Application of artificial intelligence in judicial field
In just 50 years, the rapid development of judicial (legal) artificial intelligence is amazing, especially in recent years, artificial intelligence has a tendency to replace legal persons. In foreign countries, examples of the application of artificial intelligence in the judicial field can be traced back to the 1970s. Developed countries such as the United States have developed legal reasoning systems, legal simulation analysis systems and expert systems based on artificial intelligence technology for judicial practice.
China first applied artificial intelligence to judicature in the 1980s, and Zhu Huarong and Xiao Kaiquan presided over the establishment of a mathematical model of larceny sentencing. In 1993, Professor Zhao Tingguang developed a practical criminal law expert system, which has the functions of searching and consulting criminal law knowledge, reasoning and judging criminal cases and qualitative sentencing.
Today, with the full implementation of key projects such as smart courts and smart procuratorial work in China, the Supreme People’s Court launched the "Smart Court Navigation System" and "Intelligent Case Push System" in 2018, as well as the "Judge Rui" intelligent judgment system in Beijing, the "2006" intelligent auxiliary case handling system in Shanghai, the "Smart Trial 1.0" trial auxiliary system in Hebei and artificial intelligence products launched by other local courts for the trial of judges.
As we have seen, along the way, judicial artificial intelligence is constantly innovating and breaking through the cognitive limits of human beings. Indeed, the application prospect of artificial intelligence in the judicial field is very broad, and many scholars even boldly speculate that with the revolutionary development of artificial intelligence technology, legal persons will gradually be replaced. However, although this new thing has inspired us to think infinitely about the future, we should remain sober, keep some bottom lines while promoting its development, do not have blind expectations and conduct research without direction, and objectively evaluate and use it cautiously within the limits of legal principles and ethics. We will discuss this in detail below.
Application Status of Artificial Intelligence in Judicial Field
(A) legal questions and answers, information processing data
Judicial artificial intelligence is becoming electronic and data-based in legal retrieval and information processing, and this trend will continue in full swing. Such as legal question and answer, pre-litigation consultation, electronic file generation, remote filing, etc. all belong to online information processing technology. During this period, the artificial intelligence used has no independent thinking process, and it is still operated by people in substance. Its core lies in the transformation from traditional offline case handling to online mode, which provides convenience for the parties and case handlers. Among them, the legal question answering robot seems to be closer to the artificial intelligence imagined in our minds. By searching the fixed questioning mode set in advance in the robot system to obtain the required information, it cannot generate additional answers according to different difficult questions, but it can still meet the needs of ordinary cases.
Another example is file OCR recognition, trial speech recognition, evidence recognition and so on, which are perceptual intelligence technologies, which are greatly improved compared with traditional scanning and recording technologies. Taking the trial speech recognition as an example, Linxi voice assistant in Iflytek has optimized the recognition of Chinese accent, and the speech recognition rate has reached over 90%. Compared with the court clerk’s manual input of written materials, the court trial speech recognition technology greatly improves the efficiency of court trial records. After comparative testing, the court trial time is shortened by 20% to 30% on average, the complicated court trial time is shortened by more than 50%, and the integrity of court trial records reaches 100%. In addition, the popularization of this technology can solve the biggest drawback of using audio and video recording technology to record the trial process, that is, the dialect problem in China, which avoids the difficulty of understanding caused by the difficulty of recording recognition in the later stage. Secondly, compared with the carrier of audio and video recording, the identified and transformed electronic book materials are more targeted, more convenient and faster to consult, and are playing the role of liberating clerks.
(2) Automation of document making and case pushing.
In document making and case pushing, judicial artificial intelligence plays a more intelligent role than basic information processing. For most simple cases, such as dangerous driving, small loan disputes, government information disclosure, etc., which can simplify reasoning and use elements and format judgment documents, the automatic judgment document generation system can automatically identify and extract key contents such as party information, litigation requests, case facts and so on through OCR, semantic analysis and other technologies, and generate simplified judgment documents with one click according to the corresponding template. It is even more obvious for the generation of other legal documents, and it can also correct errors automatically, thus greatly shortening the time of drafting documents, reducing the workload of judges and helping judges improve the quality and efficiency of handling cases.
