作　者：王德夫
关键词：人工智能技术；算法治理；知识产权；算法可理解；数据要素市场；数据安全；数据可信
摘　要：

人工智能算法专注于解决“如何实现智能”这一特定的技术问题，是一种特殊的技术方案，相关的法律治理也围绕这一基本属性而展开。在人工智能算法设计和自我完善过程中所产生的新知识与相关利益，可以通过知识产权制度进行确认和分配；而对于如何平衡投资利益与社会公众利益，以及如何使法律监管突破黑盒障碍等问题，则应设计专门制度予以解决。通过创新性的“算法可理解+数据可信+参数可解释”治理架构，结合算法识别、数据可信以及算法可理解等基础性规则，可以突破算法解释、平台责任等现有治理手段的局限，确保技术理性与社会发展的协调与相互促进。

 

On the Legal Attribute and Governance Approach Of Artificial Intelligence Algorithm

Wang Defu  (Wuhan University)

Abstract Artificial intelligence algorithm aims to solve the technical problem of how to achieve intelligence. As a special technical solution, "artificial intelligence algorithm" has not only the relatively traditional intellectual property path, but also the innovative special law governance path. New knowledge and relevant interests are derived in the process of artificial intelligence algorithm design and self-perfection. Specific solutions should be designed to strike a balance between investment interests and social welfare. Based on the framework of "algorithm comprehensible + data credible + parameter interpretable", the comprehensive governance system and basic rules of artificial intelligence algorithm are designed, and the "individual algorithm interpretable" and "platform responsibility" are clarified in order to ensure the coordination and mutual facilitation of technical reasoning and social development.

Key words artificial intelligence technology; algorithm governance; intellectual property; algorith‐ munderstandability; data element market; data security; data credibility

 

■作者简介 王德夫，法学博士，武汉大学法学院讲师，武汉大学知识产权与竞争法研究所、武汉大学网络治理研究院助理研究员；湖北 武汉 430072。

[PDF]（下载数：）