| In the cross field of law and artificial intelligence(hereinafter refer as “AI”),AI is used to realize legal retrieval functions such as correlate-statute-recommendation and similar-cases-recommendation.However,these applications have not achieved the expected results,and most searchers prefer to conduct legal retrieval by themselves.Despite the facts that searchers are not familiar with the emerging application of AI technology,the inaccuracy and low quality of the statute and cases recommended by AI is also an important reason.This paper intends to investigate the application status of AI in auxiliary legal retrieval,find out and analyze the obvious problems in such application,and in which part the application can be improved,then put forward specific suggestions to improve.The article is divided into five parts: the introduction is mainly to raise questions,point out that Chinese legal AI system is not unsatisfactory in the practice of legal retrieval,introduce the significance of the topic and related research results both foreign and domestic,and sketch the main research methods,structure,innovations and shortcomings of the article.The first chapter talks about the concepts of legal retrieval and AI,by distinguishing the concepts of legal adjudication and legal decision-making,it is proposed that AI can independently draw conclusions in legal search,but it is essentially still in a supplementary position,analyzes and compares the respective advantages and disadvantages of humans and AI in legal retrieval,and briefly explains the auxiliary status and specific applications of AI in legal retrieval.The second chapter introduces the history of the application of AI in China,especially the background and current situation of the development of AI assisting legal retrieval in China,and point out some problems in the process of correlate statute recommendation and similar cases recommendation.For instance,there are inherent problems in the process of legal retrieval such as the contradiction between legal provisions and reality.There are also practical problems such as the date is too little or distort which result in the low quality of AI’s deep learning and the unsatisfactory quality of legal retrieval results.Moreover,there are technical problems such as AI’s inability to accurately identify fuzzy concepts and understand the changes of social environment,which lead to the low quality of correlate statute recommendation and similar cases recommendation.The third chapter mainly introduces and analyzes the application of legal AI in foreign countries and the operation status of AI legal retrieval system and platform,analyzes and summarizes the difficulties faced by foreign legal AI companies and legal retrieval platform,and points out the lessons to be learned.Combines foreign experience to conduct necessary analysis on the aforementioned issues and make suggestions for improvement,such as adding the function of "suggesting words",putting forward suggestions for keyword replacement,establishing a legal concept sorting system,improving the two-way connection of legal articles and the function of screening for specific regions.And improving the display function of legal retrieval results,such as distinguishing paragraphs of judicial documents and cases,showing the key information at the beginning of the cases,making clear whether the case and the correlate rules are effective,adding the keyword priority and word frequency screening function.And puts forward some suggestions on the structure of the legal retrieval database and AI deep learning for legal retrieval,such as unifying the law and case coding,providing the undisclosed cases to AI for deep learning,and establishing a feedback mechanism to train AI in legal retrieval.The concluding part summarizes the article and looks forward to the future development direction of AI. |
PDF Full Text Request