Current Issues of Increasing the Explainability of AI Systems in Medicine

The paper considers current problems of increasing the explainability of artificial intelligence (AI) systems in the medical field. The main technical, methodological and ethical challenges associated with the development of transparent AI systems are analyzed. Key explainability issues are investigated, including the complexity of neural network architecture, the multidimensionality of medical data and the need to adapt explanations for different interested parties. An overview of modern approaches to increasing the interpretability of AI systems is presented, and a comparative analysis of existing methods is carried out. The critical importance of developing explainable AI technologies for ensuring trust, safety and effectiveness of medical technologies is shown.