算法伤害和解释权
Algorithmic Harms and the Right to Explanation

摘要 算法对个人和社会的影响日益明显，同时引起了一系列围绕算法决策不公、歧视与不透明的隐忧。旨在使算法透明化的解释权被认为是应对算法伤害的有效措施。然而，解释权在现有的法律资源中难以证成，在实践中也无法有效解决算法对个体造成 的损害。本文认为，无需另创解释权，GDPR的数据擦除权、数据携带权、数据保护影响评估等权利机制与范式要求便可以使算法更合理和更负责。
The growing influence of algorithm agents on individuals and the society is becoming obvious; meanwhile, they have caused a series of public concerns such as the unfairness, discrimination, and opaqueness of algorithmic decisions. The right to explanation purporting to make the algorithm decision-making transparent is thus considered as an effective measure to deal with algorithmic harms. However, the right to explanation hardly can be proved in the existing legal framework. Furthermore, it cannot effectively solve the algorithmic harms to individuals in reality. We argue that it is not necessary to create a new right of explanation since such GDPR rules as the right to erasure, right to data portability, and data protection impact assessment, can make algorithm decision-makings more reasonable and responsible