When China first introduced Smart Court reforms in 2010s, the official narrative emphasized efficiency, transparency, and modernization. Court workloads were exploding, the public wanted faster and fairer judgments, and technology promised a way forward. Judicial AI systems could streamline case management, assist judges in drafting opinions, and even suggest appropriate sentences.
But beneath this techno-optimistic façade lies a deeper story: Smart Courts are not just digital upgrades. They are political projects. This research seeks to examine how these reforms, presented as neutral innovations, actually reshape the relationship between courts, judges, and the Party-state. Far from empowering frontline judges, they are transforming courts into algorithmically managed bureaucracies—where uniformity and control increasingly trump deliberation and discretion.
China’s courts face a real dilemma: they must process millions of cases each year, yet judicial discretion—the capacity of judges to interpret law flexibly—has long been seen by the Party as a potential political risk. Under President Xi Jinping, judicial reforms have sought to address both challenges simultaneously: accelerating case processing while tightening constraints on judicial discretion, not only in politically sensitive matters but also in routine adjudication. One provincial High Court described this shift vividly: “We are moving from person-watching-person to algorithm-watching-person.” Algorithms now stand between judges and their decisions, monitoring not only what they rule, but also how they reason.
My fieldwork and document analysis identify three main technologies driving this transformation. The first is mandatory similar-case retrieval, which requires judges to search for algorithmically selected “similar cases” before issuing rulings. Systems flag deviations from these precedents, nudging judges toward doctrinal conformity. What counts as a “similar case” is often defined not by law but by internal court authorities. Second, case deviation alerts trigger warnings when a draft judgment diverges from prior rulings or administrative guidelines. Some platforms even propose “corrections” automatically. Judges must justify deviations to court leadership before proceeding, creating a presumption against independent reasoning. Third, AI-generated judgments are still experimental but growing rapidly. These systems can now draft entire judicial opinions—from facts to legal reasoning to final verdicts—based on pre-set templates. In one Jiangsu court, judges described the platform as offering “multiple-choice justice,” with little room to insert their own views before finalization.
On paper, Smart Courts promise greater access to justice. Case filing is now automated; citizens can submit complaints online; judgments are published in massive digital databases. Yet in practice, access to justice is becoming access to automation. Litigants report that complex or unconventional claims risk being rejected by keyword-based filing systems. For judges, algorithmic deviation alerts discourage nuanced reasoning. Appeals are becoming less meaningful because decisions already carry the weight of algorithmic legitimacy.
From the court user perspectives, three particular consequences stand out: procedural rights such as the right to be heard or to appeal shrink as decisions are pre-shaped by internal systems rather than open deliberation; legal reasoning becomes standardized, as judges start from algorithmic “recommended outcomes” and then search for precedents to justify them rather than reasoning from facts and principles; and substantive fairness suffers as unique or emotionally complex cases are forced into one-size-fits-all templates.
The rise of Smart Courts marks a turning point in China’s legal system. What began as a pragmatic response to overwhelming caseloads has evolved into a project that reshapes the very logic of judicial decision-making. These reforms promise efficiency and consistency, but they also redefine the role of courts and judges in ways that deserve close attention. In a closed political system like China, the stakes are especially high. Judicial independence has long been off the table, but now judicial autonomy is shrinking as algorithms embed uniformity and administrative oversight into routine decision-making. Legal reasoning increasingly risks taking a back seat to compliance with algorithmic and political mandates. Over time, public trust in the courts may erode if justice comes to feel automated, standardized, and disconnected from the complexities of real-world disputes.
Technology now reinforces both legalism and state control, often at the expense of judicial autonomy. This trajectory raises pressing questions about the future of law and justice in China. Will legal reforms continue to privilege speed and uniformity over reasoned judgment and fairness? And how much discretion will judges retain in a system governed by algorithms, administrative priorities and Party imperatives?
Jiajun Luo is a Hauser Postdoctoral Global Fellow at NYU School of Law. His research explores public law, dispute resolution, and authoritarian governance. He can be reached at jiajunlok@gmail.com.