For example, the "Smart Trial Support" system developed by Hebei Higher People’s Court includes such a document production function. It was launched in July 2016 and applied in 194 courts in Hebei. As of June 2017, a total of 110,000 cases were handled in less than one year, and 780,000 documents were generated. In this way, we can build our own case information base, search for similar cases by classifying and matching marks, and automatically screen cases with high similarity in the past when judges handle cases, so as to realize the push reminder of similar cases and provide reference for judges to try similar cases. In this way, the problems of "different judgments on similar cases" and "inconsistent application of laws" can be avoided as far as possible, which is conducive to unifying the local judicial judgment scale and preventing unfair judgment.
(C) case analysis, intelligent referee.
Artificial intelligence reduces the burden of handling trivial daily affairs. In this context, we not only hope that artificial intelligence will play a role in quantifiable and inefficient links, but also expect to further exert its intelligent advantages, thus improving judicial efficiency. Therefore, it has become an important topic to apply it to case analysis and adjudication. In the initial stage of case analysis, by setting the principle of diversion and adjusting the distinguishing elements of complexity and simplicity, the intelligent case-splitting system can handle all kinds of cases in a fine way. In the process of platform operation, the weight coefficients are integrated according to the characteristics of different cases, such as criminal, civil and administrative cases, so as to scientifically calculate the handling power required for each case, help the court to realize the diversion of complexity and simplicity of cases, rationally allocate judicial resources, and ease the pressure of "too many cases". In the in-depth analysis of the case and the auxiliary judgment, the "Judge Rui" system of Beijing court can automatically sort out the facts to be tried before the trial, generate the trial outline and push it to the trial system. The biggest highlight of Shanghai "206 System" is the guidance function of evidence standard and evidence rules, which realizes the intelligent examination of evidence materials and provides standardized guidance for case handlers. In addition, the "AI Judge" launched by Ali has established a complete set of trial knowledge maps for transaction disputes, which can quickly analyze the case and give the judge advice in a very short time.
In a word, the deep combination of justice and artificial intelligence has been discussed in the theoretical circle, and some aspects of the application of artificial intelligence products have been ahead of the theory, and the reform of the judicial system is also the general trend. But this does not mean that the current research and application are completely scientific and effective, and the application space and limit of judicial artificial intelligence need to be carefully scrutinized.
The application limit of artificial intelligence in the judicial field
(a) in breadth can not cover the whole process of judicial activities
At present, judicial artificial intelligence is concentrated in the process of repetitive and inefficient labor and simple reasoning based on the model, which can assist in dealing with transactional work. It is still very successful in areas where general technology is mature, such as online information processing (electronic information) and the application of voice recognition technology, which really provides great convenience for judicial activities. However, we find that both online information processing and speech recognition are widely used technologies in daily life, and the success of these applications depends on the universal application of mature artificial intelligence technologies. Taking speech recognition as an example, the speech input method that we usually use in smart phones is the most common embodiment of this technology, but it is more accurate and perfect in the trial. Although the scene is different, its essence is unchanged.
It can be seen that most of the artificial intelligence products that are successfully and widely used at present are transplanted into some judicial activities by slightly improving the general technology, while for other judicial links, such as the fact finding of difficult cases, the proof of evaluation evidence, legal interpretation, etc., it is impossible to meet their needs only by the maturity of general technology. It is necessary to "customize" specialized products through the infusion of informal logic and judicial experience, so that they can meet most judicial problems and thus run through the whole judicial activities. The key is whether this "customization" can be achieved? This not only means highly developed R&D level as the support, but also needs to deeply integrate artificial intelligence, legal knowledge, judicial experience and other factors, so that artificial intelligence has the logic and thinking of legal person, and its difficulty is much higher than the development of daily general technology. Therefore, if artificial intelligence technology is expected to cover the whole process of judicial activities, it is not difficult for us to conclude that this kind of "customization" is difficult or even impossible.
(B) in depth can not replace the judge’s value measurement.
