Building BRIDGEs: Coordinating Standards, Diversity, and Ethics to Advance Biomedical AI

点击翻译按钮获取中文摘要

The NIH's Strategic Plan for Data Science sets forth a grand and ambitious goal to enhance the diversity of the data science workforce and to engage underrepresented minority communities. Indeed, the vision of an ethical and equitable society supported by modern AI/ML healthcare innovations can only be achieved when our stake- holders include representatives of all communities. In lockstep with the NIH's diversity goals, this application aims to enact long-lasting change in the biomedical AI/ML community by a key leadership workforce to advance our mission in fostering the growth of a diverse next-generation multidisciplinary cohort of physicians and inves- tigators who are the driving force enabling Bridge2AI. Briefly, we recognize a major challenge in biomedical AI/ML education that limits the widespread adoption of modern AI/ML in health care and biomedical innovations. There exists a general lack of understanding regarding ethical and trustworthy AI (ETAI) in our workforce, com- bined with public anxiety surrounding the use of AI/ML applications. These concerns stifle the potential impact of these AI/ML technologies. Accordingly, we propose to establish the Bridge2AI-ENABLE Scholar Award. Essential to the successful implementation of AI/ML strategies are the emerging underrepresented in medicine (URiM); these professionals (e.g., medical students, clinical fellows, nurses, physicians) are well-trained in their medical professions and are committed to undergo further training in AI/ML and to advance clinical practice. Our proposed Bridge2AI-ENABLE Scholar Award will offer targeted AI/ML training for up to 15 URiM medical profes- sionals who are poised to become future leaders driving AI/ML innovation in health care. This URiM enrichment activity will consist of designated mentor teams composed of clinicians, AI/ML specialists, and ETAI leaders. We have organized a comprehensive 10-week long training plan that includes: 1) a well-thought-out applicant re- cruitment plan and mentee appointment protocol; 2) a mentor team with strong commitment from UCLA School of Medicine, medical informatics, and computational medicine faculty that follow our mentor selection processes; and 3) a customized curriculum tailoring each mentee to completing their training. Overall, our training platform aims to overcome anxiety surrounding biomedical AI/ML, build community trust, and empower trainees exploring biomedical AI/ML topics with entry at ground zero. Each trainee can design their personalized curriculum and complete AI education at their own pace. The majority of the curriculum is conducted via an e-learning platform, complemented by 1-on-1 mentorship and A&Q consultation hours. Each mentee will partner with the SWD Core to customize training plans and design their personalized curriculum, which enables them to complete AI edu- cation in their own space and to carry out AI/ML projects in real-world scenarios. The mentor teams will be responsible for creating milestone reports for each mentee to steer a successful career trajectory. The goal of this supplemental activity is to support URiM medical professionals to acquire necessary understanding and skills in AI/ML and to enable them to be the driving force to advance AI strategies in modern healthcare.