摘要
人工智能的发展有赖于用户能够对其理解、信任和管理,但人工智能算法的潜在不透明性导致用户与人工智能产品的信任危机日益加重。人工智能算法可能会带来法律风险,如侵犯用户的知情权、决策权与隐私权,以及算法歧视、算法对抗等不利后果。目前,我国已初步具备算法规范,但人工智能算法信任机制还未成体系。以信义法规范调整人工智能与用户间的权利义务关系,具有制度优越性。同时,应从算法备案制度着手,完善算法授权机制,及时进行算法纠偏,保障用户的算法选择权。在宏观层面,建立可解释性人工智能规范体系亦是核心举措。总而言之,人工智能算法信任机制应是系统性规范体系,并能促进人工智能产业健康有序发展。
The development of artificial intelligence depends on users being able to understand,trust and manage it,but the potential opacity of artificial intelligence algorithms has led to an increasing crisis of trust between users and artificial intelligence products.Artificial intelligence algorithms may bring about legal risks,such as infringing users'right to be informed,decision-making and privacy,as well as algorithmic discrimination,algorithmic confrontation and other adverse consequences.Although China has initially had the algorithm regulatory regime,trust mechanism of artificial intelligence algorithm has not yet been established.Regulating the relationship of rights and obligations between artificial intelligence and users by fiduciary law has many merits.At the same time,we should start with algorithm filing system,improve the algorithm authorization mechanism and timely algorithm correction,and ensure the user's algorithm choice.The establishment of an explainable AI specification system at the macro level is also a core initiative.The trust mechanism of artificial intelligence algorithm should be a systematic normative system,and can promote the healthy and orderly development of the artificial intelligence industry.
出处
《数字法治》
2024年第1期123-135,共13页
DIGITAL LAW
基金
最高人民法院2022年度司法研究重大课题“算法技术的法律规制研究”(项目编号:ZGFYZDKT202211-02)的阶段性研究成果
作者简介
狄行思,广州大学法学院讲师;宋歌,山西省大同市中级人民法院法官助理。
