Suggested ICAP Fit
This proposal connects well with the applied focus of ICAP 2026 by examining a practical training model for community health education in a low-resource setting. It relates to applied psychology through its focus on learning, confidence, feedback, skill development, and the transfer of training into practice. It also reflects the conference theme of New Directions in Applied Psychology by considering how AI-supported instruction can be combined with human-led simulation and train-the-trainer approaches.
Abstract
In low-resource and crisis-affected settings, limited access to health-care services and structured skills training can reduce individuals' ability to respond effectively to medical emergencies. Community health education therefore requires learning models that are practical, scalable, and feasible under resource constraints. Applied psychology can contribute to this challenge by informing how people learn, build confidence, receive feedback, sustain motivation, and transfer knowledge into practice.
This continuing program of work, conducted by Seattle Pacific University in collaboration with the University of Geneva, Switzerland, examines how AI-supported first aid instruction can move from digital learning into applied practice. The program includes two connected components: first, an online basic first aid module supported by AI-powered avatars; second, a human-led simulation and train-the-trainer model designed to strengthen applied skills, local facilitation, feedback, debriefing, and training capacity.
The empirical evaluation of the AI-supported first aid module used a mixed-methods design. Quantitative data included 18 matched pre- and post-training surveys and 43 post-only surveys. Outcomes included first aid knowledge, attitudes toward first aid, perceived learning efficacy, persistence, and learner experience with AI avatar-based instruction. Findings showed small positive trends in first aid knowledge, Cohen's d = 0.29; perceived learning efficacy, Cohen's d = 0.15; and attitudes toward first aid, Cohen's d = 0.22. Persistence showed a slight negative trend, Cohen's d = -0.15. These changes were not statistically significant, p > .05. Qualitative findings suggested that avatar visualization and interactive demonstrations supported learner understanding and engagement.
The applied implementation component extended this work through human-led simulation and trainer development. An AI-supported avatar lesson provided standardized foundational instruction, while four local graduates and four Kenyan health professionals participated in trainer development. Trainers learned to facilitate simulations, role-plays, feedback, debriefings, and local adaptation. Across two cohorts, 57 learners started and 38 fully attended and completed the simulation module assessment in Cohort 1; 27 started and 25 fully attended and completed the module assessment in Cohort 2.
Together, these findings suggest that AI can support consistency, access, and scale, but human facilitators remain essential for feedback, trust, contextual adaptation, and skill development. This work contributes to applied psychology by offering a practical model for responsible AI use in health education, emergency preparedness, workforce development, and technology-supported training in low-resource community health settings.