Interview by Nyu Wang

In a wide-ranging discussion, Nyu Wang, a PhD student from Virginia Tech, spoke to Zhuhao Wang, an associate professor at the China University of Political Science and Law. They covered China’s courtroom automation, AI-enabled assistive judgment systems, smart courts, and online trials in what Professor Wang frames as a new era of “e-justice.” Professor Wang also decodes China’s national policies and action plans on AI development in China’s court system, as well as its major challenges and changes it brings to the law and to legal education.

Nyu Wang (NW): Tell me about some of the biggest challenges confronting the court system in China.

Zhuhao Wang (ZW): The first challenge is that domestically, the number of judges is limited. Because of personnel reform, former judges were culled for incompetence. This problem has been historically passed down. For instance, there are some fairly senior judges who have not received a formal legal education, or judges who graduated from regular schools but are not professionally competent. The number of judges is actually decreasing because of the exam for post-based judges.

When China’s courts try cases, because they belong to the inquisitorial system’s way of thinking, they need to actively investigate, actively discover the facts, judges have [to take] initiative. Compared to US judges, they’re tired, [and] the pressures and responsibilities they undertake are bigger.

The second challenge is that the general public lacks full trust in the judicial system and the legal system. The third challenge is that China’s judicial system has fallen behind for a long time, [and] has long been in a state of learning to study either common law or civil law systems. There isn’t a very clear self-identity. There also aren’t any outstanding results to take and show to the world, that can be said to be world-leading, or to be commensurate with the currently second-biggest economy, this kind of developed judiciary.

NW: In a paper I published through IEEE, I wrote about language recognition systems, sentencing judgment models, and the “206 System,” among other [technologies used to] help judges in trials. Aside from these new technologies, do institutions have ways of resolving the issue of fewer judges and increasing caseloads?

ZW: China has many legal personnel, plus many Chinese people have studied law overseas. However, whether Chinese legal talent is cultivated abroad or at home, in the end the number of people truly willing to be judges is low because everyone knows the intensity of the work is especially high.

You can search for news on Chinese judges working to death and facing sudden deaths, of which there are many. Rates of death by overwork and sudden deaths are perhaps the highest in the country’s civil service. Furthermore, salaries are low and the demands on judges are high, so in the short term, the number of judges in China will not reach a satisfactory quantity.

Right now technology indeed plays a supplementary role. There are a few technologies that actually are pretty good at freeing up a lot of manpower, but on the whole they’re all fairly elementary. Technology itself is not a panacea. I summarized technological development and applications in law as ‘e-justice’ in a paper I published.

NW: Which specific technologies are being used in the courts to address the problems you identified?

ZW: A few legal AI technologies, such as automatic speech recognition (ASR) technology for trial transcripts that supplement court stenographers’ [speech] recognition. This improves the efficacy of clerks’ work, and takes off a lot of pressure. The main company working on this task is iFlytek. This set of systems is extremely common, it seems that over two thousand courts nationwide are using [them], ASR systems cover over 80%.

Aside from this, there’s a system for automatic notification for similar cases, mainly enabling judges’ decisions to be more consistent, allowing “like cases to be judged alike.” Similar cases that are judged similarly don’t have big gaps between their judgements. This set of systems is from a company I know to be called “Fa Xin.”

When paper documents were still used, clerks had to expend great efforts to deliver documents, maybe half their time was spent outside [of the courts] running around delivering documents. Now there’s e-service [of] judicial documents, which can save plenty of time for the courts’ general staff.

The second challenge I raised is lack of public faith in the judiciary.

In resolving the question of public trust, from 2013 the Supreme Court started to establish a legal document network, open legal documents, for free. Before, legal documents were not open [to the public], ordinary people couldn’t see them. After the legal document network [launched], there was a qualitative change. Judgments and very important documents could all be looked up on the legal document network, it saw a click rate in the tens of billions. After the legal document network was established, the Supreme Court also set up four gateway websites — including open court proceedings, open results of legal judgments, as well as document service and others. The Supreme Court mainly opened these four websites [on] the open judiciary, judicial transparency — these aspects — vastly improving efficiency, and [they] have been a great help in raising public trust in the judiciary. Numerous trials can be directly viewed on courts’ websites [via livestream], so people can watch for free.

