智能投顾算法透明的价值辩证、规制导向与实现路径

Abstract
Figure/Table
References
Related Citation (2)

Download: PDF (584 KB) HTML (1 KB)
Export: BibTeX | EndNote (RIS)

Abstract Robo-advisory using AI technology to analyze big data can better implement the principle of investor suitability and generate more suitable investment allocation schemes for investors, but they also conceal negative effects such as financial institutions’ profit-grabbing, risk mismatch of investment solutions, and hollowing out of the legal obligations of investment advisors. Algorithm transparency helps to break the tyranny of algorithms, eliminate algorithm discrimination, and lay the foundation for algorithm accountability, while algorithm confidentiality is necessary to maintain national scientific and technological strength, and to maintain the competitiveness of scientific and technological enterprises in the market. In the face of the value conflict between algorithm transparency and algorithm confidentiality, complete algorithm transparency and absolute algorithm confidentiality are both biased, and different degrees and levels of algorithm transparency should be realized for different subjects in different ways according to the different purposes of regulating algorithms. The hierarchy of algorithmic transparency is a specific application of the principle of proportionality, the function of measuring interests, and the proportional analysis paradigm of “ends-means” in the governance of robo-advisory algorithms. In the realization path of algorithmic transparency of robo-advisory, “compliance review + algorithmic filing” is used to realize algorithmic model transparency in order to prevent and control systemic financial risks, “algorithmic informing + algorithmic interpretation” is used to realize algorithmic logic transparency in order to protect the investors right to know. Afterward, it evaluates the degree of influence of algorithms on investors’ trust interests, and clarifies the subjectivity, causality and relevance of algorithmic decision-making, so as to determine and allocate algorithmic responsibilities.

Key words： robo-advisory algorithm transparency algorithm confidentiality value dialectics regulatory orienta-tion implementation path

Received: 07 October 2023

PACS:	D922.294
	D922.16

	Service

	E-mail this article
	Add to my bookshelf
	Add to citation manager
	E-mail Alert
	RSS
	Articles by authors
	LIU Bo-han

Cite this article:

LIU Bo-han. Value Dialectics, Regulatory Orientation and Implementation Path of Algorithm Transparency in Robo-advisory[J]. , 2024, 41(1): 103-113.

URL:

https://xb.szu.edu.cn/EN/ OR https://xb.szu.edu.cn/EN/Y2024/V41/I1/103

[1] 姜海燕,吴长凤.智能投顾的发展现状及监管建议[J].证券市场导报,2016,(12):4-10.
[2] (美)卢克·多梅尔.算法时代:新经济的引擎[M].胡小锐,钟毅译.北京:中信出版集团,2016.123.
[3] 金东寒. 秩序的重构——人工智能与人类社会[M].上海:上海大学出版社,2017.72.
[4] 杨东,武雨佳.智能投顾中投资者适当性制度研究[J].国家检察官学院学报,2019,(2):3-18.
[5] Theodor D.Sterling. Guidelines for Humanizing Compu-terized Information Systems:A Report from Stanley House[J].Communications of the ACM,1974,(11):609-613.
[6] 廖岷. 金融科技发展的国际经验和中国政策取向[M].北京:中国金融出版社,2017.88.
[7] 汪世虎,马瑞乾.金融数据安全背景下的智能投顾算法黑箱监管体系构建[J].社会科学辑刊,2022,(2):86-95.
[8] 刘博涵. 信用图像与金融刑法重塑[J].西部法学评论,2020,(6):25-37.
[9] 何颖,阮少凯.论金融产品销售商的投资者适当性义务[J].财经法学,2021,(1):134-145.
[10] 甘培忠,周淳.证券投资顾问受信义务研究[J].法律适用,2012,(10):33-39.
[11] 高丝敏. 智能投资顾问模式中的主体识别和义务设定[J].法学研究,2018,(5):40-57.
[12] Frank Pasquale.The Black Box Society:The Secret Alg-orithms That Control Money and Information[M].Cambr-idge:Harvard University Press,2015.8-11.
[13] 沈伟伟. 算法透明原则的迷思——算法规制理论的批判[J].环球法律评论,2019,(6):20-39.
[14] 张凌寒. 商业自动化决策的算法解释权研究[J].法律科学(西北政法大学学报),2018,(3):65-74.
[15] 陈景辉. 算法的法律性质:言论、商业秘密还是正当程序?[J].比较法研究,2020,(2):120-132.
[16] 毕文轩. 生成式人工智能的风险规制困境及其化解:以ChatGPT的规制为视角[J].比较法研究,2023,(3):155-172.
[17] 刘琳. 算法解释权与商业秘密保护的冲突化解[J].行政法学研究,2023,(2):168-176.
[18] Davide Castelvecchi.Can We Open the Black Box of AI?[J].Nature,2016,(10):20-23.
[19] 安晋城. 算法透明层次论[J].法学研究,2023,(2):52-66.
[20] 冷传莉. 比例原则私法化的体系定位与调整对象[J].比较法研究,2023,(4):99-116.
[21] 卢永欣. 利益和原则:马克思恩格斯思想的重要线索[J].贵州大学学报(社会科学版),2021,(3):10-15.
[22] 蒋红珍. 比例原则适用的范式转型[J].中国社会科学,2021,(4):106-127.
[23] 许可.驯服算法:算法治理的历史展开与当代体系[J].华东政法大学学报,2022,(1):99-113.
[24] 金小野.规范高频交易是国际证券业监管焦点[N].法制日报,2013-11-12(010).
[25] 徐凤. 人工智能算法黑箱的法律规制——以智能投顾为例展开[J].东方法学,2019,(6):78-86.
[26] 解正山. 算法决策规制——以算法“解释权”为中心[J].现代法学,2020,(1):179-193.
[27] 万方. 算法告知义务在知情权体系中的适用[J].政法论坛,2021,(6):84-95.
[28] 丁晓东. 基于信任的自动化决策:算法解释权的原理反思与制度重构[J].中国法学,2022,(1):99-118.
[29] 辛巧巧. 算法解释权质疑[J].求是学刊,2021,(3):100-109.
[30] Margot E.Kaminski.Binary Governance:Lessons from the GDPR’s Approach to Algorithmic Accountability[J]. Southern California Law Review,2019,(6):1529.
[31] Ari Ezra Waldman.Power,Process,andAutomated Deci-sion-Making[J].Fordham Law Review,2019,(2):613-632.
[32] 胡改蓉,钱程.投资者适当性管理制度中金融机构的免责机制——以“信赖利益”判断为核心[J].证券市场导报,2021,(10):61-70.
[33] 张恩典. 算法影响评估制度的反思与建构[J].电子政务,2021,(11):57-68.