Fast and Slow: How AI Is Transforming High-Stakes Clinical Problem Solving and Decision-Making

This study investigates how artificial intelligence (AI) technologies transform high-stakes clinical problem-solving and decision-making in stroke care. Drawing on a five-year qualitative study at three major stroke hospitals in the UK, we examine the adoption of a leading-edge AI application for stroke diagnosis. Our findings reveal that the pressure to save lives and solve complex medical problems triggers the emergence of two distinct sets of practices: subitizing and scrutinizing. These practices illustrate how AI transforms clinical problem-solving by enabling professionals to navigate the dual pressures of speed and accuracy in stroke diagnosis and treatment. Our ethnographic approach, involving observations, interviews, and document analysis, captures the nuanced ways in which AI mediates clinical decision-making in this high-stakes context. This study contributes to our understanding of human-AI collaboration in professional settings and offers insights into how AI can be effectively integrated into clinical practice to potentially improve patient outcomes and resource allocation.