Since the beginning of this year, government units in Shenzhen, Hangzhou, Suzhou and other places have announced the integration of the DeepSeek big model to promote the application of AI in the field of government affairs, and "AI+government affairs" has attracted attention. AI is gradually embedded in scenarios such as government offices, urban governance, livelihood services, and assisted decision-making Tian Feng, the director of the Fast Thinking and Slow Thinking Research Institute, said that "AI+government affairs" can improve decision-making efficiency. For example, Hubei Lichuan has achieved functions such as second level retrieval of policy documents and automatic correction of official document formats through the deep integration of DeepSeek big model and OA (Office Automation) system, greatly improving the efficiency of government document processing. AI+government "can also assist in precise environmental control. With the help of multimodal large models, combined with satellite remote sensing, drone monitoring, and public vehicle cameras, it can achieve perceptual analysis of urban space, which is used for environmental protection and sanitation, garbage supervision, and land conflict resolution. However, there are also some issues in the field of "AI+government affairs". Yang Haiming, Chief Technology Officer of China Unicom Digital Technology Co., Ltd., admitted that currently, the problem of "swarming" AI applications is quite serious, and the phenomenon of "using AI for the sake of using AI" is quite common. For example, some departments only announced the completion of intelligent work after integrating the full blooded version of DeepSeek big model, without seriously considering the subsequent ability improvement and in-depth application scenarios of the big model. Tian Feng also frankly stated that due to the lack of long-term strategic planning, there is also a phenomenon of "sports style" promotion of AI. Some departments lack the ability to continuously operate "AI+government" products. Some government departments collect data of low quality and lack timely updates. Some of the collected data are mostly centered around themselves, without considering market demand or mining data value based on market demand, resulting in a lot of data being idle and wasted. Tian Feng gave an example that traffic accident scene data is very valuable for intelligent driving enterprises, but such data is usually only held by government departments and rarely fed back to related enterprises. At the same time, these data will be regularly cleaned, resulting in data assets not being able to realize their high value. Tian Feng suggested that government agencies develop long-term strategies covering the application of "AI+government affairs", clarify specific action plans, and ensure effective implementation of policies. At the same time, a position similar to that of Chief AI Officer will be established to leverage the role of young innovative talents and enhance the AI literacy of civil servants. Yang Haiming suggests adhering to problem oriented and demand-oriented approaches. On the one hand, focusing on practical problems and forming an effective closed loop of "AI+government" applications; On the other hand, focusing on actual needs, conducting in-depth research on government efficiency and people's needs, and finding the next development direction. In response to data challenges, Yang Haiming suggests that the country or open source communities should build some standard datasets as a foundation, and conduct ideological checks from a technical perspective. "We should provide AI with standardized high-quality data, just like providing standardized textbooks for students." When it comes to data infrastructure construction, Tian Feng suggests integrating infrastructure resources such as computing power, networks, and energy, and building AI computing power centers and data platforms on a city or multi city basis to achieve the sharing and efficient utilization of data infrastructure resources. When promoting the application of AI in government affairs, we can first pilot high-frequency and low-risk systems, then iterate to mid low frequency and high-risk systems, establish a horse racing mechanism, and promote experience sharing and replication Tian Feng emphasized that "innovation should be encouraged and trial and error should be tolerated." Yang Haiming emphasized that potential risks such as privacy and security, ethical risks, sensitive data leaks, and erroneous judgments caused by big model illusions must also be considered when using AI for government affairs. (New Society)