医疗人工智能安全监管主体责任调查研究  

Investigation on Entity Responsibility of Medical Artificial Intelligence Safety Supervision

作　　者：茅鸯对[1] 沈茜[1] 王维标[1] 丁静[1] MAO Yangdui;SHEN Xi;WANG Weibiao;DING Jing(Zhejiang Pharmaceutical College,Ningbo,Zhejiang,China 315100)

机构地区：[1]浙江医药高等专科学校,浙江宁波315000

出　　处：《中国药业》2020年第20期9-12,共4页China Pharmaceuticals

基　　金：2019年度浙江省哲学社会科学规划课题[19NDJC404YBM];2019年浙江省教育厅访问工程师校企合作项目[FG2019006]。

摘　　要：目的探讨医疗人工智能(AI)诊断产品上市后的监管工作思路。方法采用问卷调查方式,针对医疗AI诊断产品主体责任焦点问题开展实证研究,分析相关问题,采用因子分析法构建主体责任评价指标体系。结果共发放155份问卷,回收150份,有效回收率为96.77%;医疗AI诊断主体责任相关问题包括产品潜在风险多、法律主体资格有争议、责任主体不单一、侵权责任需由多方承担等;梳理出13个医疗AI诊断主体责任评价指标,成功建立相应评价体系。结论建议赋予强医疗AI有限法律主体资格,应用时遵循权责一致原则,基于主体责任评价开展分类分级监管。Objective To explore the ideas of post-market supervision on the medical artificial intelligence(AI)diagnostic products.Methods By means of questionnaire survey and interview,the entity responsibility of medical artificial intelligence diagnosis products was analyzed,and the entity responsibility evaluation index system was constructed by factor analysis.Results A total of 155 questionnaires were sent out,and 150 were collected with an effective collection rate of 96.77%.Issues related to the responsibility of medical AI diagnosis products included numerous potential application risks,disputed legal subject qualification,the non-unique subject of legal responsibility,compound liability,etc.A total of 13 medical AI diagnosis subjects were sorted out and corresponding evaluation system was successfully established.Conclusion It is recommended to granted limited legal subject qualification to the strong medical AI,and follow the principle of unanimity of rights and responsibilities in the application,and conduct classification and grading supervision basing on the assessment of the subject responsibility.

关 键 词：医疗人工智能诊断 主体责任 评价指标体系 医疗监管 

分 类 号：R955[医药卫生—药学]