| 摘要: |
| 我国现行算法歧视的法律治理存在两个弊端，分别是算法歧视规制范围过宽和算法歧视规制工具的凌乱。弊端的消除应当追溯至歧视概念本身，通过区分歧视的法律维度和道德维度，呈现算法歧视的两副面孔。第一副面孔是“算法偏见”，在语义上采用法律维度的歧视，对其治理旨在解决算法当中的历史偏见问题，适用反歧视法的规制原理，对算法设计者和算法应用者施加基于社会法的结果义务；第二副面孔是“算法不平等”，在语义上采用道德维度的歧视，对其治理旨在解决信息时代的资源分配问题。对于“算法不平等”，既要尊重私法自治又要削弱潜在歧视风险，法律规制除了事后司法救济外还可以采取事中方案，通过价值维度和信任维度，对算法设计者和算法应用者施以适度的负担。 |
| 关键词: 算法歧视 算法偏见 历史偏见 算法不平等 资源分配 |
| DOI： |
| 分类号: |
| 基金项目: |
|
| Two Faces of Algorithmic Discrimination and its Legal Regulation |
|
Li Zhiying
|
|
|
| Abstract: |
| There are two flaws in the current legal regulation of algorithmic discrimination, one is that the regulated scope is too board, and the other is that the regulated tools are fragment. The solution comes down to the concept of discrimination in itself. Distinguishing between legal respect and moral discrimination, algorithmic discrimination has two faces. The first face is “algorithmic prejudice” which reflects the discrimination in legal respect and aims to address historical prejudice in algorithmic application, legal regulation of algorithmic prejudice connects with anti discrimination law, which impose consequentialist liability on the designer and user of algorithm. The second face is “algorithmic inequality” which reflects discrimination in moral respect and aims to address the distribution of resources in the information age. In order to maintain the balance between respecting the autonomy of the will and reducing the discriminatory risk in the issue of “algorithmic inequality”, besides ex post regulation, the law can enact an interim regulation in value and trust respect to impose appropriate liability on the designer and user of algorithm. |
| Key words: Algorithmic Discrimination, Algorithmic Prejudice, Historical Prejudice, Algorithmic Inequality, Resources Distribution |
