Pharmacovigilance is an essential system for ensuring public medication safety and for identifying, assessing, and preventing risks throughout the entire drug lifecycle. Traditional manual approaches face challenges such as low processing efficiency, high risk of missing safety signals, and complex global compliance management when dealing with massive, multi-source, and heterogeneous data. In recent years, rapid advances in artificial intelligence (AI), particularly in machine learning, deep learning, natural language processing, and generative AI, have opened new technical pathways for the intelligentization of pharmacovigilance. This article systematically reviews the current AI applications in key pharmacovigilance areas, including multi-source data collection and structuring, safety signal detection and prioritization, regulatory compliance and automated reporting, as well as individualized risk assessment and causality evaluation. It also analyzes the major challenges facing AI applications in this field, such as data quality, model generalizability, interpretability, and regulatory acceptance, and further discusses the future of intelligent pharmacovigilance systems oriented toward continuous learning, multimodal integration, and ethical compliance, so as to provide theoretical foundations and practical insights for improving the efficiency and accuracy of drug safety monitoring.