An interpretable machine learning model was reported to predict 7-, 14-, and 28-day mortality in sepsis-associated liver injury, outperforming modified SOFA and MELD-Na in an independent Chinese cohort. The work positions structured risk stratification as a core next step for sepsis organ-dysfunction management. A separate ICU-focused study described an AI model that forecasts vital signs and flags kidney injury earlier, targeting the time-sensitive window in critical care where earlier intervention can change downstream outcomes. For clinical and commercial stakeholders, these updates show growing convergence on explainability plus operational deployment—risk tools that can be integrated into ICU workflows.