On The Judicial Regulation Of Algorithmic Discrimination

With the development of Internet and information technology,artificial intelligence and its derivative algorithms have gradually become a hot and difficult issue in legal research.Among them,the governance of algorithm discrimination is the top priority of current theoretical research.In the process of dealing with algorithmic discrimination,justice is a commonly chosen remedy.Therefore,the judicial regulation of algorithmic discrimination has also become the key to solve this problem.Through the combing and analysis of relevant judicial cases,we find that solving algorithm discrimination through judicial channels must go through several necessary links,such as the collection of algorithm parameter evidence,the identification of discrimination facts and the identification of judicial responsibility.However,due to the structural,indirect and technical characteristics of algorithmic discrimination,the judicial regulation of algorithmic discrimination has some obstacles,such as difficult parameter collection,empty discrimination identification standards and unbalanced distribution of judicial responsibility.Specifically,at the level of algorithm parameter collection,due to the demands of national public security,social normal operation order and social individual information rights,the algorithm parameters can not be made public in the collection of judicial evidence to a certain extent;For the identification of algorithm discrimination,this paper analyzes the shortcomings of anti classification standard,subjective intention judgment standard and differential influence standard in dealing with the new characteristics of algorithm discrimination;This paper also discusses the simplification of the judicial responsibility system in the final determination of judicial responsibility due to the complexity of the subjects related to the algorithm.To solve these problems,this paper puts forward different degrees of evidence collection requirements for algorithms in different scenarios;Through the introduction of indirect invisible discrimination,anti stereotype theory and necessity exemption conditions,the traditional discrimination identification standardsare repaired;Based on the construction of the judicial responsibility system of algorithmic discrimination,it provides a basic dimensional reference for the judicial responsibility distribution of algorithmic discrimination,and realizes the multi-level and developable goal of responsibility distribution.Finally,we can find a feasible and effective solution for the regulation of algorithmic discrimination from the perspective of justice.