[目的/意义]算法解释是AI治理的技术基础。正确认识AI治理中的算法解释与通用意义上的可解释性 机器学习的差异性是实现负责任人工智能（Responsible Artificial Intelligence，RAI）的关键所在。[方法/过程]从AI 事故的后解释、AI公平性的局部解释、以人为中心的算法解释三个方面阐述了AI治理中算法解释的需求特征，并 结合IBM、微软、谷歌和阿里巴巴公司的治理实践分析算法解释方法的应用。[结果/结论]AI治理中的算法解释方 法包括：以特征重要性分析和可视解释为代表的后解释实现方法、以局部扰动法和反事实解释为代表的局部解释 实现方法、多目标进化优化的解释方法。常用的算法解释评价方法有以人为中心的定性评价、基于统计指标的定 量评价和基于模糊认知的模糊评价。最后，对AI治理中算法解释的主要存在问题和未来研究应重视的发展方向 进行总结。

 [Purpose/significance] Algorithm explanationis the technical basis of AI governance. Correctly under? standing the difference between algorithm explanation in AI governance and interpretable machine learning in general sense is the key to realizing responsible artificial intelligence (RAI). [Method/process] The requirement characteris? tics of algorithm explanation in AI governance are expoundedfrom three aspects: post- hoc interpretation of AI acci? dents, local interpretation of AI fairness, and algorithm explanation of human-centered. Combined with the governance practices of IBM, Microsoft, Google and Alibaba corporations, the application of algorithm explanation methods is ana? lyzed. [Result/conclusion] Algorithm explanation methods in AI governance include post- hoc interpretation imple? mentation methods represented by feature importance analysis and visual interpretation, local interpretation implemen? tation methods represented by local perturbation methods and counterfactual interpretations, and interpretation meth? ods of multi- objective evolutionary optimization. The commonly used algorithm explanation evaluation methods in? clude human-centered qualitative evaluation, quantitative evaluation based on statistical indicators, and fuzzy evalua? tion based on fuzzy cognition. Finally, the main problems of algorithm explanation in AI governance and the develop? ment direction that should be paid attention to in future research are summarized.