Smart courts and Internet courts are world-leading. Internet courts are also a part of smart courts, the Supreme Court invested a lot into this. The Supreme Court was vigorous in going and promoting these projects that Chief Justice Zhou Qiang was inextricably linked with. There’s one [idea] called “overtaking the curve [in the road],” speaking of China’s smart courts as “overtaking the curve” globally [i.e., taking advantage of a narrow opportunity to make progress], catching up to and even surpassing so-called Western countries. But I think that this isn’t overtaking the curve. [Instead] it’s driving straight on a new road — it built its own, completely new road.

NW: Have you noticed changes or reactions to these technologies from others who work in the courts?

ZW: The most direct feeling is from the beginning of these smart court and e-justice projects, the concept of online trials, arose from scratch, and what’s more, they’ve become the new normal. This is revolutionary. This set of systems began two years before Covid, they never anticipated that Covid would emerge. Covid was merely a catalyst, causing a broad swath of courts and the general public to accept online trials.

Over two thousand courts around the country are doing online trials, once they have litigants’ consent. I feel that these past two years have seen a huge change in China. Obvious evidence of this type of change was the October 19 in 2021, [when] the Supreme Court’s Chief Justice Zhou Qiang gave a report to the Standing Committee of the National People’s Congress that proposed amending the People’s Republic of China’s Civil Procedure Law. Amending the Civil Procedure Law is the highest level of Chinese legal reform. He suggested five areas in need of amending.

Of the five areas, he named “the need to perfect online litigation and rules for service [of court documents]; to clarify that online litigation and offline litigation have equal legal effectiveness.” These words carry a great deal of weight. The clarification that online trials and offline trials have equal legal effectiveness [would] need to be written into Chinese law.

Moreover, [the suggestions] to perfect the rules for e-service [of court documents] and announcements of service, [and] to shorten the announcement of service time by a reasonable amount, demonstrate the success of e-service reforms — [they are] successful enough to [merit being] added to the Civil Procedure Law.

From what I’ve seen, there hasn’t been a single country in the world that has written this into their laws. The United States and Europe are still discussing, will they use online trials post-Covid? Will it be ‘back to normal’? While they’re still discussing these questions, China is already writing online trials into its laws.

NW: Who do you think are the most disadvantaged groups that interact with the court system? How do you think these technologies will affect their experiences of the courts?

ZW: I think that the most disadvantaged groups are poor people, people without resources. E-justice noticeably helps the poor. The cost of using a notary public in China differs greatly from that of using the judicial blockchain and directly storing web pages in the system. It may even come down to a difference of 1 RMB and 1,000 RMB.¹

Online trials also save litigants money and time spent running around, such as money spent on staying in hotels. For instance, people hailing from Xinjiang, Tibet, Inner Mongolia, Qinghai — these regions are vast, and have relatively few courthouses. Perhaps if a person has to go to court, a litigant has to go to court, they have to ride a train, as well as switch modes of transportation, [or] drive for many hours until they reach a courthouse. Now their use of online trials is far more convenient.

NW: At the 2020 World Artificial Intelligence Conference I saw a video in which Chief Justice Zhou Qiang particularly emphasized that AI judgement can only assist judges, and ultimately cannot replace judges. So how do you think AI judgment might be developed in the end?

ZW: Chief Justice Zhou Qiang publicly presented AI’s orientation as mainly auxiliary, not as a replacement. But actually, the Supreme Court aggressively researched this particular part, and wants to come close to actualizing AI judging. I read their smart court work plan for 2021–2025, which promotes 50 projects, among which are several projects related to AI judging, to AI cognition. On the one hand, the tone of the policy is to set AI as assisting judges, not replacing judges. On the other hand, however, the invention of technology is very aggressive. The Supreme Court has no core technologies, and they closely cooperate with China’s tech giants, top-tier universities, and related science and technology institutions.

