Abstract: Algorithmic fairness is a topic of common concern in the field of artificial intelligence and information ethics. It comes from algorithm bias in algorithm decision systems. Although technicians have made much progress in avoiding algorithm bias, there is still no consensus on how to achieve algorithm fairness at the technical level. Based on this, this research first analyzes the methods and defects of realizing algorithmic fairness from the technology path. Then, along with the current criticism and revision of realizing algorithm fairness from the single technology path, it proposes a social research framework based on Brey's "Disclosive Computer Ethics", which provides a more reasonable solution to the problem of algorithmic fairness.

摘要： 算法公平性是当下人工智能领域和信息伦理学界共同关注的一个话题，它来自于算法决策系统中的算法歧视现象。尽管技术人员在避免算法歧视方面取得了许多进展，但对于如何在技术层面实现算法公平性，至今仍没有达成共识。基于此，本研究首先分析了从技术路径实现算法公平性的方法及其缺陷，接着，沿着当前对于单从技术路径实现算法公平性的批判及修正，提出一种基于布雷的“披露性的计算机伦理学”的社会化研究框架，为算法公平性问题提供了一个更为合理的解决方案。