走向安全的人工智能临床医生:对医疗保健中的大型语言模型越狱的全面研究
Hang Zhang1, Qian Lou2, Yanshan Wang1,3,4,5,6
1Intelligent Systems Program, School of Computing and Information, University of Pittsburgh, PA.
AMIA ... Annual Symposium proceedings. AMIA Symposium
|February 23, 2026
概括
大型语言模型 (LLM) 显示了对医疗越狱攻击的重大漏洞. 持续微调 (CFT) 提供了一个防御,强调了在医疗保健中需要特定领域的人工智能安全性的需要.
科学领域:
- 人工智能的人工智能
- 医疗信息学 医疗信息学
- 网络安全 网络安全
背景情况:
- 大型语言模型 (LLM) 越来越多地融入医疗保健.
- 在临床环境中部署引发了严重的安全问题,特别是关于错误信息.
- 在医疗环境中评估LLM漏洞对于安全的AI采用至关重要.
研究的目的:
- 系统地评估七个LLM在医疗场景中对先进的越狱技术的易感性.
- 引入一个自动化,域调整的代理管道来量化攻击的有效性.
- 调查持续微调 (CFT) 作为对抗医学对手攻击的防御机制.
主要方法:
- 通过使用三种先进的越狱破解技术,对七个LLM进行了黑盒评估.
- 开发并使用自动化,适应领域的代理评估管道.
- 评估了持续微调 (CFT) 在缓解已识别的漏洞方面的有效性.
主要成果:
- 领先的商业和开源LLM显示了对医学越狱的高度脆弱性.
- 拟议的评估管道有效量化了对抗性攻击的影响.
- 持续微调 (CFT) 在增强LLM对医疗特定攻击的弹性方面表现有前途.
结论:
- 目前的LLM容易受到复杂的医疗越狱,对患者的安全构成风险.
- 对AI临床医生来说,特定领域的安全调整和强大的评估方法是必不可少的.
- 平衡人工智能的安全性和实用性对于在医疗保健中道德和有效地部署人工智能至关重要.
相关概念视频
Issues And Trends In Healthcare Delivery System
6.3K
The issues and trends in healthcare delivery are constantly changing. The COVID-19 pandemic is one recent issue that wreaked havoc on healthcare systems, causing a shortage of healthcare workers, high demand for medicines and supplies, and increased medical expenditure due to a lack of insurance. Other issues include rising healthcare costs and care fragmentation.
Cost Containment
Payment for healthcare services has historically promoted adoption of costly and often unnecessary or inefficient...
Cost Containment
Payment for healthcare services has historically promoted adoption of costly and often unnecessary or inefficient...
6.3K
Healthcare Agencies II
1.1K
There are various healthcare agencies in the United States—some of which are managed by religious institutions and others by different government branches.
Parish nursing is a growing specialty nursing profession that focuses on holistic healthcare, health promotion, and illness prevention. It blends professional nursing practice with a health ministry, focusing on health and healing within the context of a Christian community. Parish nurses serve as health educators, referral sources,...
Parish nursing is a growing specialty nursing profession that focuses on holistic healthcare, health promotion, and illness prevention. It blends professional nursing practice with a health ministry, focusing on health and healing within the context of a Christian community. Parish nurses serve as health educators, referral sources,...
1.1K
Nursing Clinical Information System
1.3K
Nursing Clinical Information System (NCIS)
A Nursing Clinical Information System (NCIS) is a specialized type of healthcare information system tailored to meet the unique needs of nursing practice. It incorporates the principles of nursing informatics to streamline information management and improve the quality of care delivery.
Critical attributes of NCIS include:
A Nursing Clinical Information System (NCIS) is a specialized type of healthcare information system tailored to meet the unique needs of nursing practice. It incorporates the principles of nursing informatics to streamline information management and improve the quality of care delivery.
Critical attributes of NCIS include:
1.3K
Current Trends in Nursing II
3.6K
Trends in nursing are multifactorial and associated with changes in society, within the nursing profession, and in other professions. Notably, telehealth and remote nursing contribute to successful healthcare delivery for numerous patients and help reduce stress for nurses due to nursing shortages. Nurses can reach patients, monitor their conditions, and interact with them using computers, audio, visual accessories, and telephones—for example, remote patient monitoring systems. Likewise,...
3.6K
Documentation in Long-Term and Home Healthcare Setting
1.6K
Documentation in long-term care facilities and home healthcare settings is crucial for ensuring continuous, coordinated, and comprehensive care for patients. Each setting has its specific documentation processes and tools:
Long-Term Care Facilities
Long-Term Care Facilities
1.6K
Integrated Healthcare System
2.5K
An integrated healthcare system (IHS) is a set of organizations that provides for or arranges to provide coordinated and continuous service to a defined population. The IHS takes responsibility for that particular population's health status and outcome, both clinically and fiscally. An integrated healthcare system is a well-organized, well-coordinated, and collaborative network. The integrated delivery system is a network that connects different healthcare providers to deliver organized,...
2.5K


