- 数智赋能
- 文章编号：1009-6000(2025)07-0050-07
- 中图分类号：F49 文献标识码：A
- Doi：10.3969/j.issn.1009-6000.2025.07.008
- 项目基金：国家社会科学基金一般项目“我国网络空间内容风险治理体系构建与评价研究”（20BXW080）。
- 作者简介：张世政，博士，燕山大学文法学院（公共管理学院）讲师，硕士生导师，研究方向：数字治理；
马瑶瑶，苏州大学政治与公共管理学院本科生，研究方向：技术治理；
刘航，博士，华南农业大学公共管理学院副教授，硕士生导师，研究方向：数字经济。
- DeepSeek赋能政务服务：场景重构、运行逻辑与推进路径
- DeepSeek Enabling Government Services: Scene Reconstruction, Operation Logic and Promotion Path
- 张世政 马瑶瑶 刘航
- ZHANG Shizheng MA Yaoyao LIU Hang
- 摘要：
作为一种本土化开源大模型，DeepSeek 强大的中文语境理解能力、低算力需求、快速响应能力和多模态交互水平能够进一步推动政务服务的一体化、集约化和高效化建设，实现生成式人工智能赋能政务服务升级和智慧政务领域的场景落地。文章在总结DeepSeek 的技术优势及其在政府治理中的应用前景的基础上，结合地方政府政务服务的案例发现，DeepSeek 通过深化政务服务理念的实践进程，重构了政务服务的场景生态，具体包括整合和优化服务平台，实现精准化服务供给，增强治理对象的体验感和获得感，提升政务服务的效率等。由此，形成了技术运用、治理生态和“技术 + 治理”交互这3个层面的运行逻辑，实现了政务服务的场景创新、过程重构和环节优化。同时，DeepSeek 赋能政务服务也产生了诸如技术适配性导致的“成本—效率”问题、算法公平性风险、生成内容的可信度和可解释性不足等缺陷，为此需要依据政务服务业务需求实现精准应用，通过加强数据治理的制度建设和管理规定、提升模型可解释性和价值规范性、健全智能政务服务效果评估和责任监督体系等措施，夯实人工智能与政务服务深度融合的技术保障和价值基础。 - 关键词：
DeepSeek；政务服务；生成式人工智能；技术治理 - Abstract: As a localized open-source large-scale model, DeepSeek—featuring strong contextual understanding of the Chinese language, low computational requirements, rapid response capabilities, and multimodal interaction—can further promote the integration, consolidation, and effi ciency of government services, facilitating the application of generative artifi cial intelligence in service upgrading and the realization of smart governance scenarios. Based on a summary of DeepSeek’s technical strengths and its application prospects in governmental governance, and drawing from case studies of local public service practices, the article finds that DeepSeek has reshaped the ecosystem of government service scenarios by advancing the practical implementation of service concepts. This includes integrating and optimizing service platforms, enabling precise service delivery, enhancing the experience and satisfaction of governance recipients, and improving service effi ciency. As a result, an operational logic has formed across three dimensions: technology application, governance ecology, and “technology + governance” interaction, thereby promoting innovation in service scenarios, restructuring of service processes, and optimization of administrative procedures. However, DeepSeek’s empowerment of public services also brings certain challenges, such as the “cost-effi ciency” dilemma caused by technological adaptability, risks related to algorithmic fairness, and the limited credibility and interpretability of generated content. To address these issues, it is essential to ensure precise application of DeepSeek in line with specifi c public service needs by strengthening institutional frameworks and regulatory mechanisms for data governance, enhancing model interpretability and value alignment, and establishing robust evaluation and accountability systems for intelligent government services, thereby reinforcing the technological and ethical foundations for the deep integration of artifi cial intelligence into public administration.
- Key words: DeepSeek; government affairs service; generative artificial intelligence; technology governance