[author]中国法与社会研究院

[content]

Editorial

Weidong JI

China Institute forSocio-Legal Studies

The Internet of things (IOT), big data and artificialintelligence (AI) are the basic components

of the new industry and complement each other. The IOT isconstantly generating data; data are economically valuable, even considered the“oil” or means of production of the twenty-first century; theprocessing, analysis, and utilization of data require AI. Here, there is asignificant positive correlation between data and AI: the larger the scale andthe higher the quality of the data, the stronger the effectiveness of AI. Thehuge population in China and the diversified applications in e-commerce, onlinefinance, and mobile phones make China the world’s largest data-outputcountry with the largest scale of the bigdata industry. This lays thefoundation for the development of AI. The 5G mobile communication system, as ahub and device, further strengthens the interactive relationship between theIOT, big data, and AI, forming a new type of social communication and operationsystem with low power consumption. In addition, the culture of Japan and otherEast Asian countries is mostly pantheistic. They also have an optimistic andfriendly attitude towards robots. Osamu Tetsuka’s anime “Astro Boy” stands in sharp contrast to the terrifying and tragic atmospherein the science-fiction movie series Terminator directed by JamesCameron. Therefore, countries in Asia (at least East Asia) have the advantageof promoting and popularizing big data and AI.

From a legal perspective, big data and AI constitute arule-embedded system. Therefore, Professor Lawrence Lessig points out that “code is law” several times in hisclassic book on Internet law, Code. As an extension of this proposition, algorithms arethought to be laws that can govern social actions. China’s Alibaba Group launched the Sesame Credit Rating System,which shows examples of AI scoring people to determine their eligibility forloans, car rental, house purchasing, and even their employment and promotion.The Joint Punishment Mechanism of Dishonesty connects sesame credit with courtsand lawenforcement authorities, which affects the judgment and implementationof law. In Singapore and South Korea, the online-trial mechanism has developedto a relatively high level. On 2 April 2019, the “Guangzhou 5G WisdomCourt Construction Strategic Cooperation Agreement” was signed, markingthe official start of the construction of China’s first 5G smartcourt. Predictive police activities, which started in Chicago, US, have beencomprehensively applied in China, and their efficiency has been greatlyimproved due to the grid and three-dimensional management. However, the disputeover the “face-swap” agreement of image software company ZAO’s in June 2019 and the first case of face recognition inOctober 2019 revealed the risks of data capture and abuse from the perspectiveof human-rights protection, highlighting the tension between the modernrule-of-law system and AI technology.

The above facts and social trends are the background forplanning a Special Issue on big data, AI, and laws in Asia in the Asian Journal of Law and Society, and also show the academic value and practical significanceof the topic. Since the 30th Anniversary Conference of the O˜nati Institute of International Law and Sociology and theAnnual Conference of the Research Committee on Sociology of Law (RCSL) werejointly held in June 2019, I hoped to organize a symposium on “Big Data, AI and Judicial Service Across Generation” and solicit contributions, which has received positiveresponses from Professor Zuo Weimin, Professor Cheng Jinhua, Professor Yang Li,and Associate Professor Yang Fan. In order to expand the scope of participantsand make international comparisons, I sent an e-mail to Håkan Hydén, ProfessorEmeritus in Sociology of Law, Lund University, on 19 November 2018, invitinghim as co-chairman to jointly call for papers and received his full support. Inrecent years, Professor Hydén has shown a strong interest in AI algorithms as asocial norm, and has a strong willingness to promote collaborative research inthis area between Europe and Asia. A month later, he told me that several youngresearchers decided to attend the Symposium with him and deliver a speech, andsent me the abstracts of the speeches by Associate Professor Stefan Larsson,Associate Professor Pedro Fortes,Lecturer Ulrika Wennersten, PostdoctoralResearcher Ekaterina de Vries, etc. For visas and other reasons, most of theChinese-speakers were unable to attend the O˜nati Conference, butsome submitted papers later. All the European participants were present, whichensured the success of the two thematic sessions. To this end, I would like topay special tribute to Professor Hayden and his team.