As mentioned above, judicial artificial intelligence has been successfully applied in information retrieval, document making, intelligent identification, evidence guidance and so on. For some simple cases, it can also realize auxiliary reasoning and put forward sentencing suggestions. Its powerful data processing ability has greatly reduced the burden of judges engaged in daily trivial work. Although there are related studies that show that the prediction accuracy obtained by applying artificial intelligence to predict the referee results is much higher than that of human beings, can artificial intelligence replace judges, as some people pursue?
Artificial intelligence is a technology about data, which depends on human beings to design and program it. It can only absorb and process the existing knowledge and information input by human beings in a patterned way. Although its mastery of existing knowledge far exceeds that of human beings, its fatal defect lies in its inability to adapt to the ever-changing changes in human society. As an immediate reflection of the real society, the application of law is normal to integrate the general concept of social justice into the value measurement of judicial decisions, which is a very complicated process, and it is difficult for machines to learn "value judgment with human touch". Judicial judgment is about the art of experience. American Justice Holmes said that the life of law lies in experience rather than logic. He asked judges to give each case a unique judgment value according to the constant changes of social life and not to be bound by the inherent legal logic. This requires an interactive cognitive process between legal norms and life, and draws conclusions through complex value judgments. From this point of view, artificial intelligence judges can only judge cases according to the pre-set formal instructions, and follow the human control step by step. Without human participation and input, machine learning will not happen autonomously, let alone independent thinking, and it will not meet the contextual needs of judges in the process of value judgment. As Judge Holmes said, law is not pure logic and ultimate rationality, and experience, morality and warmth should also be considered in judicial adjudication. After all, people who only know the law may not be qualified as judges, let alone machines that only know the law.
Application prospect and outlet of judicial artificial intelligence
(A) judicial artificial intelligence: good steel used in the blade.
The biggest advantage of artificial intelligence is that it directly serves people. By intervening in all aspects of judicial activities, it plays an important role in information storage, data operation, simple reasoning, etc. With these auxiliary functions, case handlers can extricate themselves from repetitive and transactional work to the maximum extent, and can also find and correct some details in time, so that they can really invest their time and energy in analyzing difficult cases and realize the optimal allocation of judicial resources. Not only that, artificial intelligence intervention in the judicial process can also constrain the judge’s behavior in the process of handling cases, correct the judge’s subjective prejudice, expand the judge’s cognitive ability, and reduce the negative impact of subjective factors such as intuition on the judge’s value judgment.
In judicial practice, artificial intelligence as an auxiliary tool brings convenience, but it is also limited to serving judges and auxiliary judges. Therefore, the research and development of judicial artificial intelligence should focus on its auxiliary functions. Moreover, based on the current application of judicial artificial intelligence in the field of general technology, corresponding achievements have been made, so in the future, it should be more inclined to develop the technology in the field of judicial specialization according to the professional characteristics of judicial activities. For example, some foreign courts use artificial intelligence to predict the outcome of cases, and also apply artificial intelligence to examine the admissibility of expert testimony, providing guidance and reference for judges to judge cases. China can also develop corresponding products in combination with its own reality when researching and developing. In addition, we can further improve and popularize the functions of sentencing recommendation and early warning of trial deviation, which have been applied in some courts in China.
Technology is an important force for the progress of human society. As the main body of social development, human beings must use technology for our own use. As the saying goes, "a gentleman’s nature is not different, and goodness is better than falsehood." Only by knowing how to develop and use judicial artificial intelligence and being good at promoting its strengths and avoiding its weaknesses can the value of artificial intelligence be maximized. Only by using this piece of good steel in the cutting edge can we create more value for the judiciary and the people.
(B) artificial intelligence judges: out of sight and out of reach.
As early as the 1970s, some scholars have put forward the hypothesis that human judges can be replaced by machine judges to eliminate legal uncertainty. He Rong, vice president of the Supreme People’s Court, China, gave a clear answer: the robot judge is impossible to appear.