Picture a prosecutor in Shanghai opening their computer to review a theft case. Instead of manually searching through thousands of precedents, they turn to an AI system that instantly analyzes the evidence and recommends similar cases from a database of millions. Within minutes, the system suggests whether to detain the suspect and even predicts the likely sentence. This isn’t science fiction anymore, it is the reality of China’s “206” system, arguably one of the most ambitious experiments in AI-assisted criminal justice worldwide.
The Rise of Smart Justice in China
Since 2016, China rapidly embraced what it calls “Smart Justice” (智慧司法), integrating artificial intelligence throughout its legal system. The Shanghai “206” system represents one frontline pioneer of this transformation. Unlike AI applications in “Western” courts that focus on specific tasks like risk assessment, China’s system attempts something far more comprehensive: it assists judges and prosecutors across the entire criminal process, from pre-trial detention decisions to sentencing recommendations.
What makes this particularly fascinating is not just the technology itself, but the institutional context. China’s judicial system, operating without the constraints of judicial review traditions found in many “Western” courts, has adopted AI technologies with remarkable speed and minimal resistance. The AI-assisted system now processes hundreds and thousands of cases annually, with the “206” system expected to be utilized in all criminal cases in Shanghai.
How Does It Actually Work?
The “206” system’s core feature is its Similar Case Recommendation function, which operates like a sophisticated legal search engine on steroids. When a prosecutor inputs case details, the system uses deep learning algorithms trained on past verdicts to identify patterns and recommend outcomes. It considers over 50 variables, from the suspect’s employment history to whether victims have forgiven the accused, to generate sentence recommendations.
The system does not just help with individual decisions. For example, in plea leniency cases (which now account for 87% of criminal cases in China), prosecutors may display the AI’s predictions to negotiate with defendants. “Look,” they might say, “based on our AI system, you’re likely facing 3-5 years of imprisonment. If you plead guilty now, we can recommend the lower end according to the law.” The system has changed the dynamics of criminal justice negotiations.
The Good, The Bad, and The Algorithmic
Our research, building on exclusive data from Shanghai’s procuratorate, reveals a complex picture. On the positive side, the data suggest genuine improvements: cases processed with AI assistance took 23% less time to complete, and sentencing recommendations made with AI support were accepted by judges 75.8% of the time, whereas they only accepted 65.6% of the sentencing recommendations without AI support. The system appears to reduce arbitrary detention and increase consistency in sentencing.
However, three critical concerns emerged based on our fieldwork. First, the anchoring effect: when prosecutors see the AI’s recommendation first, it becomes incredibly difficult for them to deviate, even when case specifics might warrant a deviation from the recommended sentence. Once the prosecutor sees the recommended number on the screen, she or he ascribes a sometimes unwarranted authority to it.
Second, accountability avoidance: the system’s complexity creates a perfect excuse for passing the buck should errors occur in sentencing. If something goes wrong, was it the algorithm’s fault? The fault of the system developer? The fault of the prosecutor or judge who relied on the system? This diffusion of responsibility poses serious challenges to China’s judicial accountability reforms.
Third, and perhaps most troubling, is the compression of the rights to a proper defense. Defense lawyers have no access to the system, cannot challenge its algorithms, and often do not even know that it is being used. While prosecutors wield sophisticated AI tools funded by public money, defendants and their lawyers are left in the dark. This to some extent amplifies the already existing problems of an unequal playing field in criminal justice.
Lessons for the Global Legal Community
What can the rest of the world learn from China’s bold experiment? First, procedural design matters more than technological sophistication. Our research suggests that many of the risks associated with AI in criminal justice are not inherent to the technology but arise from how it is implemented. Simple procedural safeguards, like requiring judges to form initial opinions before consulting AI, or ensuring that the defense has access to algorithmic tools, could mitigate many concerns.
Second, transparency isn’t optional. The closed nature of China’s system, where algorithms operate as black boxes hidden behind claims of trade secrets of technology providing companies, undermines procedural justice. Any jurisdiction considering adopting AI for prosecution or adjudication work must grapple with balancing technological innovation with fundamental legal principles like the right to a fair defense.
Finally, China’s experience confirms a widely observed paradox: AI systems designed to reduce human bias and increase consistency may actually entrench existing patterns of injustice if they are trained on historical data reflecting those very biases. The algorithm does not innovate; it replicates and amplifies patterns that have previously manifested.
