Volcanic caves represent complex geomechanical environments, and their structural stability is crucial for risk prevention, conservation, and sustainable tourism. Artificial Intelligence (AI) has emerged as a promising solution for optimizing monitoring tasks and enhancing safety in subway operations. This study aims to analyze the application of AI techniques for improving the stability of volcanic caves through a systematic literature review, determine the current status of these technologies, and address the challenges that hinder their potential exploitation in this field. A systematic review was conducted using the Scopus database with a search equation to understand the advances, applications, and limitations of AI-assisted techniques in volcanic caves. Within the proposed search strategies, 10 documents were obtained for the search (AI + volcanic caves) and 321 records for the search (AI + tunnels + stability). AI technologies, such as convolutional neural networks (CNN), SLAM algorithms, deep learning, computer vision, and LiDAR systems, applied in volcanic caves have proven effective in 3D model generation, traffic analysis, structural risk detection, and autonomous route planning. In terms of future trends, the use of technologies such as Explainable Artificial Intelligence (XAI), hyperspectral sensors, and generative algorithms for stability simulations is expected to grow. However, few studies have specifically addressed volcanic caves, highlighting a gap in the current research, such as the implementation of AI for stability prediction and risk assessment.