Liability Rules for Algorithmic Discrimination

Abstract

Algorithmic discrimination represents the translation of human societal discrimination into the realm of algorithms. It manifests as algorithmic thinking, black-box algorithms, and non-neutrality, all stemming from the design of algorithmic technologies. Furthermore, it becomes apparent in the form of biased decision-making during algorithmic operations. This phenomenon not only infringes upon citizens' equal rights but also disrupts market economic order and jeopardizes social fairness and justice. To effectively address the real-world issues arising from algorithmic discrimination, it is imperative to approach this problem from a legal perspective. Algorithmic discrimination differs from traditional forms of infringement, necessitating a thorough analysis of its constituent elements, responsible entities, modes of liability, and the establishment of mechanisms for liability allocation. Through these legal regulations, a legal framework for the algorithmic era can be constructed, safeguarding the safety and development of humanity in the digital society.