Looking Ahead
As courts worldwide grapple with backlogs and inconsistencies, China’s aggressive adoption of AI offers both inspiration and cautionary tales. The technology clearly has potential to improve efficiency and consistency in criminal justice. But our research suggests that without careful attention to procedural safeguards, transparency, and equal access, AI applications risk creating a two-tiered system of justice where algorithmic efficiency trumps fundamental fairness. The question is not whether AI will transform criminal justice, but whether we can design systems that harness AI’s benefits while preserving the procedural protections that define justice itself.
The full article titled Access to technology, access to justice: China’s artificial intelligence application in criminal proceedings can be accessed here.
Wanqiang (Aiden) Wu is a Yat-sen Postdoctoral Fellow at Sun Yat-sen University Law School. He received his Ph.D. in Criminal Procedure Law (Cum Laude) from Shanghai Jiao Tong University in 2025. His research focuses on criminal procedure, empirical legal studies, and the intersection of technology and criminal justice in China. He has published extensively on China’s procuratorial system and judicial reforms in leading journals including Modern China, Hong Kong Law Journal, and International Journal of Law, Crime and Justice.Contact him via email.
Xifen Lin is Professor of Law and Vice Dean at KoGuan School of Law, Shanghai Jiao Tong University. Professor Lin is a leading scholar in Chinese criminal procedure and judicial reform, with particular expertise in prosecutorial systems and empirical legal studies.Contact him via email.
Imagine a researcher seeks to answer a fundamental question of legal fairness: Do better-resourced parties (the “haves”) achieve more favorable outcomes in Chinese courts simply because of their socio-economic status?
A naïve answer is to compare the win rates of well-resourced and less-resourced parties in litigation. But the researcher might quickly realize that well-resourced parties are likely to be represented by counsel whereas the less-resourced are likely to be self-represented. Perhaps it is the quality of legal representation that influences how judges rule, not the status of the parties. On this account, the “haves” do better than the “have nots” because they have, among other things, superior legal representation.
A solution might be to control (or adjust) for legal representation. It is common for statistical studies to control for an explanatory variable to isolate the effect of the variable of interest on outcomes. Controlling for legal representation here means, essentially, that the “haves” who are represented by counsel are compared to the “have nots” who are represented by counsel. These comparisons seem to make sense: if the researcher finds that represented, well-resourced parties still win more than represented, less-resourced parties, that indicates that status is driving the observed pattern of outcomes.
This approach can result in inaccurate inferences being drawn due to collider bias. In causal modeling, a collider is a variable that is a common effect of two other factors. A classic, non-legal example is the car that fails to start. This can be caused by a dead battery or an empty gas tank.[1] These two causes are independent; one doesn’t cause the other. Suppose you only look at cars that have broken down, i.e. you condition on the collider “car fails to start”. Then a correlation appears between the two causes: if you know a broken-down car has gas, you can be confident the battery is dead. But of course, the car having gas is not itself a cause of the battery being dead.
So how does collider bias apply to our example of the “haves” and “have nots” in Chinese courts? There is suggestive evidence that Chinese litigants are more likely to seek professional advice in harder, more doubtful, cases.[2] So, if a well-resourced party has a lawyer, it could be because they can easily afford one or because the merits of their claim are doubtful. If a less-resourced party has a lawyer, it is probably because the merits of their claim are doubtful. If a researcher controls for legal representation, she is comparing the “haves” who are represented by counsel are compared to the “have nots” who are represented by counsel. But the “haves” who are represented by counsel have, on average, stronger cases on the merits than the “have nots” who have counsel, and they might therefore prevail in more cases. The researcher may conclude from her controlled comparison that the “haves” come out ahead more often in Chinese courts because of who they are. But this inference is not necessarily warranted by the evidence.
This example demonstrates that big data does not absolve legal scholars from thinking through the causal relationships between the variables in their analyses. Indeed, qualitative and sociological methods can produce valuable domain knowledge for distinguishing plausible relationships from spurious ones. A multi-disciplinary paradigm remains critical to studying the Chinese legal system, even in the era of artificial intelligence.
