CAP: Capacity Building for Trustworthy AI in Medical Systems (TAIMS)

该项目是一个扩展人工智能能力建设试点(CAP)，专注于通过在值得信赖、道德、可解释和负责任的人工智能(AI)方面进行重要的受使用启发的研究，在Meharry医学院建立和发展与AI相关的活动，以缓解AI支持的医疗系统中的算法偏差问题。现有的临床人工智能方法在系统设计和开发过程中受到伴随的算法和社会偏见的限制，导致误诊，最终导致糟糕的治疗和健康差距。梅哈里医学院(MMC)的研究人员将通过加强对“黑匣子”人工智能算法本质的理解，为安全、隐私和可靠的无偏见临床决策开发可解释的方法和工具。该项目是梅哈里牙科学院、研究生院和应用计算科学学院(SACS)的合作努力，旨在为美国研究生和医疗专业人员提供人工智能和机器学习(ML)技术方面的高级培训，以帮助他们为未来人工智能驱动的职业道路准备技能。研究能力建设计划包括对不同教师进行使用尖端AI/ML技术的培训。在教育方面，将设立短期课程、暑期学校和辅导系列，以涵盖重要主题，如人工智能/ML伦理概述、可解释的人工智能方法、算法偏差和缓解策略以及数据隐私。几个研究主题被确定为整个Meharry园区人工智能技术的潜在增长领域，包括(I)医疗系统中的可解释人工智能，(Ii)医疗系统的伦理和负责任的人工智能。例如，可解释人工智能系统是指人类更好地理解人工智能系统决策背后的推理，而不是预测准确性和统计性能。许多从业者、临床医生、研究人员和患者都不愿使用人工智能，除非它是可解释、可验证和可信的。合乎道德和负责任的人工智能涉及为合乎道德地使用人工智能工具和技术建立定义明确的指导方针和法律法规。人工智能工具的道德使用对于维护基本人权、健康和公共安全至关重要。教育能力建设计划通过几项举措进行，包括：(I)在医疗专业人员、研究生、教职员工中传播认识的培训活动；(Ii)短期课程和教程系列，涵盖人工智能/ML中的伦理概述、可解释的人工智能方法、算法偏差和缓解；(Iii)研究生课程中侧重于值得信赖的人工智能的新课程；(Iv)面向K-12学生的高中教育工作者和暑期学校。通过为广大不同的教职员工、专业人员、教育工作者和学生提供人工智能知识和身临其境的学习体验，人工智能素养和技能发展将得到显著提高。扩展人工智能计划支持少数族裔服务机构的人工智能驱动的教育和劳动力发展、基础设施和研究，以加强和多样化美国的研究和教育途径，并在STEM职业生涯中为历史上被边缘化的社区提供新的机会。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

This project is an ExpandAI Capacity building pilot (CAP), which focuses on establishing and growing AI related activities at Meharry Medical College by conducting important use-inspired research in trustworthy, ethical, explainable and responsible artificial intelligence (AI) for the mitigation of the problem of algorithmic bias in AI-powered medical systems. Existing clinical AI methods are limited by attendant algorithmic and societal biases during systems design and development resulting in misdiagnosis and ultimately poor treatment and health disparities. Through enhanced understanding of the nature of “black box” AI algorithms, researchers at Meharry Medical College (MMC) will develop interpretable methods and tools for secure, private and reliable bias-free clinical decisions. The project is a collaborative effort between Meharry’s School of Dentistry, School of Graduate Studies, and School of Applied Computational Sciences (SACS) to provide American graduate students and medical professionals, with advanced training in AI and machine learning (ML) techniques to prepare them with skills for future AI-powered career pathways. Research capacity building plans include training of diverse faculty in the use of cutting-edge AI/ML techniques. On the educational side, short courses, summer school and tutorial series will be established to cover important topics such as overview of ethics in AI/ML, explainable AI methods, algorithmic bias and mitigation strategies and data privacy. Several research themes were identified as potential growth areas for AI technologies across the Meharry campus including (i) explainable AI in medical systems, (ii) ethical and responsible AI for medical systems. For example, explainable AI systems are ones in which humans better understand the reasoning behind decisions made by AI systems other than predictive accuracy and statistical performance. Many practitioners, clinicians, researchers, and patients are reluctant to use AI unless it is explainable, verifiable, and trustable. Ethical and responsible AI deals with establishment of well-defined guidelines and legal regulations for ethical use of AI tools and technologies. The ethical use of AI tools is critical for preservation of fundamental human rights, health and public safety. Educational capacity building is planned through several initiatives including (i) training activities to disseminate awareness among medical professionals, graduate students, staff, and faculty; (ii) short course and tutorial series to cover topics such as overview of ethics in AI/ML, explainable AI methods, algorithmic bias and mitigation; (iii) new course focusing on Trustworthy AI in medical systems for the graduate programs; (iv) outreach to high school educators and summer academies for K-12 students. By providing a broad swath of diverse faculty, professionals, educators and students with AI knowledge and immersive learning experiences, AI literacy and skill development, will be significantly enhanced. The ExpandAI Program supports AI-powered education and workforce development, infrastructure and research at Minority Serving Institutions to strengthen and diversify U.S. research and education pathways and provide historically marginalized communities with new opportunities in STEM careers.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.