自动决策算法的法律规制:以数据活动顾问为核心的二元监管路径

Legal Regulation of Automated Decision-making Algorithms: Dual Regulatory Path with Data Activity Consultant as the Core

大数据和人工智能的发展使得以自动决策为特征的算法被广泛使用,这也使得算法透明化的目标愈发紧迫。强制公开算法路径、验证或认证算法路径、个人算法解释权和算法结果控制等都是对算法监管的有益探索,但或因有违基本的制度期待,或因适用范围有限,而无法成为算法监管的有效手段。算法监管涉及政治、经济和法律三个社会子系统,应当通过数据活动顾问这一"接口岗位"实现系统间的结构耦合,从而借助系统间的协力有效监管算法活动。有鉴于此,我国应当设立以数据活动顾问为主、数据活动监管局为辅的二元算法监管机制:数据活动顾问从企业内部对算法的设计、使用和评估进行陪同控制;数据活动监管局则从外部对数据活动顾问提供必要的援助和监督。

The development of big data and artificial intelligence is making algorithms, characterized by automatic decision, widely used, which correspondently makes the need of algorithm transparency much more urgent. The idea of mandatory disclosure, verification or certification of algorithm, personal right of algorithm interpretation or the control of the algorithm result is though inspiriting, but may not be an effective way to supervise the use of algorithm, because it either violates the basic expectation of people or has a very limited application scope. Three social subsystems are involved in Algorithmic supervision, specifically political, economic, and legal system. As a Result, data activity consultant should be made to play the "interface role" to realize the structural coupling between systems, so that different subsystems would cooperate to effectively supervise the algorithm activities. China should set up a binary algorithm supervision mechanism, in which data activity consultants work primarily and data activity supervisory authority plays a secondary role to assist. Detailly, data activity consultant has an accompanying control of the design, use, and evaluation of algorithms within the enterprise;meanwhile, the data activity supervisory Authority provides data activity consultants necessary assistance and supervision from outside.