As artificial intelligence (AI) becomes increasingly embedded in critical care settings, there is a pressing need to understand how these technologies interact with human teams responsible for high-stakes decision-making. This paper introduces a tailored Input-Mediator-Output-Input (IMOI) model to conceptualize the complex, cyclical dynamics of human-AI teaming in environments such as intensive care units and emergency departments. Building on principles from team science and information processing theory, the model identifies key inputs (e.g., AI capabilities, team composition, interface design), mediators (e.g., trust, communication, coordination), and outputs (e.g., team performance, patient outcomes), while accounting for moderating factors like clinician experience, stress, and AI transparency. A critical feature of the model is its feedback loop, through which outcomes inform future team behaviors, training, and system redesign. The paper outlines practical applications for healthcare training, AI system design, and simulation-based evaluation, offering a comprehensive roadmap for integrating AI as an adaptive, trustworthy member of clinical teams. Importantly, this model is conceptual and has not yet been empirically validated; it is intended to serve as a foundation for future empirical research. This model supports ongoing quality improvement initiatives and promotes safer, more effective human-AI collaboration in time-sensitive, high-pressure care environments.
Korentsides et al. (2026) studied this question.