I personally believe that AI judging has different interpretations at different levels. For example, in key cases, cases of high importance, of course it is insufficient to completely replace human judges, such as in a few national major cases. But in the minority [of cases], [like] civil and commercial disputes, AI will definitely have considerable potential.

AI also has error rate problems. Human judges also lack 100% accuracy. Human judges’ error rate is also somewhere in the 10- 20% range. If AI judging can approach human judges’ error rate — for example, let’s say an AI’s accuracy rate is 70%, this is great, and can be used to resolve a few minor disputes like online shopping disputes on [Chinese e-commerce platform] Taobao, because they’re really convenient [and] can be judged with a single click. So from a cost-benefit [analysis], of course there’s a prospect of AI independently judging in certain fields. But at present, [this] has still not [been achieved] to a great extent.

NW: In my IEEE paper, I discussed sentencing recommendation systems that can provide judges with suggested durations of criminal sentence recommendations or suggest verdicts derived from similar previous cases. But not all cases are exactly the same. Will the sentencing recommendations slowly make judges unaccustomed to making judgements on their own?

ZW: There are certainly concerns about this. Though technology is becoming popularized quickly, we should not depend on AI to do everything. At the same time, we should not ignore the intelligence of human beings, and we should not stop developing our own intelligence amidst the development of artificial intelligence. The judges themselves still need to have professionalism to make fair judgements.

The other key issue is the error rate that I mentioned earlier. If the error rate is minimized to a certain level, AI will play a positive auxiliary role for judges and become very helpful. However, if AI’s error rate continues to be high, it will only inconvenience judges’ decision-making and judges may not like using it.

We should also avoid blind worship of algorithms. In your IEEE paper, you mentioned the example of Ke Jie and AlphaGo. This example is fundamentally different from the application of AI in the judiciary. Go is a ‘closed system,’ and algorithms can surpass humans with some training. However, the realm of judiciary is a ‘complex adaptive system.’ Judges need to find the truth and make timely decisions amidst uncertainty and ambiguity. This is a completely open and complex system, so the worship of algorithms must be avoided in this scenario.

AI is a double-edged sword. On the one hand, it does act like a legal assistant that can provide help, but its influence may also be alarming. AI is like an entry-level law clerk and can offer some suggestions. However, using it habitually, which might lead to dependency, should be avoided. In addition, there would be several engineers involved in sentencing guideline design for the AI system, and these engineers have not studied the law. In terms of developing “AI+law,” you can’t just solely leave it to the engineers to design these systems. The ‘smart court’ should reflect the wisdom of its judges.

NW: Do you think that the application of new technology will affect legal aid?

ZW: I think the overall impact is positive. There are similar apps and platforms like Uber in the realm of law. For example, I used to work at a law firm in Texas, and my boss created an app called Lawmato. As soon as this app was on the market, it attracted widespread attention in the legal services sectors in North America and China. Now, many lawyers and potential clients connect through the app, which has a great customer engagement rate. Similar to the concept of Uber, the app connects “supply” with “demand.” In the past, lawyers spent all day trying to seek clients, and clients tried very hard to find good lawyers. Due to information asymmetries, clients did not know which lawyers were good. This platform is very helpful as it has reviews about lawyers, similar to Yelp.

Chinese lawyers in North America often organize lectures on Clubhouse.² Through these public lectures, they present simple legal knowledge or resources to the public. These include relevant topics such as how to find a lawyer, how to file a lawsuit for free. Some answers to legal questions are hard to find and people can find those answers via those online lectures. Those lectures can be considered as legal aid.

NW: What impact will new technology have on Chinese law? Will it produce new departments in future courts? Does it have an impact on law school education?

ZW: Many of these impacts have happened. Zhou Qiang’s proposal to amend the Civil Procedure Law to the Standing Committee of the National People’s Congress on October 19 is a typical example.

Many new departments have started to appear in the courts. The Supreme People’s Court is now recruiting engineers. This was not imaginable before. Engineers are being hired because the Supreme People’s Court wants to master and control core technologies that are being used in court. It is also important for them to strengthen the capability of auditing AI technologies and products. They can no longer rely entirely on technology companies.

