Abstract / Summary
Artificial intelligence (AI) in gastrointestinal (GI) endoscopy has evolved from its early application in simple edge detection to highly sophisticated technologies designed to improve detection of neoplastic and preneoplastic lesions and support clinical decision-making. However, this rapid technological evolution has generated a “ data tsunami ” or “ datanami ”: an overwhelming and continuously expanding volume of information that, frequently, exceeds the ability of clinicians and patients to process, interpret, and apply it effectively. Substantial investments in AI have been driven by expectations of improved efficiency and reduced healthcare costs. Nevertheless, the metrics used to justify the adoption of AI-assisted diagnostic and therapeutic strategies remain inadequately standardized, leaving the true cost-effectiveness of these technologies uncertain. Current AI research in gastroenterology has primarily focused on improving diagnostic performance rather than evaluating long-term clinical outcomes or economic value. To address this limitation, investigators employ stochastic microsimulation models, which simulate individual patient trajectories and enable more robust assessments of cost-effectiveness. As the field progresses from AI to extended intelligence and, ultimately, intelligent assistance, the role of humans becomes increasingly central. Rather than replacing clinical judgment, AI should augment human expertise, with healthcare professionals serving as the critical link between technological innovation and meaningful clinical benefit. This review examines whether AI applications in GI diseases can be implemented cost-effectively in routine clinical practice for diagnosis and treatment. It further explores the indispensable human dimension of AI adoption, emphasizing accountability, reliability, ethical oversight, and responsible implementation as essential safeguards for current and future clinical applications.