In recent years, artificial intelligence (AI) has been increasingly integrated into endoscopic practice and is considered a priority area for the development of visual diagnostics. This analysis highlights developments aimed at automated detection and characterization of gastrointestinal tumors, assessment of invasion depth, monitoring the completeness and quality of examinations, standardization of photo documentation, risk stratification, and training support for novice endoscopists. The objective of this review is to summarize the current state of AI technologies in endoscopy, assess their clinical significance, and identify key limitations that hinder their widespread use in clinical practice. The data presented demonstrate that modern algorithms can improve diagnostic efficiency and reduce inter-specialist variability. The use of AI helps mitigate the influence of human factors and experience on the quality of visual examinations and can also serve as a tool for developing practical experience in novice endoscopists. This demonstrates the high potential of AI not only as a diagnostic tool but also as a means of process optimization and training support. However, there remains a need for further multicenter studies, the development of more comprehensive and integrated solutions, and the assessment of the long-term impact of modern algorithms on clinical practice.
Samsonyan et al. (Thu,) studied this question.