The full article, titled Data Still Needs Theory: Collider Bias in Empirical Legal Research, co-authored by Xiaohan Yin, is available here. Benjamin Minhao Chen is Associate Professor and Director of the Law and Technology Centre at the University of Hong Kong Faculty of Law. You can contact him via email.
[1] Judea Pearl, Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference (San Francisco: Morgan Kaufmann 1988).
[2] Wenzheng Mao and Shitong Qiao, ‘Legal Doctrine and Judicial Review of Eminent Domain in China’ (2021) 46 Law & Social Inquiry 826; Yali Peng and Jinhua Cheng, ‘Ethnic Disparity in Chinese Theft Sentencing’ (2022) 22 China Review 47
This is contribution #2 in our series on SMART COURTS AND SMART GOVERNANCE IN CHINA, outcome of our workshop in July 2025 at Cologne University.
In our daily lives, there’s an app for almost everything—ordering food, tracking fitness, managing work. But what if there was an app for political loyalty? For the Chinese Communist Party (CCP), this isn’t a hypothetical question. It’s a core part of its strategy for governing in the 21st century. Through a massive initiative known as “Intelligent Party Building” (智慧党建), the CCP is rolling out digital platforms to manage, educate, and discipline its 95 million members.
My research explores how this digital push is reshaping the very nature of power within the Party. It is not just about efficiency; it is about deepening control. The Party aims to achieve two seemingly contradictory goals at once: build a more modern, responsive bureaucracy while simultaneously reinforcing absolute top-down authority. This creates a new, powerful dynamic of internal governance in the digital age.
The Two Faces of State Power: An Iron Fist and a Nervous System
Despotic Power: This is the “iron fist”—the state’s ability to make decisions and issue commands without negotiation. It’s top-down, coercive, and absolute. Think of it as power over society.
Infrastructural Power: This is the state’s “nervous system”—its actual capacity to penetrate society and carry out its plans. It relies on logistics, institutions, and technology to get things done. It is power through society.
Historically, regimes were often strong in one of the types of power but weak in the other. For example, an ancient empire might have immense despotic power (the emperor’s word is law) but weak infrastructural power (it is hard to actually collect taxes in a remote province). Modern democracies often have strong infrastructural power (efficient services) but weak despotic power (leaders are constrained by law and public opinion).
The CCP under Xi Jinping, however, is trying to maximize both. “Intelligent Party Building” is a prime example of this ambition, using digital tools to build a highly efficient administrative nervous system that also serves a powerful iron fist.
Inside the “Party-Building Cloud Platforms”
So, what do these apps and platforms actually do? While they vary locally, they generally focus on four areas:
Digital Dossiers and Performance Tracking: The traditional paper dossier (dossier), a lifelong file tracking a person’s political behavior, is going digital. These platforms create a permanent, tamper-proof record of a member’s participation in Party activities. Poor performance can trigger warnings, public criticism, or even negative entries in their file that could impact their career.
Ideological Education on Demand: The platforms deliver a constant stream of ideological content. This includes live-streamed lectures, articles from state media, and materials from official education campaigns. To ensure engagement, members are often tested with gamified quizzes where they can earn points and compete on leaderboards.
Gamifying Loyalty: The platforms borrow heavily from popular apps to keep users engaged. For instance, the famous “Study the Great Nation” (学习强国) app rewards users with points for reading articles about Xi Jinping or watching videos of his speeches. This subtly transforms political indoctrination into a daily habit, much like checking social media.
Digitizing Bureaucracy: Beyond ideology, these platforms are also practical office tools. They handle routine tasks like paying Party fees, managing announcements, and approving leave requests. By integrating with daily work, they ensure the Party’s presence is not just a separate political activity but is embedded into the fabric of everyday professional life.
A Double-Edged Sword of Digital Control
The true innovation here is not the technology itself—many of these features mimic existing corporate collaboration tools like Ding Talk or WeChat Work. The innovation lies in its dual-use application for political control.
The very same feature that enhances infrastructural power (e.g., efficiently collecting data to provide personalized educational content) is used to wield despotic power (e.g., using that same data to monitor and punish members for insufficient engagement). A tool designed for convenience also becomes a tool for surveillance. A gamified quiz that makes learning “fun” is also a mechanism for ideological enforcement.
