Abstract：Public credit evaluation is the core system of credit supervision. With the iteration of algorithm technology, its advantages in public credit evaluation are becoming increasingly obvious. The intelligent fusion of “data + algorithms” is the technical logic of algorithms-embedded public credit evaluation. Promoting the modernization of government governance and improving social credit is respectively the practical logic and value logic of algorithms-embedded public credit evaluation. However, while giving full play to the technical advantages, algorithm may also cause triple risks in public credit evaluation: technical risks（mainly including black box algorithms and data breach）, legal risks（mainly including power abuse and data abuse）, and ethical risks（mainly including algorithmic supremacy and algorithmic bias）. In view of these risks, it is advisable to adopt a “holistic perspective” and a “process-based approach” in regulation. In the ex-ante stage, it is necessary to formulate technical standards for algorithmic credit evaluation, establish an algorithmic information disclosure system, and improve the algorithmic impact assessment system. When applying algorithm technology in public credit evaluation, it is necessary to use the principle of administrative rule of law to restrain administrative power, and use credit and personal information rights to check and balance algorithmic power. In the post-event stage, it is necessary to improve the public credit evaluation objection appeal system, and at the same time strengthen the algorithm audit and judicial review of public credit evaluation. In the ex post stage, it is necessary to improve the public credit evaluation objection appeal system, and at the same time strengthen the algorithm audit and judicial review of the public credit evaluation.
2022, Vol. 39 
粤公网安备44030502007936号