Intelligent Media, as an AI-driven Information Ecosystem, relies on the synergy between information and users for its operation. At present, the 3 key links of information production, dissemination and user feedback are facing new AI ethics challenges, and there is an urgent need for a compatible AI ethical governance framework to maintain the balance of the information ecosystem and protect the rights of users. Through literature research and logical discernment, the article combs through the concepts and contents of AI ethics, retraces the history of intelligent media, analyzes specific AI ethics challenges based on the dual perspectives of information and users, and puts forward a targeted governance framework that includes the positioning of main bodies and governance paths. The article argues that AI ethics include 19 items at 4 levels: AI system, social well-being, literacy enhancement, and responsibility traceability; intelligent media is undergoing an important shift from technology-driven to HAII. There are 10 AI ethics challenges in the 3 key links, such as the proliferation of false information, privacy and data protection, etc. The 5 main governance bodies, namely AI designers and developers, intelligent media operators, users, government, and social organizations, can realize intelligent media AI ethical governance through 19 paths, such as user-oriented, dataset fairness assessment, and so on.