Clinical AI operates within a dynamic triangle of Patient, Provider, and Clinical AI interactions, all occurring within a defined Care Setting.
Two interconnected spaces clash to define the Clinical AI Skill-Mix: What is the Clinical Competency (5C) and how AI Engagement (3A) happens within it?
Total Cells (5C × 3A) within Skill-Mix Cube:
67,504,639,104
9.3M clinical scenarios per provider × 84 AI engagement patterns
= 784.9M clinical AI scenarios per provider
Each Clinical Intelligence Cell (5C × 3A) represents a unique capability requiring distinct evaluation approaches, defined by:
- Clinical Competency Scenario — condition, care phase, setting, task, provider role
- AI Engagement — agent facing, anchoring layer, assigned authority
- Performance Metrics — accuracy, clinical safety, reliability, clinical utility, and skill preservation safety
Five dimensions defining WHAT clinical scenario
Three dimensions defining HOW AI engages
Each dimension represents a constituent element defining bounderies of clinical scenarios and AI engagement patterns.
Loading dimension data...
A philosophical exploration of medical cognition—mapping how clinicians and patients process information through iterative cycles of intuition, analysis, and reflection. The hypothesis serves as the central organizing concept, mediating between three cognitive processor engines in a dynamic, bidirectional exchange.
The physician's cognitive architecture: integrating clinical data, knowledge, and experience through dual-process reasoning.
The patient's cognitive journey: processing clinical information through values, context, and lived experience.
"Clinical reasoning is not merely the application of medical knowledge, but an iterative dialogue between intuitive pattern recognition and deliberate analytical thinking—a cognitive dance where the hypothesis serves as both guide and test of our understanding." — Clinical World Model Framework
Clinical data, context, and prior knowledge converge at the processor
Central concept emerges, mediating three processor engines (2-4 iterations)
System I (intuition) and System II (analysis) evaluate in bidirectional reflection
Proceed to planning/preference or loop back via data collection