Research On The Application Of Artificial Intelligence In Judicial Decision

Under the background of rapid update and iteration of AI technology,AI judicial decision-making tools are gradually applied.This paper first introduces the characteristics of AI technology,and reviews the application in the field of judicial decision-making in China and the United States.It is pointed out that in the era of weak AI,technology is only a tool to assist investigation and decision-making,and it cannot replace the judge to make decision independently.At present,the tools are mainly applied through the following ways.The similar cases push function helps the judge obtain the judgment documents of similar cases.The sentencing prediction function helps the judge obtain the sentencing results within a reasonable range.The deviation warning function corrects the judge's excessively deviated judgment results.The electronic files generation function,legal document writing function,personnel management function and other functions can facilitate documents,personnel management,etc.The specific functions mentioned above can not only improve the judicial efficiency,save the tedious steps of administrative work,leave enough time for the trial of cases,solve the dilemma of more cases and fewer people,but also provide scientific suggestions for the judgment results,maintain the consistency of judicial decision-making,and achieve the goal of similar trial of similar cases.In addition,the judgment process of the AI platform is full of marks,which is also conducive to correct the judge's mistakes and supervise the judge's unfair behavior.However,technology is a double-edged sword,and the application of AI technology in judicial decision-making also faces many challenges.First,data is the basis of platform.But data untrue,incomplete and low degree of structure make the foundation of technology application weak.Secondly,because the algorithm is unexplainable and the trade secrets is not open,the calculation of data is in the black box,and the authenticity of the calculation results is greatly reduced.Third,the algorithm results are not friendly with specific groups,and algorithm discrimination is everywhere.Fourth,the judge's decision-making will be affected by the anchoring effect,and the judgment results will unconsciously approach the statistical results.Finally,the AI system is suitable for easy cases but not for difficult cases,because it cannot complete the necessary value judgment and discretion of referees.Therefore,this paper provides the following countermeasures to deal with the risks.In view of the data problem,we should integrate the data resources,establish a cross industry big data platform,combine the characteristics of legal language to solve the problem of insufficient data structure.In view of the algorithm problem,we should ensure the neutrality of the algorithm development subject,and comprehensively consider the causes of algorithm discrimination.In order to avoid the occurrence of anchoring effect,we should recognize the tool status of AI technology,always keep vigilance and avoid the judicial view of data doctrine.Finally,we should distinguish the ordinary cases and difficult cases,and analyze the cases in different ways.