大语言模型在医疗健康领域的应用现状与前景展望.pdfVIP

医学高新技术伦理

医学与哲学2025

年6

月第46

卷第

12

期总第767期

大语言模型在医疗健康领域的

应用现状与前景展望*

1123△1△

陈紫林祝帆帆罗宇昕张伟卢龙

摘要：系统梳理了医疗大语言模型的开发方式、模型类型与应用现状，并指出其在医疗场景下所面临的独特且更为复杂的伦

理安全挑战。相较于其他人工智能技术，医疗大语言模型在数据安全与隐私保护、算法偏见与公平性、医患关系重构、责任

归属不清以及跨领域迁移中的伦理风险等方面呈现出更深层次的问题。因此，推动医疗大模型的实际落地，既需持续突破技

术瓶颈、提升生成内容的准确性与可解释性，也需正视其潜在的伦理风险与社会影响。将医疗大语言模型嵌入临床工作流程，

应遵循自主、有利、不伤害与公平四项基本伦理原则。

关键词：医疗健康大模型，实现范式，应用场景，伦理挑战

中图分类号：R-05文献标识码：A文章编号：1002-0772(2025)12-0032-06

DOI：10.12014/j.issn.1002-0772.2025.12.07

112

ApplicationsandFutureProspectsofLargeLanguageModelsinHealthcareCHENZilin,

ZHUFanfan,

LUOYuxin,

31

ZHANGWei,

LULong．1.

CollegeofInformationManagement,WuhanUniversity,Wuhan,430072China;

2.

MaynoothInterna-

tionalEngineeringCollege,FuzhouUniversity,Fuzhou350108,China;

3.

InformationCenter,GeneralHospitaloftheCentralThe-

aterCommandofthePeoplesLiberationArmy,Wuhan,430070,China

Abstract：This

paper

provides

a

systematic

review

of

the

development

approaches,

model

types,

and

current

applications

of

large

language

models

(LLMs)

in

the

healthcare

domain.

It

highlights

the

unique

and

increasingly

complex

ethical

and

safety

challenges

these

models

face

in

healthcare

contexts.

Compared

with

other

artificial

intelligence

technologies,

LLMs

exhibit

deeper

issues

related

to

data

security

and

privacy

protection,

algorithmic

bias

and

fairness,

the

reconfiguration

of

doctor–patient

relationships,

ambiguity

in

the

at