摘要
随着近年来算法问题的大量出现,人们开始思考如何规制算法。算法透明原则是学理和实践中众所周知的一项算法规制原则,许多学者对算法透明原则十分推崇。但与事后规制相比,算法透明原则作为一种事前规制方式,其规制效力有着天然的缺陷。即使算法透明原则可被用来限制“算法黑箱”的不利后果,但在大规模通过立法、行政、司法措施规制算法的时代,算法透明原则通常既不可行,也无必要。因此,就算法透明原则在算法规制谱系中的合理定位而言,其应该处于非普适性、辅助性的位置。比起本质主义色彩浓厚、以算法透明为代表的事前规制,以实用主义为导向、以算法问责为代表的事后规制是更加得当的规制策略。
In recent years,with the emergence of the problems of algorithms,people begin to think about the way to regulate algorithms.Among those thoughts,the Algorithmic Transparency Principle is,practically and theoretically,a well-known principle.It has attracted a significant number of scholars in recent years.However,as ex ante regulation,as opposed to ex post regulation,the Algorithmic Transparency Principle has its inherent limitations.Although the Algorithmic Transparency Principle can alleviate some“black box”problems,in the age of massive legislative,administrative and judicial regulation on algorithms,the Algorithmic Transparency Principle is normally neither feasible,nor necessary.As a result,the reasonable role of the Algorithmic Transparency Principle in algorithm regulation pedigree should be a non-universal and subsidiary one.Based on the critical analysis of the Algorithmic Transparency Principle,this article considers the reconstruction of the theory for algorithm regulation,and further argues that in contrast to essentialism-driven ex ante regulation such as algorithmic transparency,the pragmatism-driven ex post regulation such as algorithmic accountability should be a more appropriate regulatory strategy.
出处
《环球法律评论》
CSSCI
北大核心
2019年第6期20-39,共20页
Global Law Review
