算法商业秘密与算法正义

Algorithmic Trade Secrets and Algorithmic Justice

算法时代中,"算法黑箱""算法歧视"和"算法茧房"等非正义现象引发了算法规制的需要。但算法商业秘密的存在使算法处于"黑箱"之中,成为算法正义实现的障碍。面对算法商业秘密与算法正义的张力可以从三个角度加以考量:商业秘密的公共利益平衡;算法知情同意权、算法解释权和算法拒绝权等私权制约;算法正当程序。由于对商业秘密的公共利益考量主要局限于维系商业道德和市场秩序,无法实现对算法商业秘密的有效制约。算法私权中基于一般道德和法理的算法拒绝权能够成为制衡算法商业秘密的强理由,同时,算法正义需要建构一种基于过程的,而不是仅基于结果的正当程序观念。

In an algorithm age,injustices such as"algorithm black box,""algorithm discrimination"and"algorithm cocoon"have triggered the need for the regulation of algorithm.However,the algorithmic trade secrets put the algorithm in a"black box"and become the obstacles to the achievement of algorithmic justice.To ease the tension between secret algorithms and algorithmic justice,it can be considered from three perspectives:the public interests of trade secrets law;private rights as the check on secret algorithms such as rights to the informed consent,to the interpretation of and to the rejection of algorithms;algorithmic due process.The consideration of public interest is limited to business ethics and market order,and it is hard to effectively restrict algorithmic trade secrets.But algorithmic private rights,based on general ethics and legal principles,can be a strong ground to check and balance secret algorithms.At the same time,to achieve algorithmic justice,it needs to construct a concept of due process that is process-based,not result-based.