论人工智能的风险规制  

Regulatingg the Potential Risks of Artificial Intelligence

作　　者：周学峰[1] Zhou Xuefeng

机构地区：[1]北京航空航天大学法学院

出　　处：《比较法研究》2024年第6期42-56,共15页Journal of Comparative Law

摘　　要：人工智能风险是随着人工智能应用而生的。对人工智能风险的识别并不是单纯的技术问题,而是涉及复杂的伦理、价值与法律判断。人工智能风险具有不确定性、不透明性、泛在性、变动性等特点,从而给人工智能风险规制带来挑战。在对人工智能风险进行规制时,其立法模式宜采用总分结合的方式,在采用行为规制模式的同时应辅之以赋权模式,在选择规制方案时应采用基于风险的规制模式。人工智能风险规制的目标并不是排除所有的风险,而是将风险控制在合理的范围内,因此,人工智能致损难以完全避免,事后救济机制的存在是十分必要的。Risks may arise when artificial intelligence is put into practice.Identifying these risks is not just a technical issue;it involves intricate ethical,moral,and legal considerations.The risks associated with artificial intelligence,such as uncertainty,opacity,ubiquity,and variability,pose challenges to effectively managing these risks.In crafting regulations for artificial intelligence risks,a legislative approach that combines broad,overarching laws with specific sector-specific regulations is recommended.In addition to a regulatory framework focused on behavior,an empowerment model should be integrated.When selecting regulatory strategies,a risk-oriented approach should be adopted.It is crucial to keep the ex post remedy mechanisms,as the ex ante regulations alone cannot eliminate all risks associated with artificial intelligence.

关 键 词：人工智能 风险 规制 基于风险的规制 

分 类 号：D99[政治法律—国际法学] TP18[政治法律—法学]