| In the era of data explosion,algorithms have the ability to learn autonomously and judge independently towards massive data.This technology is widely used in business and public management fields,such as user portrait,product pricing,credit evaluation,insurance underwriting,welfare application and human resource management,which also leads to a great deal of algorithm discrimination risks,for the systematic and repeated unfair treatment of algorithm decision to specific groups,on which multiple unfavorable effects will be exert,such as the injured groups and even the whole society.The problem of algorithmic discrimination needs urgent settlement.However,due to the lack of effective response of the requirements of openness and accountability of algorithms from the current legal regulation,which also fails to effectively regulate the discrimination caused by the application of algorithms,therefore,this paper analyzes the relevant concepts,features,specific performance and legal risks of algorithm discrimination,summarizes the experience of legal regulation inside and outside of China.This paper takes the specific field of algorithm discrimination into full consideration to explore diverse regulatory paths for algorithmic discrimination in business field and public administration field.From the innovative application of private law norms and traditional market regulation law to the grafting of market regulation exclusivity system,this paper constructs various legal regulation paths from different perspectives in different fields.Apart from introduction,the main body of this paper is divided into five chapters.The first chapter ascertains the basic definition of algorithmic discrimination through determination and analysis of the concepts of algorithmic discrimination,whose characteristics and causes of concealment,structure and discrimination are also analyzed.And then the first chapter elaborates on the detail of causes and specific manifestations of algorithmic discrimination in business field and public administration field.The second chapter aims to prove the necessity of regulating algorithmic discrimination by law.Algorithmic discrimination jeopardizes the right of equality for citizens,especially the equality of opportunity represented by equal employment and the right of equality represented by equal protection by law.Moreover,due to non-public,opaque and non-acceptable to doubt characteristics of algorithmic discrimination,it is unlikely for the decision-maker to claim the right of due process in defense to the algorithm.The algorithm discrimination on price in the commercial field is a deemed to be deception to the consumer group,which infringes on the consumer’s right to know and choose.Hence we can see that many social risks caused by algorithm discrimination are illegal and need to be effectively regulated by law.Moreover,value norms and behavior standards of the current algorithm industry has yet formed,which leads to insufficient endogenous power of industry regulation amid the unimplemented ethical regulation landscape due to technical reasons,so it is necessary and urgent to regulate algorithm by law.The third chapter sorts out the response of domestic and foreign legal systems to algorithmic discrimination.From the domestic perspective,China’s direct response to the problem of algorithm discrimination is scattered in the e-commerce law without a systematic legislative norms,and the measures are still regulated from the perspective of personal data protection.As for the foreign point of view,Europe and the United States,as the forerunners of legal regulation algorithm,have given different answers in practice: the EU has formed a regulation mode focusing on "data risk prevention",which strictly prevents algorithm from causing serious detriment from the perspective of personal information protection and data management;while the United States has formed a regulation mode focusing on "algorithm review",which relies preventive algorithm review and retro-review of responsibility to strengthen the legal supervision and responsibility implementation of the algorithm.Then based on the current status that the legal system of China fails to handle the algorithm discrimination,the third chapter puts forward the path of algorithm regulation from different perspectives in different fields for selection aiming at the problem of algorithm discrimination with different causes and illegal attributes in different application fields of algorithms.The fourth chapter puts forward the regulation path of algorithmic discrimination in the business field.The most representative algorithmic discrimination in the commercial field,that is,algorithmic price discrimination,is considered as the regulatory object.In order to construct the anti-monopoly law regulation path of algorithmic price discrimination,this paper takes the factors into full consideration such as exploiting consumer surplus and infringing the equity when wealth transfers,establishes the illegal review standard of "qualitative behavior + implementation effect",reasonably expand the identification method of platform operators’ dominant market position,and formulate reasonable algorithmic price discrimination regulation strategy to construct the anti-monopoly law regulation path of algorithmic price discrimination.At the same time,algorithmic price discrimination is also an infringement of consumers’ right to know and fairly trade.Therefore,by confirming the consumption fraud attribute of algorithmic price discrimination,the responsibility of the operator for contracting fault or punitive damages is determined,so as to realize the regulation of algorithmic price discrimination in the consumer protection law.In addition to the regulation path of traditional department law tort relief,it is also necessary to strengthen the protection of data subject’s right to be forgotten,give data subject the right to refuse algorithm decision-making in order to achieve pre-regulation,and achieve normalization regulation with the improvement of algorithm supervision system.The fifth chapter of this paper proposes a legal regulation path of algorithmic discrimination in the field of public management.The algorithmic system in the field of public management has "quasi algorithmic authority",so we need to get rid of the "black box" features of opaque,non-public and low transparency of algorithms to achieve effective review of algorithms,so as to achieve effective regulation of algorithmic discrimination in the field of public management.The author thinks that through the construction of the exclusive right system of algorithm,we can achieve effective algorithm review and supervision,and then open up a regulatory path to exchange market entry control for algorithm transparency.At the same time,we can establish the corresponding algorithm discrimination review mechanism to exclude discriminatory algorithms from entering the public decision-making field. |
