Large language models in perioperative anesthesia: a risk-based roadmap for safe clinical adoption

Large language models, Perioperative medicine, Patient safetyPerioperative anesthesia is not simply another setting in which a fluent chatbot might be useful.Anesthesiologists must integrate preoperative history, laboratory results, procedural context, rapidly changing physiology, drug administration, airway events, surgical stimulation, and postoperative recovery under severe time constraints.Recent reviews have catalogued potential applications of large language models (LLMs) and broader artificial intelligence (AI) in anesthesiology [1-6].Our position is more specific: clinical readiness should be judged by the consequences of failure in the intended use, not by the fluency of the output.We therefore propose a risk-based framework in which lower-risk information-processing uses may be automated, interpretive uses should augment clinicians, and time-critical or high-consequence decisions should not currently be delegated to LLMs (Fig. 1).