生成式人工智能驱动公共服务共同生产的内在逻辑与风险应对OACHSSCD

With the continuous iteration of digital intelligence technology,generative artificial intelligence(AI)has become one of the core actors in public services co-production,forming a new joint production state of"government-public-technology"together with traditional actors.Generative AI systematically drives the co-production of public services by reshaping the goal of co-production,activating the independent participation of the public and fostering the bridging environment of subject cooperation.However,there are also some embedded risks in generative AI,such as algorithm black box,technology monopoly and digital ethics,which will bring adverse effects on the cooperation boundary,synergistic motivation and value consensus between the government and the public,and thus hinder the sustainability of public service co-production.In order to avoid the above risks and promote technology for good,it is necessary to construct countermeasures from three aspects.First,regulating technology through institutional norms,which means clarifying the action boundary of subject cooperation by improving the legal framework,perfecting the list of powers and responsibilities,and implementing the whole process of governance.Second,reshaping technology through organizational construction,which means inspiring the endogenous motivation of co-production by enhancing organizational resilience,creating a collaborative culture,and cultivating digital abilities.Third,governing technology through technological means,which means gathering the ethical consensus of the government and the public by constructing a regulatory sandbox,establishing a circuit breaker mechanism,and promoting trustworthy AI.This study focuses on the discussion of the co-production logic and process of public services driven by generative AI,responding to the theoretical and practical concerns about how to achieve subject cooperation with technology so as to promote the high-quality supply of public services in the era of digital intelligence.