Purpose of Review <p>The intensive care unit (ICU) cares for some of the most medically complex and vulnerable patients in healthcare. These patients are particularly susceptible to developing surgical emergencies, ranging from intra-abdominal sepsis in the setting of malignancy-associated neutropenia to life-threatening gastrointestinal hemorrhage related to advanced liver disease and portal hypertension. In this review, we examine emerging artificial intelligence (AI) applications that may enhance decision-making in critically ill surgical patients, particularly in settings characterized by physiologic instability, diagnostic uncertainty, and fragmented data.</p> Recent Findings <p>AI represents a promising approach for synthesizing and interpreting the large volumes of multimodal data generated in the ICU, with the goal of producing more timely and accurate clinical insights for critically ill surgical patients. Early applications of AI have demonstrated potential across several domains, including risk prediction, diagnostic support, computer vision-enabled monitoring, and triage or resource allocation. Emerging advances in multimodal and continuously learning models may further enhance AI’s ability to capture the dynamic trajectory of critical illness and adapt to evolving practices and protocols among intensivists and surgeons.</p> Summary <p>The translation of AI into critical care and emergency surgery settings will require more than technical innovation. These systems must be supported by rigorous evidence of clinical benefit, safely incorporated into high-acuity, high-stakes clinical workflows, and implemented in ways that preserve, and ultimately augment, high-quality clinician training, expertise, and judgment.</p>