Exploring Practical Pathways for the Legitimacy of Algorithm-Driven Automated Administrative Activities

Abstract

In the era of algorithms, algorithmic administration presents the dual effects of efficiency improvement and procedural risk, and the tension between its operation logic and traditional administration needs to be systematically studied. Combined with the characteristics of digital technology, the analytical framework of technical due process is constructed, which helps to establish a procedural regulation mode compatible with the characteristics of algorithm. Currently, the core elements of publicity and participation in traditional administrative procedures, such as, are facing problems such as algorithmic black box, data defects, algorithmic discrimination and technological monopoly. These problems hinder public participation, violate the principle of administrative openness, and result in automated decision-making discretionary inaccuracies. In response to these procedural crises caused by algorithmic characteristics, the development of technical due process theory can be adapted from three links: in the input link, the establishment of algorithmic evaluation system to identify whether the algorithm has a negative impact on the public and to measure the degree of its effect; in the operation link, the construction of algorithmic bias correction mechanism and vulnerable groups relief system to reduce discrimination and protect the public interest; in the output link, the strengthening of the algorithmic explanation system to enhance decision-making. In the output segment, the algorithm explanation system is strengthened to enhance the transparency of decision-making, thus realizing the contemporary development of the principle of due process and boosting the rule of law process of China's digital government construction.