| In recent years,with the maturity and wide application of big data and machine learning technology,judicial artificial intelligence has gradually become a reality.In this context,artificial intelligence judicial decision-making has gained considerable cognitive and technical possibilities.This paper aims to discuss this possibility and reality,mainly involving three closely related topics: First,the legitimacy of artificial intelligence judicial decision,mainly discussing whether artificial intelligence should be allowed to engage in judicial decision.This involves some legal philosophy issues related to morality and ethics,and a full discussion of these issues is the basis and premise for further research.The second is the methodology of artificial intelligence judicial decision,which involves the technical feasibility of artificial intelligence judicial decision.Firstly,the existing logical framework and technical scheme of artificial intelligence legal reasoning are introduced,and then based on the existing scheme,appropriate promotion is made.This is the core issue of artificial intelligence judicial decision.Third,the system supporting problems related to artificial intelligence judicial decision-making mainly involve how to combine artificial intelligence with justice and the different operational logic behind them.Technical logic and judicial logic have both consistency and conflict.Therefore,while giving full play to the role of judicial artificial intelligence,we should ensure that the judicial decision-making with artificial intelligence can meet the needs of judicial practice and conform to the basic judicial system of our country.This requires that artificial intelligence and judicial practice can shape each other.Therefore,on the one hand,it is necessary to adjust related technologies,and on the other hand,it is necessary to adjust the system.The latter is the third issue that this paper focuses on.In order to fully show the logical relationship among the above-mentioned problems and achieve the purpose of the research,this paper intends to discuss the above-mentioned three closely related problems from three main aspects.This paper is divided into five chapters.Chapter 1 is an introduction,whose main function is to lead out problems and define some important concepts.As we all know,under the theme of artificial intelligence judicial decision-making,there are two basic questions: First,should artificial intelligence judicial decision-making? That is,the question of whether or not;Second,can artificial intelligence judicial decision be carried out? That is,the question of "whether or not".The question of "whether or not" involves the legitimacy of artificial intelligence judicial decision,and the question of "whether or not" involves the technical feasibility of artificial intelligence judicial decision.Therefore,the problem of artificial intelligence judicial decision-making includes both epistemology and methodology.After answering these two basic questions,we also need to consider the question of "how is it possible" for artificial intelligence to make judicial decisions.This paper attempts to answer from two aspects: first,the possibility and technical path of artificial intelligence judicial decision-making;The second is the necessity system construction of artificial intelligence judicial decision.Generally speaking,this paper is based on the idea of "whether","whether" and then "how" in artificial intelligence judicial decision-making.After eliciting the above problems,this paper also preliminarily defines and clarifies several important concepts such as judicial artificial intelligence,intelligent justice and artificial intelligence judicial decision-making,so as to make some preparations for the following discussion.Chapter 2 mainly discusses the legitimacy of artificial intelligence judicial decision.On this issue,a popular view is that if artificial intelligence is allowed to make judicial decisions,it will not only deny the subjective status of human beings,but also lead to a series of problems related to legitimacy.Therefore,artificial intelligence should not be used to supplement or replace human judges for judicial decisions,but must be limited to a certain extent and used for some auxiliary judicial work.This thesis consists of the following four sub-arguments or arguments: First,artificial intelligence judicial decision-making will endanger the subjective status of human judges,and then lead to subjective crisis;Secondly,artificial intelligence judicial decision-making does not have the basis of public authorization,which may endanger the independent operation of judicial power and trigger the crisis of power legitimacy;Third,artificial intelligence judicial decision-making will inevitably lead to judicial errors,which will lead to a vacuum of judicial responsibility,which should not be allowed in any case;Fourth,artificial intelligence judicial decision-making will damage judicial procedural justice and some inherent attributes of justice,and then fundamentally shake the essence of justice.Through a concrete analysis of the above arguments,we can find that the above arguments are not valid or may be exaggerated.Because artificial intelligence judicial decision-making is essentially a problem of engineering realization,it follows the practical philosophy under the framework of bounded rationality.Therefore,artificial intelligence judicial decision-making will only strengthen,not weaken,the subjectivity of human judges,and will not change the public attribute and independent operation of judicial power.In terms of avoiding judicial mistakes,artificial intelligence may do better than human judges,and we should not demand the absolute correctness of artificial intelligence judicial decisions;As for the imputation of judicial errors,it can be effectively solved by distributing the responsibilities reasonably among judges,courts and the state.Finally,transparent artificial intelligence judicial decision-making can guarantee rather than damage procedural justice.Therefore,there is no prior obstacle to the legitimacy of artificial intelligence judicial decision.Chapter 3 mainly discusses the methodology of artificial intelligence judicial decision,that is,the technical feasibility and approach of artificial intelligence