Artificial intelligence in gastrointestinal endoscopy: Current applications, clinical impact, and future perspectives

Artificial intelligence (AI) is increasingly transforming gastrointestinal (GI) endoscopy by supporting lesion detection, characterization, and clinical decision-making.This narrative review summarizes current evidence on AI-assisted GI endoscopy, focusing on clinical applications, diagnostic performance, implementation challenges, and future directions.Machine learning and deep learning underpin computer-aided detection systems that identify suspicious lesions and computer-aided diagnosis systems that support their characterization.In colorectal cancer screening, the reviewed evidence highlights improvements in adenoma detection and the potential of AI to support optical diagnosis of colorectal polyps.In upper GI endoscopy, promising applications include the detection of Barrett's esophagus-associated neoplasia, early gastric cancer, and other premalignant lesions.Integration with image-enhanced endoscopy further supports real-time tissue characterization and risk stratification.Beyond lesion recognition, AI applications extend to procedural planning, quality assessment, and clinical decision support.Across these applications, data quality, external validation, and algorithm transparency remain central considerations when interpreting diagnostic performance and assessing clinical applicability.The review therefore emphasizes that successful implementation requires both technical development and careful evaluation within clinical workflows, alongside appropriate ethical and regulatory oversight.AI has substantial potential to advance precision endoscopy; however, its contribution to improved patient outcomes requires further clinical evaluation.