陈桂月, 贾春岩, 陈智伟, 白慧君. 生成式人工智能在医学科研管理中的应用进展研究. 2026. biomedRxiv.202610.00048
生成式人工智能在医学科研管理中的应用进展研究
通讯作者: 白慧君, camsacc@cams.cn
DOI:10.12201/bmr.202610.00048
Application Progress of Generative Artificial Intelligence in Medical Research Management
Corresponding author: bai huijun, camsacc@cams.cn
-
摘要:〔摘要〕目的/意义 系统调研生成式人工智能在医学科研管理中的技术演进与应用现状,分析其在数据治理、流程优化和决策支持方面的进展与难点,为其应用提供参考。方法/过程 采用系统文献综述方法,对生成式人工智能在科研管理工作应用从技术演进、管理流程应用、数据治理与隐私保护、决策可信度等维度开展系统分析,并选取三个典型案例深入剖析其技术路径与治理实践。结果/结论 生成式人工智能已在科研管理文本解析、资助数据分析和工作流自动化等场景显示应用价值,检索增强生成、知识图谱和人机协同机制有助于提升输出的可追溯性与可靠性;但仍面临模型幻觉、跨机构数据标准不一、部署成本较高及权责机制不清等问题。建议由行业主管部门、科研机构和技术支撑部门协同推进语义治理、分级风险管理和应用评价,加快数据术语建设、概念映射与业务规则的语义治理,明确人机分工边界,并加强专业人才的AI素养培训。
Abstract: 〔Abstract〕Purpose/Significance: To systematically review the technological evolution and application of generative AI in medical research management, analyzing its progress and challenges in data governance, process optimization, and decision support to provide references for its application. Method/Process: A systematic literature review was conducted across technological evolution, management processes, data governance and privacy protection, and decision credibility, with three typical cases analyzed for their technical pathways and governance practices. Result/Conclusion: Generative artificial intelligence has demonstrated practical value in scenarios such as textual parsing for research administration, grant funding data analysis, and workflow automation. Retrieval-augmented generation, knowledge graphs, and human-AI collaboration mechanisms help improve the traceability and reliability of outputs. Nevertheless, challenges remain, including model hallucinations, inconsistent cross-institutional data standards, high deployment costs, and ambiguous accountability frameworks. It is recommended that industry authorities, research institutions, and technical support departments collaboratively advance semantic governance, tiered risk management, and application evaluation. Efforts should be accelerated to develop data terminology standards, implement concept mapping, and enforce semantic governance of business rules, clarify the division of responsibilities between humans and AI, and strengthen AI literacy training for professionals.
Key words: Generative artificial intelligence; Medical research management; Large language model; Data governance; Research management informatization提交时间:2026-10-21
版权声明:作者本人独立拥有该论文的版权,预印本系统仅拥有论文的永久保存权利。任何人未经允许不得重复使用。 -
图表
-
孟冬晴, 吕文娟. 生成式人工智能赋能医学图书馆知识服务的路径研究. 2025. doi: 10.12201/bmr.202510.00039
中国科学技术信息研究所, 北京市海淀区复兴路号. 基于生成式人工智能的医学视频创作路径探究. 2025. doi: 10.12201/bmr.202508.00005
弓孟春, 李雨杭, 马永慧, 弓凯, 刘超, 欧阳自豪, 戴辉. 生成式医学人工智能(GMAI)的临床伦理治理:三维协同路径与中国实践. 2025. doi: 10.12201/bmr.202511.00057
史森中. 生成式人工智能(GAI)基层医疗应用风险及治理策略分析. 2024. doi: 10.12201/bmr.202408.00053
张世红, 史森, 白玲, 杨小冉, 韩冬, 胡红濮. 人工智能在公共卫生应急管理的应用场景分析及挑战. 2025. doi: 10.12201/bmr.202511.00063
芮晨, 陈玥琪, 李金斌, 张胜发. 人工智能在基层2型糖尿病健康管理中的应用进展与趋势. 2025. doi: 10.12201/bmr.202506.00072
郏欣茹, 施彦, 姚东伟, 林鹏耀, 林呈飞. 生成式人工智能辅助教学法在眼科临床教学中的应用研究. 2026. doi: 10.12201/bmr.202603.00124
王蕾, 刘苗, 王茜, 赵琬清, 胡佳慧, 方安. 生成式人工智能医学语料库:数据风险、合规义务及应对方案. 2025. doi: 10.12201/bmr.202512.00083
向艾宁堃, 田靖雪, 胡德华, 刘海霞. 生成式人工智能对于老年糖尿病应答效能的比较研究. 2025. doi: 10.12201/bmr.202503.00019
马傲, 葛小玲. 人工智能大模型在医疗健康领域应用的研究. 2024. doi: 10.12201/bmr.202408.00039
-
序号 提交日期 编号 操作 1 2026-06-23 10.12201/bmr.202610.00048V1
下载 -
-
公开评论 匿名评论 仅发给作者
引用格式
访问统计
- 阅读量:50
- 下载量: 0
- 评论数:0

登录
注册




京公网安备