| 22 | 0 | 12 |
| 下载次数 | 被引频次 | 阅读次数 |
人工智能概念提出以来,医学便是其核心应用领域。我国多项政策推动医疗人工智能落地应用,当前已深度融入智慧医院各流程,但基座模型漏洞、数据安全等系统性风险制约其发展。本研究通过分析医疗人工智能应用现状、技术特性及部署模式,识别模型算法、数据、合规伦理等多层风险,提出构建“四道防线”全域安全防护架构,配套“算力-数据-模型”协同管理策略与“组织-技术-制度”协同治理机制。实现医疗大模型全生命周期安全加固,为其在安全可控、伦理可信基础上稳健可持续发展提供支撑,助力医疗人工智能高质量发展。
Abstract：Since the concept of artificial intelligence(Al) was proposed,medicine has been one of its core application fields.Driven by a number of national policies,medical AI has now been deeply integrated into all processes of smart hospitals.However,its development is constrained by systemic risks such as foundation model vulnerabilities and data security issues.By analyzing the application status,technical characteristics and deployment modes of medical Al,this paper identifies multi-level risks in terms of model algorithms,data management,compliance and ethics.It further proposes the construction of an all-round security protection architecture of the "Four Lines of Defense",coupled with a collaborative management strategy of computing power-data-model and a synergistic governance mechanism integrating organization,technology and institutions.This research is expected to realize the full-life-cycle security reinforcement of medical large models,provide support for their steady and sustainable development on the basis of safety,controllability and ethical credibility,and facilitate the high-quality development of medical Al.
[1]吴敏敏,王鑫钰,王伟炳.医学人工智能的技术发展和场景应用[J].复旦学报(医学版),2025,52(3):470-474.
[2]国家卫生健康委办公厅,国家发展改革委办公厅,工业和信息化部办公厅,等.关于促进和规范“人工智能+医疗卫生”应用发展的实施意见:国卫办规划发[2025] 30号[EB/OL].(2025-10-20)[2025-12-10].https://www.gov.cn/zhengce/zhengceku/202511/content 7047018.htm.
[3]国家卫生健康委办公厅,国家中医药局综合司,国家疾控局综合司.国家卫生健康委员会办公厅关于印发卫生健康行业人工智能应用场景参考指引的通知:国卫办规划函[2024]420号[EB/OL].(2024-11-06)[2025-01-26].https://www.nhc.gov.cn/guihuaxxs/c100133/202411/3dee425b8dc34f739d63483c4e5c334c.shtml.
[4]国家卫生健康委办公厅,国家发展改革委办公厅,工业和信息化部办公厅,等.关于促进和规范“人工智能+医疗卫生”应用发展的实施意见:国卫办规划发[2025] 30号[EB/OL].(2025-10-20)[2025-12-10].https://www.nhc.gov.cn/guihuaxxs/c100133/202511/d1a42ae835c743b9b3e83ac0253c3e9f.shtml.
[5]国务院办公厅.国务院办公厅关于加快场景培育和开放推动新场景大规模应用的实施意见:国办发[2025] 37号[EB/OL].(2025-11-02)[2025-11-27].https://www.gov.cn/zhengce/zhengceku/202511/content 7047420.htm.
[6]方明旺,郭玲,黄应德,等.生成式人工智能在医院管理领域的应用前景与挑战探讨[J].医学信息学杂志,2024,45(10):18-21.
[7]BROWN T,MANN B,RYDER N,et al.Language models are few-shot learners[J].Advances in Neural Information Processing Systems,2020,33:1877-1901.
[8]GALLIFANT J,FISKE A,STREKALOVA L A Y,et al.Peer review of GPT-4 technical report and systems card.[J].PLoS Digital Health,2024,3(1):e0000417.
[9]CHOWDHERY A,NARANG S,DEVLIN J,et al.Palm:Scaling language modeling with pathways[J].Journal of Machine Learning Research,2023,24(1):113.
[10]TOUVRON H,LAVRIL T,IZACARD G,et al.Llama:Open and efficient foundation language models[J].arXiv preprint arXiv:2302.13971,2023.
[11]TOUVRON H,MARTIN L,STONE K,et al.Llama 2:Open foundation and fine-tuned chat models[J].arXiv preprint arXiv:2307.09288,2023.
[12]DUBEY A,JAUHRI A,PANDEY A,et al.The llama 3 herd of models[J].arXiv preprint arXiv:2407.21783,2024.
[13]RAY P P.ChatGPT:A comprehensive review on background,applications,key challenges,bias,ethics,limitations and future scope[J].Internet of Things and Cyber-PhysicalSystems,2023,3:121-154.
[14]LIU Y,HAN T,MA S,et al.Summary of chatgpt-related research and perspective towards the future of large language models[J].Meta-radiology,2023,1(2):100017.
[15]萧文科,宋驰,陈士林,等.中医药大语言模型的关键技术与构建策略[J].中草药,2024,55(17):5747-5756.
[16]BENGESI S,EL-SAYED H,Sarker M K,et al.Advancements in generative AI:A comprehensive review of GANs,GPT,autoencoders,diffusion model,and transformers[J].IEEe Access,2024,12:69812-69837.
[17]SHEN T,LI Y,GAO Y,et al.Rapid deployment of large language model DeepSeek in Chinese hospitals demands a regulatory response[J].Nature Medicine,2025,31(10):3233-3238.
[18]薛飞,陈彬,刘静,等.自智网络全栈部署技术研究与实践[J].电信科学,2023,39(9):63-74.
[19]张姝艳,皮婷婷.医疗领域中人工智能应用的可解释性困境与治理[J].医学与哲学,2023,44(3):25-29,35.
[20]陈子瑜,程国斌.医疗人工智能中的算法黑箱及其核心伦理问题[J].医学与哲学,2024,45(12):6-10.
[21]ISO/IEC 42001:2023 Artificial intelligence—Management system[S].Geneva:ISO/IEC,2023.
[22]邱景,赵笑尘,胡徐茜,等.大数据医疗时代的人工智能与患者个人隐私保护的技术方案研究[J].国外电子测量技术,2025,44(7):294-300.
[23]罗妍,刘宇炀,李晓瑛,等.面向医学大模型的体系化人工智能框架构建与应用[J].北京邮电大学学报,2024,47(4):98-104.
[24]国家互联网信息办公室.生成式人工智能服务管理暂行办法[EB/OL].(2023-07-10).https://www.cac.gov.cn/2023-07/13/c 1690898327029107.htm.
[25]World Health Organization.Ethics and governance of artificial intelligence for health:large multi-modal models.WHO guidance[M].World Health Organization,2024.
基本信息:
中图分类号:R197.32;TP18
引用信息:
[1]张懿文,韩作为.医疗机构人工智能安全风险与大模型安全加固体系构建研究[J].中国卫生信息管理杂志,2026,23(02):195-201.
基金信息:
深圳市南山区医疗集团总部与中国医学科学院阜外医院医疗人工智能心电判读团队合作框架协议(2025-SMGC01)
2026-04-20
2026-04-20