Abstract: As generative artificial intelligence (AI) technologies like ChatGPT become more widely used, they are increasingly being integrated into government governance, triggering a series of transformations. The special nature of government governance requires an open attitude to adopt emerging technologies to enhance efficiency, but it also necessitates cautious consideration of the potential risks and challenges posed by technological applications. By constructing a “Technology-Institution-Value” analytical framework, this paper explores the prospects, potential risks, and preventive mechanisms for the integration of AI into government governance. The integration of generative AI can effectively gather comprehensive information for government decision-making, improve operational efficiency, enhance the human-machine interaction experience in public service provision, and increase the accuracy, objectivity, and fairness of government governance. However, it is crucial to be vigilant about potential risks such as data security issues, over-reliance on technology, social security threats, weakening of public authority, algorithmic discrimination, and the absence of accountability in governance. Following the “Technology-Institution-Value” theoretical framework, this paper proposes risk prevention mechanisms through strengthening security technology research, enhancing algorithm interpretability, improving accountability systems, improving leadership systems, enhancing the digital capabilities of cadres, eliminating algorithmic bureaucracy, and adhering to inclusive and prudent principles to achieve the high-quality development of technology, ensuring it is used wisely and effectively for “good governance” and “intelligent governance”.