-handling activities. ¹¹⁴ Court leaders who overstep their authority or cause “serious consequences” due to gross negligence shall “bear responsibility in accordance with the law.” The local Discipline Inspection Commission may even get involved. ¹¹⁵

A year later, in 2018, the Yibin Court issued the Measures for the Supervision and Management of “Four Types of Cases” (for Trial Implementation), stipulating the scope and content of the “Four Types” and how to supervise them. ¹¹⁶ Copying Article 24 of the 2015 SPC Opinion on Judicial Responsibility, the “Four Types” are (1) cases involving group disputes that may affect social stability; (2) cases that are difficult and complicated and have a significant impact on society; (3) cases that may conflict with a decision of the court or a higher-ranked court; and (4) cases that involve reports of violations by the adjudicating judge. ¹¹⁷ The 2018 Measure goes into more detail than the 2015 SPC Opinion and clarifies their meaning and scope: the “Four Types” are cases ranging from involving a large number of litigants relating to issues such as labour disputes; to cases involving criminal gangs; to cases in which one of the litigating parties is a government department of the same administrative level as the legal court; to cases in which there are problems in the application of the law, and so forth. ¹¹⁸

The 2018 Measure also clarifies the start-up procedure, giving both the case-docketing department and adjudicating judge the responsibility to flag a case as a “Four Types” and submit it to a court leader to initiate the “supervision and management” procedures. Court leaders also have the responsibility to initiate procedures when they identify a case as a “Four Types” during their routine work. Other departments must do the same in specific circumstances. ¹¹⁹

Articles 13 to 17 stipulate the “trial supervision and management” measures: court leaders may perform so-called “silent supervision”—that is, they may monitor the progress of a case, push relevant decisions and typical cases for reference to the adjudicating judge, review case material and consult case files, attend trial hearings, and even change the adjudicating judge or collegiate panel when necessary. They may also instruct the adjudicating judge to report on the case’s progress. However, where they object to the way the trial process is going or to an adjudicating judge’s ruling, they are not allowed to directly change the adjudicating judge’s or collegiate panel’s rulings but must submit the case to a professional judges’ meeting or adjudication committee. ¹²⁰ The time, content, stage of the process, and the results of the “trial supervision and management” must be permanently recorded on the digital trial management platform. These measures may not interfere with the autonomy of the individual judge or collegiate panel. ¹²¹ Lastly, where the relevant people are found to have neglected the reporting of a “Four Types” case, or where court leaders have neglected to exercise or have improperly exercised their duties, they will be held accountable in accordance with relevant regulations. ¹²²

4.3.2 Digitizing and automating supervision procedures

Based on these measures, the Yibin Court built the “Entire Court, All Staff, Full Process” Supervision and Management Platform (the “Yibin Platform”). The court designed lists with specific items according to criminal, civil, and administrative jurisdiction to classify cases requiring supervision. Based on these items and a decision-tree model, the Yibin Platform “intelligently” identifies, flags, and pushes the cases that require supervision. ¹²³

The report explains the main functions of the Yibin Platform. The first function is intelligent recognition and indexing. It screens docketed cases; extracts, cleans, and indexes the case data; and finally determines the type of case. Where it identifies a case as a “Four Types,” the Yibin Platform flags them to a court leader, providing the necessary case information. It uses a colour scheme to visually indicate supervision progress and give court leaders easy status updates. A court leader must review the flagged case and initiate supervision measures through the platform. A court leader cannot ignore these automatic warnings as the Yibin Platform records non-response. It also records decisions not to launch supervision procedures. The platform provides the “supervision and management” measures a court leader may take, which they initiate by selecting from a list of options. ¹²⁴ Additionally, the Yibin Platform records who adjudicated the case, the reasons for identifying a case as “Four Types,” the reasons of the court leader for accepting or rejecting the case, the measures the court leader took, and the results of the supervision. ¹²⁵

The second function is called “hierarchical early warning supervision.” The Yibin Platform can send warnings to the court one level immediately above the court in which it discovered a “Four Types” case. Each court can predefine the degree of sensitivity of cases according to the actual situation of the court. Certain sensitive cases are immediately reported to a higher level depending on the severity. ¹²⁶ The higher-ranked court decides whether to leave supervision to the reporting court or initiate supervision itself and provide guidance through the Yibin Platform. ¹²⁷

The third function is called “full process supervision and management”—that is, the Yibin Platform records the entire judicial process. At any given stage, the adjudicating judge, a court leader, or someone from the case-docketing department or the trial management department can initiate the supervision and management procedures with one click. In addition, it allows adjudicating judges to request their superior to initiate “trial supervision and management” procedures. However, to avoid abuse and the shirking of adjudicating responsibilities, a court leader can naturally suspend this process where they deem the case not to be a “Four Types.” ¹²⁸

