Dilemmas and countermeasures for ethical governance of large multimodal models in the medical field
10.12026/j.issn.1001-8565.2025.09.06
- VernacularTitle:医疗领域大型多模态模型伦理治理的困境与策略
- Author:
Yanna MAO
1
Author Information
1. School of Health Management， Southern Medical University， Guangzhou 510515， China
- Publication Type:Journal Article
- Keywords:
large multimodal models;
ethical governance;
patient safety;
accountability framework;
sustainable development
- From:
Chinese Medical Ethics
2025;38(9):1133-1140
- CountryChina
- Language:Chinese
-
Abstract:
The application of large multimodal models （LMMs） in the medical field has significantly enhanced efficiency， yet it is also accompanied by ethical and safety risks. In terms of medical quality assurance， LMMs face risks such as poor data quality， over-reliance， and skill degradation. Regarding patient safety， there are risks of information misunderstanding and privacy infringement. With respect to responsibility and accountability， there are risks of vague clinical responsibility boundaries and incomplete legal frameworks. As for sustainable medical development， there is a lack of governance and evaluation systems and insufficient monitoring of environmental impacts and climate change. To address these challenges， this paper proposed a series of ethical governance countermeasures. In terms of medical quality assurance， it was necessary to improve the quality of training data， optimize data protection strategies， strengthen cross-departmental collaboration and supervision， enhance the professionalism of data annotation， and establish ethical review mechanisms. Regarding patient safety， it was crucial to improve public digital literacy， establish transparent data processing mechanisms， promote public participation in ethical governance， increase the transparency and interpretability of output results， and encourage empirical research in the academic community. With respect to responsibility and accountability， it was essential to establish a clear framework for responsibility allocation， strengthen the distinction between clinical and AI responsibilities， reinforce legal regulations and policy support， and develop a qualification certification system for users. As for sustainable medical development， it was essential to establish a systematic governance evaluation system， promote social experiments and third-party monitoring， predict the human resource needs of medical development， monitor and warn about the impacts of climate change on health， and enhance cooperation and communication among stakeholders. In summary， the application of LMMs in the medical field requires comprehensive consideration of factors such as ethics， safety， responsibility， and sustainable development. Through multi-party co-governance and systematic governance， it ensures that medical quality and patients’ rights and interests are safeguarded while advancing technology.
