Artificial intelligence applications in cardiology: a review

The review article considers key applications of artificial intelligence (AI) in cardiology. The review includes subsections devoted to weak and strong AI used in clinical practice and cardiology health provision. The article describes the application options for AI in the analysis of electrocardiography, echocardiography, sonography, computed tomography, magnetic resonance imaging, and positron emission tomography of the heart data. The article briefly describes the aspects of using machine learning and artificial intelligence to process ambulance calls from patients with cardiac complaints, and considers AI applications in preventive cardiology. The review considers the potential of AI in the analysis of data arrays obtained during tonometry, pulse wave velocity measurement, and in biochemical studies. The paper also formulates the principles of strong AI (large language models) in cardiology health provision, identifies the main problems and difficulties in implementing the latest technology, and provides a conceptual scheme for implementing AI technology in a cardiology center. This paper highlights the key limitations of the large language model technology, such as the lack of standard algorithms for collecting and reviewing data, lack of understanding of the context, the inability of models to form expert conclusions, and the emergence of many problematic ethical characteristics when using large language models.