The fourth function is a key node control function. The Yibin Platform records every action an adjudicating judge takes at every procedural step and can report procedural non-compliance, such as exceeding deadlines. When such procedural issues are detected, it can freeze the case, triggering supervision. The adjudicating judge must report to their court leader, who can unfreeze the case after review. ¹²⁹ The platform can also perform automatic searches and provide decision-making references. For cases under supervision, it can perform a preliminary analysis of case material, find similar or related cases, display applicable laws and regulations, and supposedly provide “more accurate and scientific reference material” for judges and court leaders. ¹³⁰

4.3.3 Discussion

In sum, the Yibin Intermediate Court drafted new procedures for supervision, and used technology to enforce compliance. The report claims the system has strengthened the transparency of internal case supervision and ensures proper supervision by court leaders in accordance with procedures and the law. Therefore, an important contribution of the Yibin Platform is that it facilitates and standardizes supervision and “watches the watchdogs.” Recording and monitoring the court leader’s “supervision and management” actions helps to enforce procedures. There can also be no confusion about what measures or how to conduct supervision since the Yibin Platform only has a predetermined, limited list of options that a court leader can take.

Here, again, we observe how technology is envisioned and operationalized to resolve a lasting tension in the work of court leaders. It is also presented as such: the report mentions that with the new judicial responsibility reforms, there was a lot of unclarity and unwillingness to comply. In line with the increase in vertical control as part of judicial reform, the Yibin Intermediate Court increased surveillance to ensure more compliance. The report narrates how previously opaque and discretionary mechanisms to exercise supervision responsibilities are fully codified into a rigid procedural framework. Accordingly, it contains clear conditions for action, clear descriptions of the specific actions to be taken when conditions are met, and consequences in case of non-compliance. In the next step, this clear and rigid structure, akin to an IF-THEN chain set, allows the automation of this mechanism.

Just like the previous case, the Yibin Intermediate Court presents their platform as a system of total surveillance that cannot be escaped. Especially the Yibin case underscores the fact that every single procedural node in the judicial process is monitored: there is no possibility for abuse. The next section compares the two cases in more detail.

5. Drawing justice into an “digital iron cage”

The case-studies pertain an intermediate and provincial high court. This was done because the main purpose of this article was to give a first account of smart court reform, and focusing on different levels of courts gives a more complete picture. Therefore, there exist substantial differences in the purpose and functions of both systems. Nonetheless, it merits a short comparison: there exist some points of overlap and difference between the two that are worth discussing.

The case-studies showcase the different priorities between higher- and lower-ranked courts. The Jiangxi High Court not only focuses on codifying supervision, but is mainly about facilitating and institutionalizing the way lower- and higher-ranked courts communicate with each other. As a provincial high court, their main concern is controlling and ensuring consistency across lower-ranked courts. The system allows them to have a better overview of the cases in their province, and therefore improve uniformity.

In contrast, the Yibin Intermediate Court is more concerned with clarifying supervision responsibilities and codifying these procedures. It creates a rigid procedural framework, which it then enforces through an all-encompassing digital and automated system. In this sense, it is a difference in priorities: lower-level courts are more concerned to get an immediate grasp on sensitive cases and manage them carefully. Ideally, there is no need for extensive communication with higher-ranked courts: if local court leaders perform their supervision responsibilities correctly, an appealed case will not be remanded or revised by a higher-level court.

Next to the difference in rank and priorities, there are also a few points of overlap. A first point is that both systems are supposed to make the detection of sensitive cases, namely cases in which the state’s prerogative is at stake, more efficient. To eliminate the possibility of potential cases slipping through the net, the courts have standardized and automated this detection process. It shows how the judiciary’s primary concern is not necessarily adjudication. Rather, as an administrative institution, it is likely that they perceive technology as a way to reduce bureaucratic errors.

A second point is that both systems have far-reaching surveillance functions. They are presented in the reports as encompassing every node of the judicial process, monitoring and registering every action undertaken. Whether this is actually true remains to be seen, but the idea that is presented shows how the judiciary thinks about judges’ work: it is far more important to be able to control judges’ work and determine who is responsible for what action than to allow judges to do their work well.

A third point that is also connected to the judicial responsibility reform is that smart courts reform enables the judiciary to “collectivize” responsibility. Empirical research found that the reform has given individual judges more autonomy, but it has made adjudication more inconsistent. ¹³¹ The systems help with ensuring procedural compliance and improving consistency, not only by their surveillance capacity, but also by enhancing the co-operation between different hierarchical levels of both courts and judicial staff. When the system flags a case for supervision, it draws a second person into the adjudication process. The Jiangxi Platform also facilitates conveying a judges’ meeting, which has become one of the go-to ways to exercise supervision so as to spread out responsibility for decisions. ¹³²

In sum, technology allows the Chinese party-state to have their cake and eat it too: it has a functioning normative system with procedures embedded into it that allow it to intervene to protect its bottom line, without upending the entire normative system.