The Supreme Court established laboratories in its office. The informatization office has also been expanding for the past two years.

China’s law schools have “law + computing” majors and programs. China University of Political Science and Law (CUPL) is the first school to offer a double bachelor’s degree in law and computer science. Other universities, such as Tsinghua University and Renmin University of China, have “data + law” majors. New professional associations have also been established, with a focus on new “technology + law” and “data + law.” Major law journals in China have also created special columns for publication on this topic, so it’s been a big change.

The Supreme People’s Court is playing the leading role. Zhou Qiang, the chief justice of the Supreme People’s Court, is one of the most knowledgeable experts in this field. The Cyberspace Administration of China (CAC), though not a judicial agency, is very active in this field. For example, the establishment of the Internet Court required the approval of the CAC, and the CAC approved it immediately. It also helps with resource integration and connects the Internet Courts with domestic telecom operators.

A lot of law schools and jurists are researching and exploring new legal fields and interdisciplinary areas that have not existed before. In addition, provincial courts, such as Beijing High People’s Court, Shanghai High People’s Court, Jiangsu High People’s Court and Zhejiang High People’s Court also play very important roles. They integrate resources to do technology (AI) projects, such as the 206 System in Shanghai. There is healthy competition among local courts: courts in Hebei, Guangxi, and Guizhou are all doing “AI + law” projects, and the Supreme People’s Court will give rewards to successful projects accordingly. Therefore, provincial courts have played very active roles in the realm of “AI + law.” The three Internet Courts are testing labs for new technologies. If the projects or technologies yield good results [in the three Internet Courts], then they can be rolled out across the country.

NW: Are there any projects and opportunities for lawyers or law firms to participate in the development of “AI + Law”?

ZW: Opportunities are relatively limited. One of the outstanding projects was developed by Tiantong law firm. Tiantong has a search system similar to wenshu.com, which is China’s official adjudication document network. In some ways, it’s even better than wenshu.com. The key person who drove these efforts was Jiang Yong, who was also one of the leaders of the firm. Unfortunately, Jiang Yong passed away because of overwork. Their efforts in “AI+Law” dropped dramatically after Jiang Yong’s death.

NW: What are the biggest challenges that China is facing in terms of “AI + Law” in the future?

ZW: I think there are three aspects. The first aspect is policy continuity. The Supreme People’s Court is the leader of China’s “AI+ law” initiative, with Chief Justice Zhou Qiang playing a key role in shaping the court’s policy. But according to China’s official system, Zhou can only serve two terms as chief justice. He is still young and needs to rotate to other positions. The chief justice will be replaced before 2025. Though his work plan on “AI + Law” will be developed through 2025, we can’t tell whether the next chief justice will continue the work plan or path of their predecessor.

The second challenge is that the Supreme Court lacks the control of core technologies. The Supreme Court has ideas for AI development, but it relies heavily on tech companies and corporations for technical support to actualize these ideas.

The third challenge is that each step of “AI+ law” requires a large amount of financial support, but the Supreme Court is not a profit-making institution. Its projects are dependent on state funds, so this is definitely a big challenge.


Professor Zhuhao Wang serves as councilor and associate executive director of the International Association of Evidence Science, and is a member of the International Association of Procedural Law. His teaching and research areas include evidence law, theories of juridical proof, legal AI, civil procedure, constitutional criminal procedure, transnational dispute resolution, Online Dispute Resolution (ODR), and collaborations between evidence law and forensic science.

Nyu Wang is a PhD student in the Virginia Tech science and technology program. Her research covers AI and justice, data privacy, as well as AI ethics. Her recent publication on robot judges and China’s court automation is available on IEEE: https://ieeexplore.ieee.org/document/9462216.

This interview was commissioned and edited by Shazeda Ahmed, and co-translated by Nyu Wang and Shazeda Ahmed.


Footnotes

[1] 1 RMB = $0.16 USD, 1000 RMB = $157 USD.

[2] Lectures are in Mandarin and aimed at helping Chinese audiences better understand North American legal systems.