医疗领域ChatGPT-4o的法律归责问题研究  

Research on legal imputation of ChatGPT-4o in medical field

作　　者：徐梦洁 陈秀萍[1] XU Mengjie;CHEN Xiuping(School of Law,Hohai University,Nanjing Jiangsu 211100,China)

机构地区：[1]河海大学法学院,江苏南京211100

出　　处：《卫生软科学》2025年第3期42-48,共7页Soft Science of Health

基　　金：河海大学中央高校基本科研业务费项目(2019B36914)。

摘　　要：介绍了医疗领域ChatGPT-4o凭借实时视频对话能力和卓越的检索生成能力,于医疗数据管理、辅助临床决策及推动数字医疗普及方面展现出的广阔应用前景;分析了在其应用中伴随的数据泄露、不当处理风险和智能决策错误致人损害风险等问题。提出以下建议:在事后责任承担层面上,应依据民行刑共治和多元主体的理念,要求医务人员、医疗机构和医疗人工智能研发者在符合条件的情况下分别承担民事责任、行政责任和刑事责任;在事前风险预防层面上,应构建全生命周期医疗数据合规管理机制、算法解释权制度和算法备案体系,将算法设计、运行逻辑和隐患纳入法律监管框架以全面防范风险发生。This article introduces broad application prospects of ChatGPT-4o in medical data management,assisting clinical decision-making,and promoting the popularization of digital healthcare by virtue of its real-time video conversation capability and exceptional retrieval and generation capabilities.The problems associated with its application are analyzed,such as data breaches,risk of improper handling and risk of human damage caused by intelligent decision-making errors.The following suggestions are put forward.In terms of responsibility after the event,medical personnel,medical institutions and developers of medical artificial intelligence should be required to bear civil liability,administrative liability and criminal liability respectively under certain conditions based on the concepts of joint administration of civil,administrative and criminal laws and multiple subjects.In terms of pre-event risk prevention,it is necessary to build a compliance management mechanism of medical data full lifecycle,system of algorithm interpretation right,and algorithm filing system.Algorithm design,operational logic and potential risks are incorporated into the legal and regulatory framework to comprehensively prevent risks.

关 键 词：ChatGPT-4o 医疗人工智能 责任承担 民行刑共治 风险预防 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] D92[自动化与计算机技术—控制科学与工程] R197[政治法律—法学]