《生命科学》 2025, 37(11): 1445-1452

医疗人工智能公正性风险及其伦理治理研究综述

杜严勇^* , 朱梦皎

同济大学人文学院，上海 200092

摘　要：

人工智能技术在医疗领域的应用日益广泛，在产生大量积极效应的同时，也引发了包括公正性在内的各种伦理风险，引起了社会各界的关注。医疗人工智能引发的公正性问题主要包括人工智能作为医疗资源的分配公正、作为健康公平工具的潜在风险、在医学影像等领域产生的偏见问题，以及人工智能辅助手术的可及性等。国内外专家从提升人工智能医疗资源的公正分配、关注全过程伦理审查与监管以及积极推进并发布伦理框架与专家共识等多个方面出发，提出了一系列解决公正性风险的建议与策略。通过政府部门、科研人员、医护人员与用户在内的所有利益相关者共同努力，可以在很大程度上规避医疗人工智能的公正性风险，使其更好地为人类健康服务。

通讯作者：杜严勇 , Email:yanyongdu@tongji.edu.cn

Review of the risks of justice in medical artificial intelligence and its ethical governance

DU Yan-Yong^* , ZHU Meng-Jiao

School of Humanities, Tongji University, Shanghai 200092, China

Abstract:

Artificial intelligence (AI) has been increasingly permeating the medical domain, yielding substantial  benefits while simultaneously giving rise to a spectrum of ethical risks, particularly concerning justice and fairness. This has garnered attention from diverse sectors of society. The justice problems caused by medical AI mainly include the distributional justice of AI as a medical resource, the potential risk of AI as a tool for health equity, bias emerging in medical imaging and other diagnostic fields, and the gap of accessibility of AI-assisted surgery.  Researches have proposed a range of strategies to avoid justice risks from  multiple aspects, such as improving the equitable distribution of medical AI resources, enhancing the whole-process ethical review and oversight, and  actively promoting and issuing ethical frameworks and expert consensus. Through the concerted efforts of all takeholders, including government  departments, researchers, medical staff and patients, large part of the justice risks of medical AI can be effectively mitigated. This collaborative approach will enable AI to better serve the overarching goal of promoting human health.

Communication Author：DU Yan-Yong , Email:yanyongdu@tongji.edu.cn