In other words, judicial reform has been about unifying the opposing requirements of courts’ dual role in China’s party-state. On the one hand, they need to function as institutions that resolve legal disputes according to law. In this sense, they present the normative state. On the other hand, they also need to act as agents of the party-state, defending its interests where required. In this sense, they present the prerogative state. To unify this, judicial reforms have focused on improving procedures and restructuring accountability and control so as to create a more synchronous operation between the normative and prerogative states.

The case-studies show how smart courts are envisioned and operationalized to facilitate and institutionalize the procedural pathways for the prerogative state to enter into the domain of the normative. The systems enhance procedural mechanisms by enforcing compliance through recording and monitoring. Drawing these processes into a digital environment, the systems create a closed-choice architecture, where every discretionary decision is heavily circumscribed and monitored. The normative process has become fully transparent for the prerogative state.

The cases also show how technology allows the party-state to “proceduralize” prerogative intervention, while guaranteeing the normative state—that is, the judges’ autonomy. The system does not allow a court leader to get involved when it is not required. Conversely, where the system identifies a case requiring supervision, court leaders must get involved within predetermined boundaries, ensuring they do not overstep their responsibilities. Through the all-round surveillance, it prevents abuse of these mechanisms, where individual agents of the state pursue their own interests in the name of protecting the state’s prerogative. Therefore, technology allows the creation of a discretionary space within the normative process, where the state’s prerogative can be protected without this discretionary space being abused. Automation of justice, then, does not necessarily need to be conceptualized as the reduction or removal of human input in adjudication. Rather, automation of justice could refer to reducing human discretionary decision-making during the process. While the human input in the process remains more or less the same, the decisions that they make are based on pre-determined codification.¹³³ In a dual state such as the PRC, this kind of automation is especially convincing.

6. Conclusion

Balancing the contradicting purposes of serving political objectives on the one hand and providing judicial services and legality on the other has been a decades-old challenge for China’s judiciary. The judicial responsibility system attempts to resolve this tension between courts’ political and legal tasks by giving front-line judges more individual autonomy. Simultaneously, it codifies intervention by clarifying procedures regarding “trial supervision and management.” Smart courts are envisioned and operationalized as a pathway to further unify these contradicting purposes.

The case-studies of the Jiangxi High Court and Yibin Intermediate Court have shown how technology is envisioned as further strengthening these reforms by automating and digitizing the “trial supervision and management” mechanism. Many Chinese courts have designed all-encompassing and comprehensive digital environments to monitor every judicial step. This closed-choice environment enforces procedural compliance. It enables court leaders to get involved in cases and align the outcomes with the political and social considerations according to the party-state’s bottom line. It allows sufficient space for discretionary decision-making in politically or socially sensitive cases, but this discretion remains tightly codified and monitored.

Ultimately, smart court reform is only part of a bigger reform in which the party-state is increasing and enhancing channels of vertical control over its judiciary. An implication may be that judges will turn into pure law-applying bureaucrats rather than law-interpreting adjudicators. This issue is related to a central question in the literature on automated justice: How much discretion does a legal system want to grant its adjudicators? Smart courts and broader judicial reform reveal that the Chinese party-state grants its judges only a minimum of autonomy, and the digital environment highly circumscribes even this minimum. Whereas ethical questions related to algorithmic justice have heavily occupied Western debates, the Chinese party-state has made its choice without much concern for such questions. The case of China’s smart courts showcases how governments can leverage technology to encroach on judicial authority and independence.

Another question remaining is whether this resolves China’s courts’ precarious position? Answering this question may serve as a cautionary tale of how external observers assess political and legal reform in China and complete the puzzle of what “reform” actually means in the context of the PRC. Despite all reform rhetoric, efforts have mainly focused on increasing supervision and reducing discretionary decision-making. Therefore, reform is not a fundamental rethinking of existing structures, but rather more of the same—that is, increased vertical control facilitated by ubiquitous technological innovations. Ultimately, the judicial system remains firmly embedded in the country’s administrative hierarchy that heavily favours co-operation and where courts are often the weakest actors.

Reforming these fundamental characteristics as a pathway to increasing courts’ credibility and effectiveness is significantly harder than increasing control mechanisms through automation and digitization. Moreover, broader developments in China’s political and legal landscape, where Xi Jinping has underscored the importance of party control and political loyalty, also indicate that these kinds of structural reforms do not fit the party-state’s vision.

Lastly, smart courts are only one of many digitizing and automating efforts across different governance areas in China, such as the social credit system and smart cities. As with smart courts, local government officials conceptualize the technology of the social credit system somewhat like a “cheat code” that will enable them to solve decade-old issues in Chinese governance and justice without fundamentally rethinking the structure of China’s politico-legal system. In this sense, both embody the prevalent techno-optimism or “technological solutionism” among Chinese reformers. From their viewpoint and the central party-state’s viewpoint, smart court reform is developing as intended, and its problematic consequences are, in fact, logical consequences of automation and digitization.