Identification of Algorithmic Evidence in Administrative Punishment Cases

Abstract

In the field of administrative punishment, algorithmic evidence is the immediate result obtained through the established algorithm in the operation steps of automated decision-making by the government. The intelligibility of this kind of evidence will continue to disappear with the development of artificial intelligence technology. Compared with traditional evidence, algorithmic evidence is highly technical and complicated, and it has the endorsement of public authorities. In judicial practice, only such evidence is reviewed legally. Judges often evade reasoning on technical issues as laymen, resulting in administrative disputes that cannot be substantially resolved. In the face of off-site law enforcement, judicial decisions should jump out of the original evidence review framework, ensure that evidence is not misidentified in the evidence collection stage, implement the burden of proof of administrative subjects and technical subjects in the evidence collection stage, and adopt different identification standards according to the nature of administrative acts in the cross-examination stage, to balance the efficiency of judicial review and the effective rights and interests of administrative counterparts.