Building early career pathways in AI and medicine: a six-year pilot evaluation of the EDIT AI program

Artificial intelligence (AI) is rapidly reshaping biomedical research and clinical medicine, including cancer diagnostics, therapeutics, and healthcare delivery. Maximizing its translational impact will require a workforce capable not only of applying AI tools but also of understanding their underlying limitations, sources of bias, and clinical implications. Although early research experiences can strongly influence long-term participation in STEM fields, few pre-college programs integrate computer science with medicine– an emerging intersection with growing relevance. The Emerging Diagnostic and Investigative Technologies (EDIT) AI program was established in 2020 to provide high school students with mentored summer research experience at the interface of AI and cancer medicine. In this work, we describe the programmatic components of EDIT AI and evaluate three primary questions during its six-year pilot period: (1) whether participation improves familiarity with biomedical AI concepts and their clinical applications; (2) whether participation influences student interests toward emerging areas of computational oncology; and (3) whether a fully virtual research environment can sustain meaningful engagement and scholarly outcomes. We conducted a longitudinal internal review of the EDIT AI program from 2020 to 2025, during its ongoing pilot period. EDIT AI is delivered entirely virtually through a structured three-track model (Skills Development, Advanced Research, and Peer-Mentor), supported by a tiered mentorship network and the EDIT AI Virtual Student Laboratory, where students analyze real-world clinical research datasets. Educational activities include seminars, team-based research, office hours, mentor meetings, and an end-of-summer research symposium. Evaluation included participant demographics, pre- and post-program familiarity surveys, seminar attendance, thematic analyses of student interests and scholarly outputs, participant feedback, and alumni outcomes. During the six-year pilot period, the program has matriculated 284 student participants from 24 states. Participants demonstrated significant improvements in self-reported familiarity with biomedical AI concepts, with project interests and completed research increasingly reflecting emerging areas of computational oncology, including spatial biology. Participants produced 39 peer-reviewed publications, presented at national and international conferences, earned scholarships, and matriculated into college and graduate programs, with many alumni returning as near-peer mentors. This six-year internal evaluation provides preliminary evidence supporting the feasibility of a fully virtual, mentored AI-in-medicine research program for high school students. EDIT AI offers a framework for preparing early-stage learners to engage with interdisciplinary biomedical AI research while strengthening the future biomedical AI workforce.