At the O˜nati Conference held in June 2019, I met with ProfessorShozo Ota from the Law School of the University of Tokyo in Japan, whom I hadnot seen for a long time. We discussed the possibility of conducting China–Japan collaborative research in the field of laws, bigdata, and AI, and I introduced the idea of a special issue in the Asian Journal of Law and Society. He warmly introduced the Japanese researchers ProfessorKatsumi Nitta and Professor Ken Satoh and promised to invite them to submitcontributions. Later, he proposed that we jointly organize a group session onAI and justice, entitled “AI-Assisted Court System: How AI Can Help Judges, Lawyersand Litigants” at the Asian Law and Society Association (ALSA) OsakaSeminar in December 2019. Professor Cheng Jinhua, Yang Li, and I went to Osakato attend the group session and delivered speeches. Speakers from Japan, in additionto Professor Ota from the law major and Professor Yoshinobu Kano from theinformation-science major, introduced the research findings of usingjudicialexamination data and legal provisions for machine learning to explorethe deep structure of legal reasoning.

Based on the speeches at several group sessions of theabove international symposiums, we selected eight papers for the Special Issue.Professor Håkan Hydén’s masterpiece “AI, Norms, Big Data,and the Law” makes clear the theme from the very beginning byexamining the significance and scope of the sociology of law research on bigdata and AI from a macro perspective, and puts forward such new basic conceptsas technical norms that are in contrast to social norms, including algo-normsand the second order of normativity involving law and the order of varioussocial subsystems, and a series of issues that have led to changes in socialgovernance. Algorithms can affect people’s daily life as norms,but people cannot influence algorithms through democratic procedures. In thissense, digital information technology, represented by AI, is a revolutionarytechnology that is causing profound changes in the relationship between thestate and the individual. Professor Hydén believes that, in order to understandthe corresponding social changes, it is necessary to promote the sociology oflaw research on algorithms, so as to expand the scope of theoretical frontiers andempirical analysis to explore the impact of digital technology on systems andorder from the perspective of social science. The goal pursued by the modernlegal order is predictability and certainty, but the networking and in-depthlearning of AI make unexplainability and uncertainty major features ofalgorithm norms. This means that a paradigm shift must be carried out in socialgovernance and institutional design, and more attention should be paid to therole of trial and error in the legal order. The establishment of the “Special Economic Zone” is an importantinvention of China’s reform and opening-up. Japan has learned from thisexperience and applied it to the development of AI. Professor Hydén sees such aspecial zone (Tokku) as a “living lab” for decision-making.From the perspective of the trial-and error process, the developmentorientation of AI and algorithms should not foster a regulated economy, but be market-friendly.

Associate Professor Stefan Larsson’s paper “On the Governance of Artificial Intelligence throughEthics Guidelines,” based on the Guidelines for Trustworthy Artificial Intelligence presented by the European Commission’s High-Level Expert Group in April 2019 and the European Commission White Paper onDigital Strategy and Artificial Intelligence released in February 2020, analyzes the basic concepts,main content, and impact on the legal system of European AI governance,especially the combination of hard law and soft law. This paper focuses on themanifestation of human-centred AI governance in the social structure and interactionprocess, points out the main challenges of technological innovation to legal andsocial change, and emphasizes the relationship between big data and AI, and theneed for interdisciplinary research on the transformation of social-governanceparadigms. This article also cites the ethics, policies, and legal norms of AIgovernance in China and Japan as examples to compare the basic framework andmechanism design between Europe and Asia. On 25 May 2019, the Beijing Consensus on ArtificialIntelligence was jointly releasedby Beijing Academy of Artificial Intelligence (BAAI), Beijing University, TsinghuaUniversity, Institute of Automation of Chinese Academy of Sciences, Institute ofComputing Technology of Chinese Academy of Sciences, and ArtificialIntelligence Industry Technology Innovation Strategic Alliance (AITISA). Fromthe three aspects of research and development (R&D), use, and governance,the following 15 guidelines were proposed in the Beijing Consensus on Artificial Intelligence: benefiting people, serving people, being responsible,controlling risks, being ethical, being diverse and inclusive, being open andsharing, using AI wisely and properly, informed consent, education andtraining, optimizing employment, harmony and co-operation, adaptation andmoderation, refinery and implementation, and long-term planning. The Rule ofLaw Forum of the World Artificial Intelligence Conference held on 30 August2019 also released the Blue Book onWorld Artificial Intelligence Rule of Law and Guidelines forthe Security and Legal System ofArtificial Intelligence. Japan’s AI R&D guidelines put forward five major concepts: human-centred,international sharing, benefit-and-risk balance, technology neutrality, and emphasison soft law. By comparing Europe with Asia, it can be found that theinternational community has reached some basic consensus on the principles andpolicies of AI R&D and governance.

