How should Administrative Law Respond to Generative Artificial Intelligence——Based on Algorithms,Training Data and Content
Generative artificial intelligence(GAI)employs more complicated algorithms,requires massive training data,and enables collaborative content creation between producers and users.This exacerbates governance challenges at the administrative law level,including issues such as algorithmic opacity,decentralized control over algorithms,scarcity of high-quality training data,legality of training data,and content governance.Although the government has formulated the"Interim Measures for the Management of Generative Artificial Intelligence Services",its effectiveness is limited in addressing these administrative law governance challenges.Therefore,it is necessary to establish obligations for algorithm transparency towards users and a new algorithm security assessment system.Additionally,regulatory frameworks should be developed based on the relationships between algorithmic control entities.Regarding training data,enhancing the open access to public data,establishing a system for recording training data,and ensuring its legality are essential.For content governance,regulations should differentiate between governance during content production and dissemination stages,advocating for a lenient approach towards content production and a strict one towards content dissemination.
NETL
NSTL
万方数据
