As a leading technology in the digital intelligence era， generative artificial intelligence finds its application in policy dissemination， which is prominently demonstrated in aspects such as the intelligent upgrade of each element within the policy dissemination structure and the continuous optimization of the overall dissemination structure. Specifically， it transforms the subject form of policy dissemination， driving their development towards the direction of human-machine interaction， human-machine co-construction， and human-machine symbiosis. It also alters the generation mode of policy dissemination content， improves the matching efficiency between the supply and demand of policy dissemination information， promotes the development and intelligence of integrated media， and comprehensively improves and optimizes the policy dissemination structure， thereby promoting the healthy development of policy dissemination. However， the application of generative artificial intelligence in policy dissemination is extremely complex and will also face non-negligible risks， such as the weakening of the “gatekeeper” status of policy dissemination subjects， the reinforcement of “information cocoon” of policies， the out-of-control spread of “policy public opinion”， and the continuous emergence of “policy rumors”. To avoid these risks， it is imperative to heighten the in-depth and comprehensive understanding of the entire populace regarding digital technologies represented by generative artificial intelligence and the social transformations they instigate； strengthen the construction of scientific and technological ethics and morality in the whole society， and focus on improving the technological ethics literacy of policy dissemination subjects and audiences； attach importance to the unity of the instrumental value and humanistic value of policy dissemination； and regulate the governance of generative artificial intelligence technology in accordance with the law.