Associate Professor Pedro Fortes’s paper, “Paths to Digital Justice: Judicial Robots, AlgorithmicDecision-Making, and Due Process,” analyzes the impact ofinformation technology, big data, and algorithm-based decision processes onjustice, including development of legal AI such as online dispute-resolutionsystems, criminal recidivism risk assessment and early-warning technology, androbot judges. The author believes that, although the digitization of justice isnecessary and practical, it is not necessary to radically advocate the automationof legal judgments, but to analyze and monitor the algorithm in accordance withthe principle of procedural justice. To this end, he, based on the USCorrectional Offender Management Profiling for Alternative Sanctions (COMPAS),reveals the systematic deviations of big data and the resulting algorithmicdiscrimination or otherwise in the risk assessment. He points out that the keylies in using the principle of procedural justice to avoid the algorithms’ black box and the neo-collectivist labelling ofconvictions and sentencing, so as to ensure controllable and interpretable AIand tailored justice.

The above threepapers provide a general analysis framework and an international comparative perspective.Next, we will look at the progress of research on big data, AI, and law inAsia. As China attaches great importance to the application of AI in judicialand law enforcement, with unique advantages in data collection and analysis, ithas been very active in the research of this area in recent years. The papersof several scholars analyze and discuss China’s experience and itstheoretical significance from different aspects. Japan leads the way in theresearch and production of robots, and has long-term deep research onlegalreasoning expert systems and legal information technology. Singapore hashugely promoted the digitization of justice, and South Korea also has some goodpractice. Unfortunately, we have not found a suitable contributor for the timebeing. I hope we can make up for it in the future.

The paper co-authored by Japanese scholars Katsumi Nittaand Ken Satoh comprehensively introduces the Japanese experience of applying AIto the legal field. They first introduce the legal-expert-system researchincluded in the national project of the fifth-generation computers launched inJapan in 1982, and the research project of the legal expert system initiated byProfessor Hajime Yoshino in 1985, focusing on the analysis of his team’s contribution to the development of algorithms for legalreasoning and an intelligent consulting system for patent law. Since 2007,JURISIN, an international workshop on legal informatics, has replaced the aboveprojects as the main platform for AI and legal research in Japan, and, since2014, it has jointly organized COLIEE, an innovation contest on legal informatics.This series of organized research activities developed several auxiliary AIsystems to support legislation, justice, and legal services.

The paper byChinese scholar Professor Weimin Zuo and his collaborator Chanyuan Wangexamines China’s judicial big data and legal research based on big data.The authors grasp the significance of big-data legal research from theperspective of legal empirical research. They believe that judicial big datasuch as online judgments formed in the context of transparent decision-makingwill become a new resource for empirical research and will cause arevolutionary change in the legal-research paradigm. However, they emphasizethat large amounts of data and structured data after official processing arenot equivalent to big data. It is necessary to pay attention to the science ofthe mining and analysis of big data, strengthen the correction of incompletelarge amounts of data, and emphasize the complementary relationship between “small data” obtained throughsampling surveys and big data, and the significance of statistical analysis,machine learning, and other methods.

My paper, “The Change of JudicialPower in China in the Era of Artificial Intelligence,”focuses on calmthinking in the trend of AI. The lawsuit explosion and the unification of the legal system are important reasons for China’s judicial authorities to actively adopt new information technologies such as the Internet, big data,cloud computing, and AI. From Shanghai to Guizhou, courts across the nation are tryingto ease the backlog of cases by sorting out simple cases from complicated ones, verifyingthe maximum annual case-load of judges, strengthening assessment accountability, andadjusting the ratio of judges to trial assistants, and reducethe burden of mechanical labour and improve the speed and quality of processing materials and data by using computerinformation-retrieval systems and other auxiliary means. The “smart courts” are indeed conduciveto improving judicial efficiency and the justice of “treating like cases alike.” However, if AI isallowed to go beyond the scope of auxiliary means to try cases and even replacethe judgment of judges to a large extent, it is likelyto lead judicial power astray.

