As an important entrance of mobile internet to intelligent internet, face recognition technology widely permeates various industries, continuously supports multiple needs of massive users in complex scenes, and actively empowers large-scale, multilevel and fragmented innovation activities of the industrial chain, which has become the core force to consolidate and expand intelligent industry solutions. During the whole life cycle of face recognition application, from capturing face data to completely deleting it, all kinds of violations infringe on privacy, freedom, equality and security interests of the target group. It is urgent to construct and perfect control strategies on infringement risks of face recognition under different application scenarios, so as to dynamically balance public welfare and private benefit, promote the smooth landing of National Development Strategy of Artificial intelligence.