Artificial intelligence-driven triage and decision-making in trauma systems/settings

PURPOSE OF REVIEW: Trauma care demands rapid decision-making under uncertainty and time pressure, where maintaining situational awareness becomes challenging. Artificial intelligence offers potential to augment clinical judgment by processing complex physiological signals, predicting clinical trajectory, and enhancing shared mental models across care phases. This review examines recent developments in artificial intelligence-enabled triage and decision support across prehospital, emergency department, and mass-casualty settings. RECENT FINDINGS: Machine learning models now predict lifesaving intervention needs from early prehospital data with performance comparable to expert judgment. Deep learning systems detect intracranial hemorrhage in real time, while dynamic risk scores continuously update mortality predictions during resuscitation. Artificial intelligence-enabled worklist triage accelerates critical imaging interpretation and procedural team activation. In mass-casualty scenarios, wearable sensors with automated triage algorithms maintain live patient classification, while drone-based computer vision enables contactless vital sign assessment. However, prospective evaluations reveal variable translation to measurable clinical benefit, requiring careful attention to usability, workflow integration, and ongoing calibration monitoring. SUMMARY: Artificial intelligence has the potential to strengthen trauma triage by supporting cognitive work under pressure, but realizing this benefit requires rigorous evaluation extending beyond discrimination metrics to include calibration, clinical utility analysis, and human factors assessment. The most meaningful role for artificial intelligence is augmentation rather than automation, protecting clinicians' cognitive capacity for sound judgment.