Artificial intelligence (AI) is increasingly recognized as a valuable tool for improving the diagnosis of endometrial pathologies, particularly where conventional imaging and histopathological methods remain limited by subjectivity, interobserver variability, and restricted access to expert evaluation. This narrative review presents current technological advances in machine learning (ML) and deep learning (DL) applied to hysteroscopic imaging, cytology, ultrasound, magnetic resonance imaging (MRI), and multi-omics datasets. AI-based diagnostic systems achieve high accuracy in detecting and classifying endometrial polyps, hyperplasia, and malignant lesions, with several models performing at levels comparable to experienced clinicians. Significant progress has been made in areas such as hysteroscopic optical biopsy, automated cytological assessment, and MRI-based segmentation. Despite these achievements, contemporary studies are often limited by small datasets, single-center methodologies, insufficient external validation, and heterogeneity in imaging protocols, which reduce their generalizability and clinical applicability. Additional challenges include limited model interpretability, reduced robustness in visually subtle or atypical cases, and difficulties integrating AI systems into routine diagnostic workflows. Future development should focus on large-scale prospective validation, explainable AI methods, and multimodal diagnostic frameworks that combine imaging data with clinical and molecular information. These approaches have the potential to enhance early detection, diagnostic reproducibility, and individualized risk assessment in endometrial disease.