Background: Pharmacovigilance (PV) the science of detecting, assessing, understanding, and preventing adverse drug reactions (ADRs) is a critical component of post-market drug safety surveillance. Traditional PV systems, including spontaneous reporting databases such as FDA's FAERS and WHO's VigiBase, are hampered by significant underreporting, delays, and resource-intensive manual processing. Artificial intelligence (AI) has emerged as a transformative solution to overcome these limitations. Methods: A systematic literature review was conducted using PubMed, Scopus, Embase, and WHO Global ICSR databases. Studies published between 2012 and 2024 using AI, machine learning (ML), or natural language processing (NLP) for pharmacovigilance applications were included. A total of 312 studies met inclusion criteria after title, abstract, and full-text screening. Results: AI demonstrated significant improvements in ADR detection across multiple data sources: social media (83%), electronic health records (75%), scientific literature (70%), spontaneous reports (87%), and patient forums (63%). NLP-based systems showed superior performance in extracting ADR mentions from unstructured text. Deep learning models achieved F1-scores above 0.80 in signal detection tasks. AI enabled real-time global pharmacovigilance at scale, previously impossible with manual methods. Conclusion: AI represents a paradigm shift in pharmacovigilance, enabling proactive, real-time, and comprehensive drug safety monitoring. Standardization of data formats, regulatory harmonization, and explainability of AI models remain key challenges. Collaborative frameworks between regulatory agencies, pharmaceutical companies, and AI developers are essential for responsible implementation.