[Purpose/significance] Artificial Intelligence (AI) has become a central issue in global technological devel⁃opment. The sociotechnical imaginaries formed in different countries, that is, the collective visions of the future of AI are profoundly shaping national policy formulation and development planning. This study conducts a comparative analy⁃sis of different region’s AI policy documents to uncover emotional orientations and policy trajectories. The findings aimto provide references for the formulation and refinement of AI policy in China. [Method/process] A total of 35 key AI policy documents published between 2016 and 2024 by China, the US, the UK, India, and the EU were selected for analysis. Using computational topic modeling and sentiment analysis, the study explores the sociotechnical imaginaries and sentiments reflected in these policy texts. [Result/conclusion] China, the US, the UK, India, and the EU placed strong emphasis on AI chip research and shared aspirations for AI to empower social development and enhance human well-being. Meanwhile, concerns about data misuse, algorithmic bias, and technical robustness and security vary in in⁃tensity. The Global South focus more on technological inclusion and economic transformation, while those in the Global North place greater emphasis on regulatory frameworks and risk mitigation. The Global North has generally exhibited a long-standing preference for a "governance-first" paradigm, while the Global South is transitioning from a "develop⁃ment-first" stance toward a more balanced approach that integrates both development and governance. Overall, the sen⁃timent expressed in AI policies has gradually shifted from optimism to a more neutral and cautious outlook.