金融科技的算法风险及其法律规制

随着算法在金融领域应用的深入,算法固有的缺陷和特性也逐渐与金融本身的风险和逻辑发生耦合,形成了算法歧视、算法绑架、算法趋同等新型金融风险。然而,现有的金融科技算法规制框架存在主体、对象和工具层面的诸多问题,导致金融科技算法风险规制的低效和失灵,因而需要进行规制路径的优化和调适。从主体层面出发,应构建以金稳会为基础的规制协调机制和监管者、行业协会、企业分工合作的多元规制格局;从对象层面着手,应在将所有算法应用业态纳入统一规制框架的基础上,确立以"数据规模"为标准的规制门槛;从工具层面调整,应构建以"快捷沙盒"为制度载体的算法测试制度和可选择的算法披露机制,并积极探索自动化监管模式和代码规制的落地和应用。

Algorithm Risk of Fintech and its Legal Regulation

With the sustained and in-depth application of algorithm in the financial industry,the inherent defects and characteristics of algorithm are gradually coupling with the risks and logic of finance itself,and consequently forming new types of risks such as algorithm discrimination,algorithm abduction,algorithm convergence and so on.However,the existing regulatory framework and legislation on algorithm of fintech have many problems at the level of subject,object and tool,which lead to the inefficiency and failure of regulating algorithm risk of fintech.Therefore,it is necessary to restructure and adjust the regulatory approach and model.At the subject level,it is imperative to construct a regulatory coordination mechanism based on the Council for Financial Stability and Development of the State Council,and a diversified regulatory framework with cooperation among supervisors,industry associations and enterprises.At the object level,a regulatory threshold based on"data scale"should be established on the basis of integrating all algorithmic application formats into a unified regulatory framework.In terms of tools,it is necessary to establish an algorithm test system based on"sandbox express"and an optional algorithm disclosure mechanism,and actively explore the implementation and application of automatic supervision and regulation by code when regulating the algorithm risk of fintech.