"支持AI融合赋能"是上海促进服务业创新发展的重要任务，也是推进AI技术产业化、商业化及应用场景 探索的关键。民革上海市委会在调研中发现，当前上海大模型企业在拓展服务业智能化转型市场时，大 多采用传统"点对点、单对单"模式，与大模型技术特点不匹配，且AI复合型人才严重缺失制约技术商业 化与场景探索。 民革上海市委会建议，提升服务业标准化水平，强化行业级高质量语料供给，鼓励行业协会联合AI企 业及服务业龙头制定细分领域标准；强化复合型人才支撑，打造全链条人才体系；打造多层次交流平 台，激发企业转型积极性。 实施"人工智能+"行动 上海两会求解发展与治理双重命题 中新网上海2月4日电(范宇斌)正在举行的上海市两会上，以大模型为代表的通用人工智能引发的科技革 命与产业变革成为热议焦点。 上海正全力构筑人工智能发展高地。今年上海市政府工作报告提出，将深入实施"人工智能+"行动，加 强算力设施、行业语料、垂类模型等布局建设，积极培育智能原生新模式新业态。 "人工智能驱动科学研究"(AI for Science，简称AI4S)被视为发展新质生产力的前沿领域。上海市人大代 表、优刻得科技股份有限公司董事长季昕华受访时表示， ...

当前，以大模型为代表的通用人工智能正引发新一轮科技革命和产业变革，其中"人工智能驱动科学研究（AI4S）"是孕育颠覆性创 新、发展新质生产力的核心前沿。上海在产业生态、技术积累和应用场景方面构建了显著优势，已经形成从芯片算力、算法框架到 行业应用的全栈人工智能产业链。 为此，季昕华建议，在上海率先建设科学智能创新示范区，为上海科创中心建设和人工智能产业发展作出积极贡献。一方面要在现 有市级算力调度体系内，专项扩容建设"AI4S公共算力服务池"，整合高校、科研院所及部分企业的闲置算力，统一纳管调度。以现 有成熟的专业孵化平台（如启迪之星等）与特色产业园区为抓手，提供容器化、预置多学科软件、集成开发环境、项目组管理等功 能的云服务门户，向科研团队和中小企业提供普惠性算力支持。并且发放算力补贴，鼓励和补贴科研团队优先使用基于国产AI芯片 的算力进行科学计算与模型训练。 上海市人大代表、优刻得董事长季昕华对《国际金融报》记者表示，在推动AI深度赋能高风险、高价值的科学研究领域时，仍面临 系统性挑战。季昕华建议，在上海率先建设科学智能创新示范区，把杨浦打造成为AI4S创新示范区，因为杨浦区内人工智能企业呈 集群式发展， ...

Core Insights - The global technological competition is advancing towards foundational research and interdisciplinary fields, with AI for Science (AI4S) becoming a core engine for achieving high-level technological self-reliance and reconstructing scientific research paradigms [1] Group 1: Challenges in AI4S Development - The field of condensed matter science faces challenges in developing AI4S due to its reliance on researcher experience and intuition, leading to high trial-and-error costs and long R&D cycles [2] - The transition from a "human-intensive" trial-and-error approach to a "data and intelligence-intensive" rational design and prediction is expected to significantly enhance R&D efficiency and shorten the cycle from basic research to industrial application [2] Group 2: Data Quality and Governance - High-quality scientific data is the foundation of AI4S, with researchers working to build data aggregation and integration platforms to optimize AI model performance [3] - Current issues in condensed matter science data include resource scarcity, severe data isolation, insufficient data volume, inconsistent standards, and a lack of effective data aggregation mechanisms [3] - Data governance technology is a driving force for AI4S, with research teams developing efficient data processing algorithms and tools to enhance data governance and support scientific breakthroughs [4] Group 3: Innovation Ecosystem - A data innovation ecosystem is essential for the sustainable development of AI4S, with efforts to create research platforms and communities that support data sharing and collaboration [5] - Existing practices include standardized evaluation systems for models related to crystal material topology and XRD intelligent structure analysis, which support large-scale, collaborative data-driven technological innovation [5] Group 4: National Strategies and Initiatives - The digital transformation of condensed matter science has become a national strategy for major economies, with initiatives like the U.S. "Genesis Project" and the U.K. "National AI Strategy" focusing on the deep integration of condensed matter science and AI [6] - China is actively exploring this field, with the Chinese Academy of Sciences' Condensed Matter Science Data Center making systematic progress in integrating experimental, theoretical, and computational data [6] Group 5: Future Directions for Development - To further promote AI4S in condensed matter science, efforts should focus on building standardized, high-quality national data foundations and enhancing data collection, governance, application, and sharing services [7] - Creating an open, collaborative, and sustainable data ecosystem is crucial, along with fostering deep integration between industry, academia, and research [7] - Breakthroughs in data technology will enhance China's independent innovation capabilities and international competitiveness in condensed matter science [7]