Ultimately, the CCP’s push for “Intelligent Party Building” reveals a fundamental tension. While it seeks to modernize its internal management to become more efficient and responsive, it remains unwilling to give up the coercive, top-down control that defines its Leninist roots. The result is a system where the iron fist is now wearing a digital glove, able to reach further and grip tighter than ever before.
Ningjie Zhu is a researcher at the Center for Advanced Security, Strategic and Integration Studies (CASSIS) at the University of Bonn. You can reach him at nzhu[at]uni-bonn.de.
China has given this movement a name—or at least rebranded the product of the alignment of technology and the courts: Smart courts, zhihui fayuan 智慧法院. The name suggests an alignment on the ground that is both linguistic and textual in the operational spaces of courts. Over the last decade or so, and through its Supreme People’s Court, Chinese officials have led a national effort to modernize the judicial system through the use of emerging technologies. Like other modernization pathways elsewhere, the goals include enhancing access to justice and ensuring that access provides pathways toward just outcomes. Since December 2024, these efforts also include an artificial intelligence platform to help judges improve work efficiency.
But names sometimes are a distraction. And that appeared to be the case with Chinese smart courts. The name became a vessel into which people could pour their larger fears about the transformations they feared most—that the people would no longer be their own masters but would serve technology even as technology appeared to serve them. It is no surprise, then, as Susan Finder relates in her examination of the Supreme People’s Court 2024 Work Report to the National People’s Congress, that the term “smart courts” appears to have been dropped.
Nevertheless, “smart courts” have become not merely a symbol of digital reform but a mirror reflecting deeper ideological and systemic transformations. What appears at first to be a techno-administrative modernization effort quickly reveals itself to be an exercise in high-stakes governance theory. The central question I pursued: Can a digitally advanced judiciary maintain alignment with a ruling party that is not itself digitally transformed? In other words, can a smart court operate effectively without a smart Party?
From Robot Courts to Zombie States?
The study of “smart” or “intelligent,” or “wise” courts can be approached from a large number of perspectives. I start from the ordering premise that these “smart” courts can be understood as an object, and also as a symbol or signified conception, and lastly as the set of objects and behaviors that produces its own meaning through its own dialectical phenomenology—that being by doing. This amalgamation of objects and symbols is a matter central to the continued evolution, in human society, of the notion and practice of judging, and of institutions of judging to which it is both attached and to which it lends meaning. But an object and symbol of what?
The term “smart court” evokes both utopian promise and dystopian anxiety. While the ambition of the People’s Republic of China has been to develop courts that are faster, more accessible, and more consistent, the term has also sparked deeper fears—especially outside China—of robot judges, automated justice, and dehumanized legality. This isn’t merely science fiction. Rather, as I suggested during my talk, these fears can be metaphorically grouped into a three-course cautionary tale.
First, courts risk being consumed by the very technology meant to assist them. Their core identity shifts from a site of judgment to a platform for automated processing. Second, courts may begin to consume their stakeholders—litigants, judges, and lawyers—by reducing them to data points in algorithmic workflows. Third, courts may consume themselves, becoming mechanisms of predictive governance rather than instruments of legal deliberation.
Such risks are not unique to China. But within China’s governance model, they raise particularly intense contradictions—especially the one between technology-led modernization and Marxist-Leninist political control.
Semantics Matter: What Is “Smart”?
Much of the misunderstanding about smart courts, I argue, stems from the loaded semantics of “smartness” itself. In English, “smart” blends quick wit, technological capacity, and sometimes pain (its etymology rooted in “to sting”). In Chinese, however, the distinction is sharper. Zhìnéng (智能) points to technical capability—what we associate with AI and data-driven systems. Zhìhuì (智慧), by contrast, suggests discernment, judgment, and wisdom.
This duality—between instrumental intelligence and human wisdom—is crucial. Smart courts, if they are to serve justice rather than mere efficiency, must retain a core of hui: the human capacity to judge wisely. In the Chinese political imagination, this is ideally embodied by the judge and the collective judiciary. But what happens when the source of wisdom—traditionally human—is threatened by ever-smarter systems?
