算法决策规制——以算法“解释权”为中心

算法决策正成为经济与社会体系的一部分,一方面,它创造了显著的社会与经济价值,但另一方面,不公不义的预测或推断会损及个人自主与尊严从而使算法备受质疑。由于算法无法解决自身导致的妨害问题,且算法控制者与数据主体间存在明显的信息或权力不对称,因此,有必要赋予个人一项具体的数据权利——算法“解释权”,以强化其对于己不利的算法决策提出异议的权利,进而促进算法正义、保护个人自主与尊严。不过,利用“解释权”对抗算法妨害虽然必要但并不充分,其在技术上面临可解释性难题,且与商业秘密存在紧张关系。因此,算法决策需要统合规制,需要进一步增强算法决策的社会控制,优化算法应用监管。

Regulating Algorithmic Decision:Focusing on the Right to Explanation of Algorithm

As a part of economic and social system,algorithmic decisions are creating remarkable social and economic value;on the other hand,they are often questioned by the public,because some unjust predictions or deductions have been made by them,even individual autonomy and dignity have been impaired by their decisions.Seeing that algorithm can’t resolve nuisance created by itself,and that there is information or power asymmetry between algorithm controller and data subject,it is necessary that right to explanation of algorithm should be granted to data subject,so as to strengthen its right to object adverse algorithmic decision,then to promote algorithmic justice and to protect individual autonomy and dignity.However,right to explanation may be necessary but insufficient,because it’s hard to be interpreted and there are tensions between right to explanation and trade secrets.Because of this,algorithmic decision-making should be synthetically regulated,including the enhancement to social control and optimization of algorithmic application regulation.