Uncertainty and Risks in Algorithm Governance and Their Countermeasures

算法治理中的不确定风险及其应对

Abstract

Algorithmic governance is a significant trend in the transformation of public governance or government management in the digital age. As an emerging technology, algorithmic applications face many unknowns, uncertainties, and unintended consequences, thus generating a special type of risk: "uncertainty risk." Uncertainty risk in algorithmic applications can be categorized into various types based on its generation domain, degree, and nature. Analysis from the perspectives of tools, subjects, and contexts in algorithmic governance shows that the characteristics of algorithmic technology, the ambiguity of decision-making subjects, and the complexity of public governance are sources or causes of uncertainty risk. Given the particularity of uncertainty risk, a shift from "deterministic thinking" to "uncertainty thinking" is necessary. Based on this, it can be addressed through participatory algorithmic design, the principle of algorithmic publicness, and adaptive governance mechanisms.