金融科技公司算法黑箱风险的生成机制与多维治理  

The Generation Mechanism and Multi-Dimensional Governance of Algorithm Black Box Risk in Financial Technology Companies

作　　者：程雪军 CHENG Xue-jun

机构地区：[1]同济大学法学院,上海200092

出　　处：《金融论坛》2023年第8期25-34,46,共11页Finance Forum

基　　金：上海市哲学社会科学规划项目“金融科技公司算法风险的系统治理研究”(2022EFX001);中央高校基本科研业务费专项资金资助项目“超级平台算法垄断的系统治理机制研究”(22120230353)。

摘　　要：计算机与人工智能等技术的深度演进,推动金融科技公司的供给结构变化。金融科技公司将算法技术深度嵌入金融服务,大幅提升金融产品的运行效率与普惠效果。然而,算法技术总是存在着一定的黑箱,并形成新兴的算法黑箱风险。从金融科技公司算法黑箱风险的内部结构来看,算法黑箱具有结构不透明风险、技术不成熟风险、信息不对称风险。中国需要有针对性地构建金融科技公司算法黑箱风险的多维治理:从市场端健全金融科技公司算法实质性审查标准,从技术端强化金融科技公司算法黑箱的科技治理,从消费者端赋予金融消费者的算法同意权与解释权。The deep evolution of technologies such as computers and artificial intelligence has promoted the changes in the supply structure of financial technology(fintech)companies.Fintech companies embed algorithm technology into financial services deeply,and improve the operational efficiency and inclusive effect of financial products greatly.However,there is always a certain black box in the algorithmic technology,and it results new algorithmic black box risks.As for the internal structure of algorithmic black box risks in fintech companies,algorithmic black boxes have the risks of structural opacity,technological immaturity,and information asymmetry.China needs to build targeted multi-dimensional governance for the algorithmic black box risk of fintech companies:improve the substantive review standards for fintech companies'algorithms from the market side,strengthen the technological governance of fintech companies'algorithmic black boxes from the technology side,and grant financial consumers the algorithmic consent and interpretation rights from the consumer side.

关 键 词：金融科技公司 算法黑箱 信息不对称风险 生成机制 多维治理 

分 类 号：O17[理学—数学] O13[理学—基础数学]