JMIR Publications

ABSTRACT

Background:

Artificial intelligence (AI) is increasingly integrated into healthcare and medical education. However, undergraduate curricula have largely positioned medical students as users of AI enabled tools rather than as developers of them. No code platforms, which allow software to be built through visual interfaces and plain language instructions instead of programming, could change this by enabling learners without technical training to create their own functional AI powered educational tools. Whether this approach is feasible in undergraduate medical education has not yet been evaluated.

Objective:

his mixed methods pilot feasibility study asked a simple question: can undergraduate medical students without prior programming experience build AI powered surgical simulation applications using no code platforms during a structured workshop? We also looked at how participants and faculty experienced the resulting prototypes.

Methods:

Eight senior undergraduate medical students at Mohammed Bin Rashid University of Medicine and Health Sciences (MBRU) took part in a single session "appathon": a 45 minute discussion followed by 1 hour of hands on development. Three teams, each using a different no code platform, built an AI enabled surgical education application. Feasibility meant the workshop ran to completion and every team produced a working prototype. Faculty independently rated each prototype using a modified System Usability Scale, a clinical accuracy and educational value checklist, and a global assessment. Participants completed a six domain questionnaire after the workshop, and open ended responses were analyzed thematically

Results:

All eight participants (8/8, 100%) finished the workshop, and all three teams (3/3) shipped working prototypes: a visual recognition tool for laparoscopic instruments (LIRA), a chatbot guided procedural simulation (Bedside Procedure Simulation), and an interactive case based decision making tool (SurgicalDx). One faculty surgical consultant evaluated all three applications (n=3 evaluations) and endorsed their clinical accuracy, educational value, and potential for further development; usability scores ranged from 3.0 to 4.5 out of 5. Every developer questionnaire came back (8/8, 100%), and agreement was near uniform across development experience, learning outcomes, AI specific learning, mentorship, perceived value, and satisfaction. Qualitatively, students described feeling empowered by independently building an application, connecting clinical knowledge to educational design, and picking up prompt engineering as a newly acquired skill.

Conclusions:

Undergraduate medical students with no prior programming experience were able to design and deploy working AI enabled surgical simulation prototypes on no code platforms, inside a single workshop, and both faculty and learners rated the results favorably. Whether this translates into measurable educational gains is a question for larger, controlled studies using objective outcome measures.