judicial decision.For this kind of problem,there have been many researches in academic circles,and three main technical approaches have been formed,including classic rule approach,classic case approach and data approach.However,the above three technical approaches all have problems in different degrees,which can not meet the requirements of judicial decision-making.The classical rule approach adopts symbolism method,which understands the realistic rule system as a complete logic system,but it is difficult to realize rule reasoning based on logic because it can not fully express legal knowledge logically;The classical case approach is also difficult to carry out analogical reasoning effectively because of the problem of knowledge representation.Data access greatly improves the practicability of judicial artificial intelligence and shows its methodological innovation through the statistical analysis of big data.However,the concealment of neural network reasoning makes it difficult for legal reasoning based on data access to make normative interpretation.Although the later scholars reformed the data approach based on the argumentation theory scheme,its case-oriented technical characteristics determined that it lacked institutional authority in decision-making basis.In order to solve the above problems,this paper proposes a rule-oriented mixed technology approach,which integrates rules,cases and neural network reasoning methods,and forms an artificial intelligence legal reasoning mechanism under the framework of connectionism,which is dominated by rules and supplemented by cases.Specifically,in the stage of rule reasoning,we try to express symbolic rules as neural rules,and carry out legal reasoning by matching the elements of rule conditions with the elements of case facts;In the stage of case-based reasoning,it combines the increasingly mature legal text parsing technology with legal argumentation,excavates arguments and argumentation structures in cases through machine learning,and applies judgment results and argumentation to pending cases.Generally speaking,this mixed technical approach can meet the technical and legal requirements of practicality,authority and interpretability,and can better fit the judicial system which is similar to our country.Chapter 4 discusses the system construction of artificial intelligence judicial decision.This part mainly discusses the system supporting problems related to the above two problems,in order to ensure the implementation of artificial intelligence judicial decision-making at the institutional level,and its core lies in determining how to coordinate the relationship between artificial intelligence and justice at the institutional level.In this regard,our basic judgment is that artificial intelligence and judicature can shape each other,not just one-dimensional regulation of artificial intelligence by judicature as some people imagine,nor unilateral change of artificial intelligence to judicature.This means that in the process of integrating artificial intelligence into justice,it is required that artificial intelligence technology has judicial practicability,so as to play the role of artificial intelligence in judicial decision-making,and to ensure the compatibility between judicial system and artificial intelligence.This part will systematically discuss the impact,shaping,adjustment and reconstruction of the judicial system caused by the fact that artificial intelligence participates in judicial decision-making,so as to ensure the practical effectiveness,system legitimacy and social acceptability of artificial intelligence judicial decision-making.Under the current technical conditions,artificial intelligence judicial decision-making is possible,but it needs necessary system adjustment and system construction,including open system,interpretation system,responsibility system,data management system,and systems related to the operation of judicial artificial intelligence.The fifth chapter is the conclusion,which mainly summarizes the legitimacy of artificial intelligence judicial decision-making,technical path and system construction.As far as legitimacy is concerned,judicial artificial intelligence should be understood as a technical project to meet the needs of judicial practice,and its evaluation should be based on its function in judicial practice rather than other prior standards.As long as it can achieve good practical results,it should not be rejected blindly on the grounds of legitimacy.In terms of technical feasibility,we believe that the current data approach has realized the methodological innovation of the original technical approach,which makes the judicial artificial intelligence have obvious technical feasibility,but it also has the defects of lack of interpretability and authority.Therefore,we put forward a hybrid legal reasoning scheme dominated by rules,in order to make up for the defects of current data access by integrating rules,cases and neural networks,so as to better adapt to China's judicial system.In the aspect of system construction,this paper puts forward the corresponding system design aiming at the tool positioning,operation principle,procedure requirements and data management of artificial intelligence judicial decision-making,and puts forward the solution of "mixed responsibility system" for the judicial responsibility of artificial intelligence judicial decision-making.In a word,the judicial decision-making of artificial intelligence has considerable legitimacy and technical feasibility.Although there are still many problems in it,these problems are not a priori,which can be completely solved by necessary technological innovation and institutional adjustment.In a word,the wide application of artificial intelligence in the judicial field has bright prospects and will soon become a reality in the near future.The wide application of artificial intelligence in the judicial field is not a denial or elimination of human judges,on the contrary,it is a liberation of human judges: it frees many judges from the dilemma of "there are too many cases but too few",so that they can have more time to solve difficult cases.In this sense,we should not be entangled and depressed for artificial intelligence judicial decision-making,but should cheer for it!... |
PDF Full Text Request