Loading...
Artificial intelligence (AI) now touches almost every aspect of clinical care (see ). In 2025, researchers published several provocative studies that were focused specifically on bedside applications — large language model (LLM) clinical decision support and ambient AI scribes.
In a randomized study of 92 clinicians, those who used ChatGPT-4 alongside conventional resources outperformed those who used standard tools alone on five complex clinical vignettes. GTP-4 users demonstrated superior diagnostic, management, and context-specific decision-making. However, LLMs still confabulate with confidence rather than stating “I don’t know.” In a set of 300 physician-validated vignettes containing deliberately fabricated clinical details, ea…
