Survey data reveal widespread informal AI use among hospitalists, but cautionary experience with electronic records echoes a major warning: tools meant to streamline may intensify workload if integrated haphazardly. The study finds that AI’s benefit rides less on headlines about adoption rates and more on fit with clinical reasoning, workflow design, and ongoing support. Training clinicians in how and when to rely on diagnostic AI, coupled with implementation science, is emerging as the neglected engine of actual impact. With stuttering digital rollouts fresh in memory, the bar for meaningful improvement may rest less on technical advances and more on whether health systems invest in the unglamorous gruntwork of workflow adaptation and evaluation.
Real-World AI Use by Hospitalists: Implementation Challenges Outpace Adoption
Does the rapid adoption of LLM-based AI tools by hospitalists translate into improved care, or does implementation quality still limit their clinical impact?