Tech as Instrument, or as Actor?
China’s digital judiciary remains in a transitional phase—digitisation more than full digitalisation. The emphasis is still on improving efficiency: filing systems, access to records, online hearings. Yet, the horizon is shifting. Predictive analytics, caselaw modeling, and AI-assisted adjudication point to an emerging reality where tech not only facilitates justice but begins to shape its substance.
This introduces a profound conceptual tension. As technology moves from being “smart” (responsive and efficient) to potentially “wise” (autonomous and analytical), it also shifts from being a tool to being an actor. This challenges long-standing assumptions about who—or what—gets to decide within a legal system.
The Smart Court Needs a Smart Party
This transformation becomes most consequential in China, where courts are not isolated institutions but deeply embedded within a Party-led governance model. The CPC is not just a political overseer but the ideological architect of the judiciary’s function. Here, smart courts demand something deeper: a smart Party.
By “smart,” I mean a Party apparatus that itself incorporates digital technologies not only in surveillance and administration, but in its very processes of leadership, assessment, and ideological guidance. Without such a transformation, an asymmetry emerges: the courts grow in techno-capacity, while the Party lags in digital adaptability. That gap threatens to destabilize the very premise of Party-led governance.
Rethinking Interpenetration: Court and Party
Chinese governance is structured around interpenetration—the mutual embeddedness of Party and State institutions. Historically, this interpenetration has been managed through personal-bureaucratic forms: overlapping roles, dual appointments, and ideological campaigns. In the digital age, however, interpenetration is reconfigured through data flows, predictive modeling, and feedback loops.
The smart court, then, is not just a site of dispute resolution but a generator of political data—inputs and outputs that reflect the health of Party ideology and administrative discipline. To oversee such a system, the Party must itself become a digitally competent, analytically capable, and ideologically precise actor.
This is no small task. It means building a digitally-enhanced Party apparatus that can assess court behavior, monitor ideological conformity, and even model the likely impact of judicial decisions—all without becoming a mere appendage of the technologies it deploys.
The smart court exemplifies both the achievements and the contradictions of China’s New Era. On the one hand, it reflects the success of socialist modernization: the integration of productivity-enhancing technologies into governance. On the other hand, it surfaces a contradiction between human-led ideological guidance and machine-augmented decision-making. Two key contradictions define the current moment. First, the contradiction between the leadership of the Party and its capacity to lead in a tech-driven environment. Second, the contradiction between technology as instrument and technology as autonomous force. Both must be addressed if the CPC is to retain its position as the core of the political-economic order.
Ultimately, one must come to understand, or at least consider the plausibility, of a principle that under New Era Chinese Marxist-Leninism, the state apparatus can only be as “smart,” intelligent” and “wise” as it is in the capacity and operations of the Party to do likewise. In the presence of asymmetry two fundamental contradictions must be addressed. The first is the contradiction between the leadership of the Party and its capacity to lead. The second is between the techno-instruments through which Party capacity is undertaken and the ability of the Party apparatus to steer, guide, assess, control and utilize these instruments in the performance of its own duties and responsibilities. The fundamental issue of instrumentalization and capacity remains undisturbed—the more autonomous the technology, the greater the risk that the relationship between instrument and its wielders will be reversed, at least in part. In the absence of a capacity to understand and manage those contradictions, either organs better capacitated to wield techno-instrumentalized applications and processes will drive human collective systems, or human collective systems may become an instrument through which techno-wisdom intelligence may realize its own vision for techno-human perfectibility.
Implications Beyond China
While my analysis focuses on the Chinese context, the underlying challenges are global. Whether in Europe, the U.S., or elsewhere, legal systems face similar dilemmas: How to preserve human judgment in algorithmic environments? How to ensure accountability when decisions are guided by machine learning? How to maintain institutional integrity when data becomes both input and output? China’s smart court project offers a provocative case study. It forces us to confront not only what technology can do for justice, but what it might do to justice—and who, in the end, will be wise enough to decide.
The full contribution is available here. Larry Catá Backer is the W. Richard and Mary Eshelman Faculty Scholar and Professor of Law and International Affairs at Pennsylvania State University. His work focuses on Chinese governance, transnational law, and political theory.
