Algorithm Regulation:From Normative Regulation to Hierarchical Regulation
There is a fundamental problem with regulating algorithms using traditional legal norms,particularly when defining algorithms with legal concepts:there is a lack of direct correspondence and fluent communication between the norms and the algorithms.This is because concepts express human behaviors,while algorithms represent machine decision-making.As concepts are textual expressions based on human language,they can not confine another form of digital expressions based on machine language.Consequently,the challenge encountered by the algorithm regulation will get intensified continuously when regulation develops further,which is how to match and communicate between conceptual expression and algorithms.This requires algorithm regulation demonstrates variable capabilities according to the different forms of algorithms like generality and specificity,and different levels of algorithms,as phenomena or hidden layers.Constructing a hierarchical structure for the regulation of algorithms will help us to maintain a clearer cognitive judgment of algorithms and regulation about them,and define the power and limitations of algorithm regulation.Only by transforming the normative regulation based on legality judgments into a hierarchical regulation aimed to promote the credibility of algorithms can we regulate algorithms effectively.
万方数据
维普
