算法决策下行政程序的风险以及法律规制

Risks of Algorithmic Decision-Making in Administrative Procedures and Their Legal Regulation

摘要: 数字政府自动化决策在提升政务效率的同时，也因其技术特性引发了程序正当性、权力结构异化及法律规制滞后三重法治风险。这些风险根源于技术工具理性与法律价值理性的内在冲突。本文主张构建法律价值引领下的双轨规制框架：技术层面需强化数据安全与算法透明度、公平性；法律层面则需明确自动化决策的适用边界，并将公开、公平、基本权利保障三大原则嵌入技术应用全流程，通过立法分层赋责、算法再造嵌入正义参数、建立问责机制等路径实现。最终强调“法律为体、技术为用”的治理逻辑，确保数字政府自动化决策在法治轨道上运行。

Abstract: While automated decision-making in digital governments enhances administrative efficiency, its technical characteristics also trigger threefold rule-of-law risks: procedural legitimacy issues, alienation of power structures, and lagging legal regulations. These risks stem from the inherent conflict between technical instrumental rationality and legal value rationality. This paper proposes constructing a dual-track regulatory framework guided by legal values: technically, it is necessary to strengthen data security, algorithmic transparency, and fairness; legally, it is essential to clarify the applicable boundaries of automated decision-making and embed the three principles of openness, fairness, and fundamental rights protection throughout the entire process of technical application. This can be achieved through paths such as hierarchical legislative responsibility, reengineering algorithms to embed justice parameters, and establishing accountability mechanisms. The study ultimately emphasizes the governance logic of “law as the essence, technology as the means” to ensure that automated decision-making in digital governments operates within the orbit of the rule of law.