Abstract:Obesity has become a global public health challenge, and metabolic and bariatric surgery (MBS) remains one of the most effective treatments for severe obesity. However, substantial variability in patient characteristics, surgical complexity, and postoperative adherence leads to heterogeneous outcomes. The rapid evolution of artificial intelligence (AI) offers new opportunities to address these limitations. By integrating multidimensional clinical, imaging, and longitudinal follow-up data, machine learning and large language models support key aspects of MBS, including candidate selection, surgical decision-making, perioperative risk prediction, skill assessment, and long-term outcome management. Recent studies have demonstrated notable progress in decision support, complication forecasting, robotic surgery optimization, patient counselling, and postoperative weight-trajectory prediction. Nevertheless, challenges remain regarding model generalizability, ethical and regulatory oversight, data privacy, and transparency in AI-assisted decision-making. This review summarizes current advances, limitations, and future directions of AI applications in MBS, providing a reference for clinicians seeking to understand and apply these emerging technologies.