As researchers at Stanford’s Human-Centered AI Institute havenoted, generalist large language models show promise in clinical settings but remain vulnerable to contextual errors and overgeneralization when deployed beyond controlled demonstrations. AnotherNature Medicinestudydemonstrated that both generalist and “medical-domain” large language models are not ready for autonomous clinical decision-making because they often fail to follow workflow instructions and are sensitive to input variations, underscoring the importance of structured oversight.