In fact, it is difficult for both ordinary people and experts who have been deeply involved in the field to distinguish the concepts of computers, artificial intelligence and robots in detail. However, as we all know, their operation is essentially a calculation process of "receiving information-processing information-outputting information", and each step needs an exact definition, and there is a "unique positive solution". However, the charm of judicial adjudication is that it often does not pursue black or white, and each specific case has its own unique side. Besides pursuing objectivity and legality, the implied worldly wisdom and ethics are complicated, which requires a judge with judicial wisdom and experience to conduct evaluation and discretion. Although the law is cold, the process of applying the law contains warmth. For example, in the case of Yu Huan, the plot of revenge for the mother by killing the "humiliated mother" is full of irrational emotions, which is not unrelated to the etiquette concept of "the enemy of the father, sharing the sky with the husband" since ancient times in China. These human feelings can only be understood and grasped by human judges who have experienced life, received education and felt edification, and are separated by artificial screens. In addition, artificial intelligence also faces the problem of "black box" algorithm. The process of judicial adjudication itself is a process of paying attention to debate and reasoning. In the trial, not only the angry words between the two sides will affect the judge’s factual judgment, but also an interactive look and a subtle expression will become the capture point of the judge. In the Western Zhou Dynasty, there was a record of "listening to prison proceedings with five voices and seeking people’s feelings" in China.However, the decision-making process of the artificial intelligence judge will be the data processing and operation within the system, and the judgment is the result of a "black box operation", and we know nothing about the operation process. The closed decision-making of the "black box" directly violates the principle of judicial openness, and may also trigger algorithmic dictatorship and discrimination, which will inevitably lead to public doubts about the judgment result and lead to social chaos.
More importantly, China’s Constitution clearly stipulates that all the power of the state belongs to the people, the power of the judicial organs is endowed by the people, and the judicial power exercised by judges is essentially public power under the theory of people’s sovereignty. There is such a legal proverb in the ancient west: "A judge has only one boss, and that is the law." In this context of public authorization, the judge enjoys absolute authority in the trial, and no other subject can influence the judge’s final judgment. From this point of view, if artificial intelligence replaces judges, it will transfer the public power entrusted by the people to a machine, which is tantamount to giving up the public utilities of the people and leading to the rule of robots. We don’t know whether the robot can recognize the law as his boss, but it will inevitably dispel the significance of judicial judgment and undermine judicial public trust, which is unacceptable and tolerated by society.
To sum up, the vision of artificial intelligence judges replacing human judges is beautiful at first glance, but it has fallen into the omnipotence of artificial intelligence. Judicial judgment is a skill of understanding and an art of value measurement, which contains experience and wisdom that artificial intelligence can’t understand. It is a distant fantasy that artificial intelligence replaces judges. In the future, the development of judicial artificial intelligence can’t break through the bottom line. Uncontrolled use of artificial intelligence to transform the trial system will only bring irreparable trauma to the modern rule of law.
conclusion
Although the upsurge of artificial intelligence has swept through again and again, constantly breaking through imagination and impacting cognition, we should not blindly pursue it, especially in the judicial field, we should be cautious and rational. There is no doubt that some routine and repetitive work in the judicial process should be handled by artificial intelligence, giving play to its super integrated operation function and improving judicial efficiency; The core power in judicial activities, such as judicial power, must be independently exercised by judges, and artificial intelligence can only be in a subordinate position because of its own cognitive and thinking defects. In addition, if the judicial decision-making power is handed over to artificial intelligence, the algorithm black box and algorithm discrimination will inevitably impact judicial justice, and it will also dispel people’s sovereignty and modern rule of law system.
In the future, we should stick to the dominant position of judges and the auxiliary role of artificial intelligence, and the research and development of judicial artificial intelligence should continue to advance on the road of serving judges and assisting judges, instead of going astray to replace judges. The deep combination of artificial intelligence and judicature is the general trend. It is necessary to do a good job in the addition of judicial artificial intelligence according to the situation, so that "artificial" and "intelligence" can be in their proper places, take what they need, and unite strongly, so as to bring the greatest value to justice and society, and promote judicial efficiency and judicial justice together.
Source: People’s Rule of Law magazine
Author: Jia Zheyu Lv Zhaoshi Sun Xiaopu
Original title: "Application of Artificial Intelligence in Judicial Field"
Read the original text