生成式人工智能的行政监管困境及其纾解

Administrative Supervision Dilemma of Generative Artificial Intelligence and Its Solutions

生成式人工智能的开放性、复杂性和高速迭代性给行政监管带来对象难以确认、主体能力不足以及过程较为封闭的治理难题。人工智能监管的现行法律应对其负起责任,认真对待监管对象转移、组合式监管乏力以及监管程序缺漏的问题。对此,应引入整体性治理理论,将监管对象整合为生成式人工智能的大模型产品,将监管资源整合至网信部门,由其主要负责生成式人工智能的监管,并通过构建整体性的程序机制,实现监管程序的民主化与规范化。

The openness,complexity and high-speed iteration of generative artificial intelligence bring administrative supervision with difficult to confirm objects,insufficient subject capacity and relatively closed process.The current law on AI regulation should take responsibility and seriously treat the transfer of regulatory objects,the lack of combined supervision and the lack of regulatory procedures.In this regard,we should introduce the holistic governance theory,integrate the supervision objects into large model products of generative artificial intelligence,integrate the regulatory resources into the cyberspace administration which will be mainly responsible for the supervision of generative artificial intelligence,and realize the democratization and standardization of the supervision procedure through the construction of an holistic procedure mechanism.