Practices such as allowing AI to automatically generatejudgments and correct deviations of legal decisions based on big data are boundto inevitably form a dual structure of trial subjects, and even lead to themultiplexing of decision-makers. A situation in which programmers, software engineers,data processors, and information-technology companies will jointly make a decisionwith a judge may even appear. Once the trial subject and the decision-maker aredifficult to specify, the power boundary becomes blurred, and the judicialaccountability system is likely to become a matter of form; at least, thepossibility of passing the buck has been greatly increased. More importantly,big data and AI will become the “guillotine” of court debates, creating an atmosphere that “everything depends on the established algorithm, while face-to-faceconversational argument is not important,” making China’s weak legal reasoning, legal argumentation, and legalinterpretations even more insignificant. This leads to a fundamental change inthe structure and function of the modern judicial process.

Yaohui Jin and Hao He have proposed in their paper on “An Artificial Intelligent-Based Semantic Assist Frameworkto Judicial Trial” an AI-based trial semantic assistance framework based onthe practice of speech recognition, text processing, and image classification insome local courts in China, allowing information extraction and machinelearning to achieve coherent and consistent logic in standardized texts, casescenarios, sanction conditions, and the reasons for the judgment. The computingand characterization capabilities of AI are continuously strengthened, but theinterpretability of models declines. This paper aims at the above practicalissue, especially the legal issues that cause the failure of the accountabilitymechanism, trying to break through the bottleneck of the algorithm manipulationand provide the necessary logical interpretation for the data processing andoutput so as to achieve the goal of interpretable AI.

The starting point of the paper on “Big-Data Measurement-Model Research about Judges’ Actual Workload in China” by Li Yang, Junlin Yi,and Hui Peng is the practical needs of increasing the numberof litigation cases and the judicial staffing-system reform in China in recent years. In order to improve judicial efficiency anddetermine a reasonable number of judicial posts, it isnecessary to measure and evaluate the trial workload and performance. Since cases vary in difficulty and the judicialestablishment and funding vary greatly from place to place, it isvery complicated to determine the number of judicial posts and assessment standards based on workload. This paper proposes a modelfor calculating the weight of cases and the workload of judges, especially theaverage annual workload based on judicial big data, through theinvestigation and study of the actual practice of the local courts and the analysis of the measurement standards of the SupremeCourt, and tries to serve as the basis for thestaffing-system reform.

The focus of George G. Zheng’s paper on “China’s Grand Design of People’s Smart Courts” is that, while other countries are actively usinginformation communication technology (ICT) for informal dispute resolution andfocusing on the development of online dispute resolution, China is focusing onthe Internet and AI to improve the formal dispute resolution, namely thejudicial field. The author believes that the algorithm obviously helps tostrengthen the rigidity and uniformity of the application of the law, and alsostrengthens the binding force of past cases under the guidance of the principleof the same judgment of similar cases. Especially in criminal trials, throughthe application of ICT, the processes of collection, proof, and argumentationof evidence have become more standardized, improving the precision andintegration of justice. It is necessary to point out that, although there are technologicalinnovations such as the Libra Chain agreement in judicial aspects, the overall resultof legal-technology innovation is to further strengthen the inherent structuralattribute of judicial hierarchy in China. Digital approaches such as big dataand AI seem to make the pyramidal control of trial activity more efficient,through case assignment, performance appraisal, and judicial accountability.

The worldwide epidemic of the COVID-19 virus sinceJanuary 2020 has severely impacted the global economic system and theinternational order. A series of emergency measures to prevent infectiousdiseases have suddenly made isolation and segregation a feature of daily lifein today’s society. Against this background, existing bureaucraticorganizations appear to be in a dilemma, and emerging ICT further performs thefunction of pooling and distributing information, resources, and materials. InChina, it is mobile payment, online shopping, takeaway, self-media, MOOC, videoconferencing and online offices, etc. that have reconnected self-isolatedpeople with quarantined people, thus forming some kind of flexible organizationand virtual community, and constructing the platform of community-basedgovernance in an interconnected manner. The investigation of travelers fromepidemic areas and suspected patients, the monitoring of quarantined persons,the analysis of treatment cases, and the prediction of the development of theepidemic all require the use of big data, AI, blockchain protocols, distantthermometers, drones, and other technological means, so as to make theoperation of the government more and more intelligent.

In a certain sense, the outbreak of the COVID-19 virushas promoted a great transformation, and the innovation of national governancefor e-government and network government is speeding up. This institutionalchange further proves the practical significance of our Special Issue. We hopeto take it as an opportunity to further promote the application of newtechnologies such as big data and AI in social governance, and explore in depththe possibilities of the state- and legal-paradigm innovation.