全球药品监管机构药品AI应用监管策略与方法  

Regulatory Strategies and Methods for Pharmaceutical AI Applications from Global Drug Regulatory Agencies

作　　者：骆实 曹凤朝 韩亮 LUO Shi;CAO Fengchao;HAN Liang(Beijing Shangzhi Hegui Science and Technology Co.,Ltd.,Beijing 100101,China)

机构地区：[1]北京尚质合规科技有限公司,北京100101

出　　处：《中国医药导刊》2025年第2期128-134,共7页Chinese Journal of Medicinal Guide

摘　　要：随着人工智能(AI)技术在制药行业的广泛应用,全球主要药品监管机构面临着前所未有的AI应用监管挑战。本研究在梳理AI技术在药品生命周期中的应用及全球主要监管机构AI监管法律法规和指南文件体系的基础上,重点分析了各监管机构在AI监管中的具体策略与方法,包括美国FDA将药品生命周期AI监管分为3个核心模块,包括人类主导的治理、责任和透明度;数据的质量、可靠性和代表性;模型的开发、性能评估、监控和验证。欧盟EMA强调AI技术在药品生命周期中的应用风险,特别是在数据质量、模型开发、性能评估和可解释性方面的监管要求。WHO则从全球视角出发,提出AI在药品研发和供应中的主要风险,并划分6个AI监管领域,包括文档和透明度、风险管理和AI系统开发生命周期方法、预期用途和临床验证、数据质量、隐私和数据保护等。通过对比分析,本研究总结全球主要药品监管机构在AI监管中的共通点与差异,提出基于风险的分类与分级、基于设计的监管策略、AI输出内容的标识、适用于AI的验证流程、持续性监管以及人类责任制等监管策略与办法,其实施将有助于建立一个更加稳健和可信的药品AI应用环境,确保药品的安全性、有效性和质量可控性,同时促进制药行业的创新发展。With the widespread application of artificial intelligence(AI)technology in the pharmaceutical industry,global major drug regulatory agencies are facing AI application regulatory challenges.Based on reviewing the application of AI technology in the drug life cycle and the system of AI regulatory laws,regulations and guidelines of major regulatory agencies around the world on regulation of AI application,this study analyzes the specific strategies and methods adopted by these agencies in AI regulation.For instance,the U.S.FDA divides into 3 core modules about the drug life cycle AI governance,including human-led governance,accountability,and transpar-ency;data quality,reliability and representativeness;model development,performance evaluation,monitoring and validation.The European EMA emphasizes the risks associated with AI applications in the drug life cycle,particularly in terms of data quality,model development,performance evaluation,and interpretability.Meanwhile,from a global perspective,WHO identifies the main risks of AI in drug development and supply,and divides AI regulation into six key areas,including documentation and transparency,risk manage-ment and AI system lifecycle approaches,intended use and clinical validation,data quality,privacy,and data protection,etc.Through comparative analysis,this study summarizes the commonalities and differences in AI regulation among global major drug regulatory agen-cies and proposes regulatory strategies and methods such as risk-based classification and grading,design-based regulatory strategies,labeling of AI-generated outputs,AI-specific validation processes,continuous regulation,and human accountability.The implementa-tion of these strategies will help establish a more robust and trustworthy environment for AI applications in pharmaceutical,ensuring drug safety,efficacy,and quality control,while promoting innovation in the pharmaceutical industry.

关 键 词：药品监管 药品生命周期 人工智能 监管策略 监管方法 

分 类 号：R95[医药卫生—药学]