Abstract
Atrial fibrillation is the most common heart arrhythmia globally, leading to life-threatening complications, reduced quality of life, a high financial burden, and significant healthcare resource utilization. Artificial intelligence is increasingly being integrated into medicine, enhancing clinicians’ ability to screen for, diagnose, and treat various conditions. In recent years, artificial intelligence models have been successfully applied to predict atrial fibrillation by analyzing 12-lead electrocardiogram waveforms, imaging features derived from computed tomography, cardiac magnetic resonance imaging, and echocardiography, as well as other clinical risk factors. The aim of this study is to synthesize current evidence, highlight emerging trends, and identify future directions in this field.