Abstract
Artificial intelligence (AI) in arthroplasty care enhances preoperative planning, intraoperative preparation, and postoperative management across hip and knee replacement pathways. Preoperatively, AI recognizes implant designs on plain radiographs, supports revision planning, and can estimate the hip joint center, improving measurements and logistics for complex cases. Clinically, models assist with patient selection and risk stratification, helping forecast readmission or reoperation and guiding resource use. Postoperatively, supervised and ensemble learning predict complications, pain trajectories, dissatisfaction, attainment of minimally clinically important differences in patient-reported outcome measures (PROMs), and even prolonged opioid use, enabling earlier, personalized interventions. Reported benefits include faster, more accurate implant identification for revision surgery and tailored care pathways that can reduce time and cost in typically older, comorbid populations. However, widespread adoption requires rigorous external validation, transparent performance reporting, and reproducible evaluation standards to ensure models are calibrated, transportable, and safe across settings. Overall, the evidence indicates that integrating AI with clinical and imaging data can streamline decisions from preoperative risk assessment to postoperative follow-up, while highlighting the need for stronger multicenter studies, data standards, and governance to translate promising results into reliable, equitable clinical tools.