Research On The Legal Regulation Of Algorithm Automatic Decision-making

With the rapid development of big data and artificial intelligence,algorithmic automatic decision making has been widely used in various fields of social life as an efficient way to solve problems.The application of algorithmic automatic decision making is increasingly rich,and algorithmic automatic decision making is increasingly reshaping digital social life,bringing digital convenience but also inducing a series of legal risks.Therefore,how to regulate algorithm automatic decision-making has become a hot topic in academia.This paper is divided into six parts.The first part is the introduction.Firstly,to summarize the research background of this paper,we are stepping into an "algorithmic society" controlled by automatic decision-making of algorithms.Then,the literature on algorithm automatic decision is reviewed and analyzed,and the two parts that are most closely related to this paper are analyzed specifically: the study of legal risk and the study of the overall regulation path.The second part is an overview of algorithm automatic decision-making.Before studying the legal regulation of algorithm automatic decision-making,we must first have a clear concept of algorithm.As one of the cornerstones of automatic decision-making,algorithm has a long history.In the context of algorithmic automatic decision-making,algorithm refers to the accurate and complete description of problem solving scheme,which is a series of clear instructions to solve the problem.The limited autonomy,high efficiency,concealment and difficulty of accountability of automatic decision-making of algorithm help data controllers and system operators to achieve their own goals,but also bury legal risks.The third part combs out the legal risks caused by automatic decision-making algorithm.In general,algorithmic automation can improve the efficiency and fairness of decision-making,but it will also lead to legal risks such as weakening human autonomy,increasing privacy disclosure,algorithmic discrimination,information cocoon and damaging the rights and interests of vulnerable groups.From the objective point of view,the original "black box" of the algorithm,data quality defects,and alienation risk of system upgrade may lead to the generation of risks.From the subjective point of view,the improper operation and excessive dependence of application subjects on automatic decision-making of algorithms is also one of the causes of the above five special risks.The fourth part analyses the current situation and dilemma of legal regulation of automatic decision-making of our algorithm.The traditional regulation path has implementation difficulties and is not enough to deal with the legal risks caused by algorithm automatic decision-making.The existing legislation is scattered and not comprehensive,the effectiveness of relevant provisions is low,and the content needs to be improved;Lacking specialized regulatory bodies and scientific regulatory concepts,algorithms with complex technologies are often unable to cope.Meanwhile,the pressure of regulation rises abruptly with the expansion of automated decision-making algorithms.The fifth part compares and analyzes the regulation paths of the European Union and the United States.The EU focuses on embedding algorithmic automatic decision-making regulation within the logical framework of data protection,empowering individuals;However,the United States tends to regard automatic decision making by algorithms as the direct regulatory object and pursue fairness and rationality in external accountability.By comparing the experience of foreign legal regulations,it will provide a reference for our legal regulations of algorithm automation,and explore the path of regulation of algorithm automation with Chinese characteristics.The sixth part elaborates the legal regulation path of improving algorithm automatic decision.Starting from the concept,we should insist on the subject position of human,and put forward the ethical requirements for automatic decision-making of algorithm.In the part of legislative suggestions,it expounds the necessity of legislation and puts forward some frame suggestions for legislation.In dealing with regulatory difficulties,it is suggested to set up a special regulatory body and introduce other regulatory forces such as industry associations and social supervision.It is proposed to promote the impact assessment system of automatic algorithm decision making,improve the transparency of automatic algorithm decision making,and solve the lag problem of post-accountability.Finally,a clear accountability mechanism should be established to ensure the healthy development of automatic algorithm decision-making in the economic and social fields.