Artificial intelligence in metabolic and bariatric surgery: current applications and future direction from the ASMBS Emerging Technology and Procedures Committee.

Abstract

Background: Artificial intelligence (AI) is increasingly integrated across the perioperative continuum of metabolic and bariatric surgery (MBS), yet adoption is outpacing specialty-specific guidance on safe, effective, and equitable implementation.

Objectives: To synthesize current evidence on AI applications in MBS, identify key implementation challenges, and provide recommendations for responsible clinical adoption.

Setting: American Society for Metabolic and Bariatric Surgery (ASMBS) Emerging Technology and Procedures (ET&P) Committee.

Methods: Surgeon-led narrative review of published literature on AI in MBS and relevant adjacent surgical and chronic disease fields, organized by preoperative, intraoperative, postoperative, and administrative domains, with recommendations developed by committee consensus.

Results: Preoperatively, machine learning models outperform conventional risk indices for predicting complications and readmission and can forecast long-term weight trajectories. Intraoperatively, computer vision systems recognize procedural steps in sleeve gastrectomy and Roux-en-Y gastric bypass with accuracy comparable to expert annotators, and AI-enabled video and virtual reality platforms support objective skills assessment. Postoperatively, wearable-based remote monitoring and predictive models may identify early deterioration, weight regain, and comorbidity recurrence. Administratively, ambient AI scribes, prior authorization automation, and operating room scheduling may reduce clerical burden. Most evidence remains retrospective and single-center. Key challenges include large language model hallucination, automation bias and deskilling, algorithmic bias, and data privacy. Eight operational recommendations address clinical role, MBS-specific validation, equity auditing, documentation, governance, technical safety, workflow integration, and learning health systems.

Conclusions: AI can enhance the safety, efficiency, and scalability of MBS when clinician-led, rigorously validated, equitable, and transparent. Prospective, multicenter evaluation is needed before broad adoption.